Compare commits

..

93 Commits

Author SHA1 Message Date
Kim Morrison
437b83e240 . 2025-01-06 21:09:02 +11:00
Kim Morrison
ceed9867d3 feat: Array lemma alignment; fold and map 2025-01-06 21:08:39 +11:00
Kim Morrison
78ddee9112 feat: release checklist script (#6542)
This PR introduces a script that automates checking whether major
downstream repositories have been updated for a new toolchain release.

Sample output:
```
% ./release_checklist.py v4.16.0-rc1

Repository: Batteries
   On compatible toolchain (>= v4.16.0-rc1)
   Tag v4.16.0-rc1 exists

Repository: lean4checker
   On compatible toolchain (>= v4.16.0-rc1)
   Tag v4.16.0-rc1 exists

Repository: doc-gen4
   On compatible toolchain (>= v4.16.0-rc1)
   Tag v4.16.0-rc1 exists

Repository: Verso
   Not on target toolchain (needs ≥ v4.16.0-rc1, but main is on leanprover/lean4:v4.15.0)

Repository: ProofWidgets4
   On compatible toolchain (>= v4.16.0-rc1)

Repository: Aesop
   On compatible toolchain (>= v4.16.0-rc1)
   Tag v4.16.0-rc1 exists

Repository: import-graph
   On compatible toolchain (>= v4.16.0-rc1)
   Tag v4.16.0-rc1 exists

Repository: plausible
   On compatible toolchain (>= v4.16.0-rc1)
   Tag v4.16.0-rc1 exists

Repository: Mathlib
   On compatible toolchain (>= v4.16.0-rc1)
   Tag v4.16.0-rc1 exists

Repository: REPL
   Not on target toolchain (needs ≥ v4.16.0-rc1, but master is on leanprover/lean4:v4.14.0)
```
2025-01-06 06:37:01 +00:00
Leonardo de Moura
2ed77f3b26 feat: attribute [grind] (#6545)
This PR introduces the parametric attribute `[grind]` for annotating
theorems and definitions. It also replaces `[grind_eq]` with `[grind
=]`. For definitions, `[grind]` is equivalent to `[grind =]`.

The new attribute supports the following variants:

- **`[grind =]`**: Uses the left-hand side of the theorem's conclusion
as the pattern for E-matching.
- **`[grind =_]`**: Uses the right-hand side of the theorem's conclusion
as the pattern for E-matching.
- **`[grind _=_]`**: Creates two patterns. One for the left-hand side
and one for the right-hand side.
- **`[grind →]`**: Searches for (multi-)patterns in the theorem's
antecedents, stopping once a usable multi-pattern is found.
- **`[grind ←]`**: Searches for (multi-)patterns in the theorem's
conclusion, stopping once a usable multi-pattern is found.
- **`[grind]`**: Searches for (multi-)patterns in both the theorem's
conclusion and antecedents. It starts with the conclusion and stops once
a usable multi-pattern is found.

The `grind_pattern` command remains available for cases where these
attributes do not yield the desired result.
2025-01-06 03:05:20 +00:00
Sebastian Ullrich
76f883b999 fix: remove unused -static-libgcc MinGW linker arg (#6535)
This PR avoids a linker warning on Windows.

The argument may have been superfluous ever since the initial
implementation.
2025-01-05 21:57:56 +00:00
Leonardo de Moura
675244de76 feat: [grind_eq] attribute for the grind tactic (#6539)
This PR introduces the `[grind_eq]` attribute, designed to annotate
equational theorems and functions for heuristic instantiations in the
`grind` tactic.
When applied to an equational theorem, the `[grind_eq]` attribute
instructs the `grind` tactic to automatically use the annotated theorem
to instantiate patterns during proof search. If applied to a function,
it marks all equational theorems associated with that function.

```lean
@[grind_eq]
theorem foo_idempotent : foo (foo x) = foo x := ...

@[grind_eq] def f (a : Nat) :=
  match a with
  | 0 => 10
  | x+1 => g (f x)
```

In the example above, the `grind` tactic will add instances of the
theorem `foo_idempotent` to the local context whenever it encounters the
pattern `foo (foo x)`. Similarly, functions annotated with `[grind_eq]`
will propagate this annotation to their associated equational theorems.
2025-01-05 21:38:05 +00:00
Leonardo de Moura
fd091d1dfe feat: pattern normalization in the grind tactic (#6538)
This PR ensures patterns provided by users are normalized. See new test
to understand why this is needed.
2025-01-05 19:35:52 +00:00
Leonardo de Moura
7b29f488df fix: E-matching thresholds in the grind tactic (#6536)
This PR fixes different thresholds for controlling E-matching in the
`grind` tactic.
2025-01-05 17:34:56 +00:00
Leonardo de Moura
fb506b957c fix: allow projections in E-matching patterns (#6534)
This PR ensures that users can utilize projections in E-matching
patterns within the `grind` tactic.
2025-01-05 16:20:54 +00:00
Leonardo de Moura
dc5c8097b5 feat: add term offset support to the grind E-matching modulo (#6533)
This PR adds support to E-matching offset patterns. For example, we want
to be able to E-match the pattern `f (#0 + 1)` with term `f (a + 2)`.
2025-01-05 02:20:17 +00:00
Kim Morrison
9dcbc330fd chore: fix signature of perm_insertIdx (#6532) 2025-01-04 23:43:23 +00:00
Leonardo de Moura
d22233fc7b fix: let_fun support in grind (#6531)
This PR fixes the support for `let_fun` in `grind`.
2025-01-04 22:22:55 +00:00
Leonardo de Moura
a5b1ed906c fix: nondeterministic failure in grind (#6530)
This PR fixes nondeterministic failures in the (WIP) `grind` tactic.
2025-01-04 21:40:32 +00:00
Leonardo de Moura
ad2c16dade feat: add support for let-declarations to grind (#6529)
This PR adds support for `let`-declarations to the (WIP) `grind` tactic.
2025-01-04 19:24:56 +00:00
Leonardo de Moura
37127ead07 fix: missing propagation in grind (#6528)
This PR adds a missing propagation rule to the (WIP) `grind` tactic.
2025-01-04 18:45:55 +00:00
Kitamado
31435e9cd1 doc: fix broken code blocks in RELEASES.md (#6527)
just fix markdown
2025-01-04 08:08:12 +00:00
Kim Morrison
639e6e92a4 chore: cleanup imports in Lean.Lsp (#6523)
This PR splits a definition out of `Lean.Lsp.Basic`, with the effect
that material about JSON is not needed for `Lean.Meta.Sorry` and its
dependencies.
2025-01-04 04:33:24 +00:00
Kim Morrison
9080df3110 chore: import cleanup in Init (#6522)
This PR avoids unnecessarily importing "kitchen sink" files.
2025-01-04 04:13:13 +00:00
Kim Morrison
cdeb958afd chore: add plausible to release checklist (#6525) 2025-01-04 04:08:21 +00:00
Kim Morrison
d2189542b5 chore: upstream some List.Perm lemmas (#6524)
This PR upstreams some remaining `List.Perm` lemmas from Batteries.
2025-01-04 04:04:13 +00:00
Leonardo de Moura
ad593b36d9 feat: add support for match-expressions to grind (#6521)
This PR adds support for activating relevant `match`-equations as
E-matching theorems. It uses the `match`-equation lhs as the pattern.
2025-01-04 02:18:43 +00:00
Kim Morrison
28a7098728 feat: add script for generating release notes (#6519)
This PR adds a script to automatically generate release notes using the
new `changelog-*` labels and "This PR ..." conventions.

Usage:
```
script/release_notes.py v4.X.0
```
where `v4.X.0` is the **previous** release, i.e. the script will process
all commits *since* that tag.
2025-01-04 01:31:02 +00:00
Kim Morrison
d991feddad chore: cherry-pick release notes from releases/v4.15.0 and releases/v4.16.0 (#6520)
These release notes were automatically generated by the script in #6519.
2025-01-04 01:25:33 +00:00
Henrik Böving
58d178e68f fix: cond reflection bug in bv_decide (#6517)
This PR fixes a slight bug that was created in the reflection of `bif`
in `bv_decide`.

Tagged as changelog-no as the code in question isn't in an RC yet.
2025-01-03 19:35:58 +00:00
Leonardo de Moura
7b496bf44b feat: improve cases tactic used in grind (#6516)
This PR enhances the `cases` tactic used in the `grind` tactic and
ensures that it can be applied to arbitrary expressions.
2025-01-03 15:54:55 +00:00
Mitchell Lee
10b2f6b27e feat: bdiv and bmod lemmas (#6494)
This PR proves the basic theorems about the functions `Int.bdiv` and
`Int.bmod`.

For all integers `x` and all natural numbers `m`, we have:
- `Int.bdiv_add_bmod`: `m * bdiv x m + bmod x m = x` (which is stated in
the docstring for docs#Int.bdiv)
- `Int.bmod_add_bdiv`: `bmod x m + m * bdiv x m = x`
- `Int.bdiv_add_bmod'`: `bdiv x m * m + bmod x m = x`
- `Int.bmod_add_bdiv'`: `bmod x m + bdiv x m * m = x`
- `Int.bmod_eq_self_sub_mul_bdiv`: `bmod x m = x - m * bdiv x m`
- `Int.bmod_eq_self_sub_bdiv_mul`: `bmod x m = x - bdiv x m * m`

These theorems are all equivalent to each other by the basic properties
of addition, multiplication, and subtraction of integers.

The names `Int.bdiv_add_bmod`, `Int.bmod_add_bdiv`,
`Int.bdiv_add_bmod'`, and `Int.bmod_add_bdiv'` are meant to parallel the
names of the existing theorems docs#Int.tmod_add_tdiv,
docs#Int.tdiv_add_tmod, docs#Int.tmod_add_tdiv', and
docs#Int.tdiv_add_tmod'.

The names `Int.bmod_eq_self_sub_mul_bdiv` and
`Int.bmod_eq_self_sub_bdiv_mul` follow mathlib's naming conventions.
Note that there is already a theorem called docs#Int.bmod_def, so it
would not have been possible to parallel the name of the existing
theorem docs#Int.tmod_def.

See
https://leanprover.zulipchat.com/#narrow/channel/217875-Is-there-code-for-X.3F/topic/bdiv.20and.20bmod.

Closes #6493.
2025-01-03 09:37:02 +00:00
Leonardo de Moura
19078655bc perf: avoid unnecessary assert/intro pairs in grind (#6514)
This PR enhances the assertion of new facts in `grind` by avoiding the
creation of unnecessary metavariables.
2025-01-03 02:16:08 +00:00
Leonardo de Moura
df9ed20385 feat: ite and dite support in grind (#6513)
This PR adds support for (dependent) if-then-else terms (i.e., `ite` and
`dite` applications) in the `grind` tactic.
2025-01-03 01:05:44 +00:00
Leonardo de Moura
3e2f1faebf feat: add user-defined fallback procedure for the grind tactic (#6512)
This PR introduces support for user-defined fallback code in the `grind`
tactic. The fallback code can be utilized to inspect the state of
failing `grind` subgoals and/or invoke user-defined automation. Users
can now write `grind on_failure <code>`, where `<code>` should have the
type `GoalM Unit`. See the modified tests in this PR for examples.
2025-01-02 23:56:51 +00:00
Leonardo de Moura
9d622270a1 feat: custom congruence rule for equality in grind (#6510)
This PR adds a custom congruence rule for equality in `grind`. The new
rule takes into account that `Eq` is a symmetric relation. In the
future, we will add support for arbitrary symmetric relations. The
current rule is important for propagating disequalities effectively in
`grind`.
2025-01-02 22:08:19 +00:00
Leonardo de Moura
e46b5f39bf fix: congruence closure in the grind tactic (#6509)
This PR fixes a bug in the congruence closure data structure used in the
`grind` tactic. The new test includes an example that previously caused
a panic. A similar panic was also occurring in the test
`grind_nested_proofs.lean`.
2025-01-02 19:08:02 +00:00
Leonardo de Moura
3cba17140f fix: missing case in checkParents (#6508)
This PR fixes a bug in the sanity checkers for the `grind` tactic. See
the new test for an example of a case where it was panicking.
2025-01-02 16:06:52 +00:00
Lean stage0 autoupdater
092449adb8 chore: update stage0 2025-01-02 12:29:37 +00:00
Joachim Breitner
e9f069146c feat: partial_fixpoint: monotonicity tactic (#6506)
This PR adds the `monotonicity` tactic, intended to be used inside the
`partial_fixpoint` feature.

Part of #6355.
2025-01-02 11:07:05 +00:00
Joachim Breitner
7d0c0d4d92 feat: partial_fixpoint: theory (#6477)
This PR adds the necessary domain theory that backs the
`partial_fixpoint` feature.

Part of #6355.
2025-01-02 09:39:18 +00:00
Eric Wieser
9eb173e444 fix: ignore no_index around OfNat.ofNat in norm_cast (#6438)
This PR ensures `norm_cast` doesn't fail to act in the presence of
`no_index` annotations

While leanprover/lean4#2867 exists, it is necessary to put `no_index`
around `OfNat.ofNat` in simp lemmas.
This results in extra `Expr.mdata` nodes, which must be removed before
checking for `ofNat` numerals.
2025-01-02 09:10:27 +00:00
Leonardo de Moura
8d9d81453b feat: grind simple strategy (#6503)
This PR adds a simple strategy to the (WIP) `grind` tactic. It just
keeps internalizing new theorem instances found by E-matching. The
simple strategy can solve examples such as:

```lean
grind_pattern Array.size_set => Array.set a i v h
grind_pattern Array.get_set_eq  => a.set i v h
grind_pattern Array.get_set_ne => (a.set i v hi)[j]

example (as bs : Array α) (v : α)
        (i : Nat)
        (h₁ : i < as.size)
        (h₂ : bs = as.set i v)
        : as.size = bs.size := by
  grind

example (as bs cs : Array α) (v : α)
        (i : Nat)
        (h₁ : i < as.size)
        (h₂ : bs = as.set i v)
        (h₃ : cs = bs)
        (h₄ : i ≠ j)
        (h₅ : j < cs.size)
        (h₆ : j < as.size)
        : cs[j] = as[j] := by
  grind


opaque R : Nat → Nat → Prop
theorem Rtrans (a b c : Nat) : R a b → R b c → R a c := sorry

grind_pattern Rtrans => R a b, R b c

example : R a b → R b c → R c d → R d e → R a d := by
  grind
```
2025-01-02 05:36:01 +00:00
Leonardo de Moura
a08379ce2e fix: proof generation for grind tactic (#6502)
This PR fixes a bug in the proof assembly procedure utilized by the
`grind` tactic.
2025-01-02 04:42:31 +00:00
Leonardo de Moura
f0c59364f4 feat: add PersistentHashSet.toList (#6501)
This PR adds `PersistentHashSet.toList`
2025-01-02 03:52:21 +00:00
Leonardo de Moura
c0d67e2a65 fix: bug in markNestedProofs within grind (#6500)
This PR fixes a bug in the `markNestedProofs` used in `grind`. See new
test.
2025-01-02 02:14:45 +00:00
Leonardo de Moura
a8d09dad1b fix: proof canonicalizer in grind (#6499)
This PR fixes the proof canonicalizer for `grind`.
2025-01-02 01:26:39 +00:00
Leonardo de Moura
f7c4edc2b7 feat: dependent forall propagator in grind (#6498)
This PR adds support in the `grind` tactic for propagating dependent
forall terms `forall (h : p), q[h]` where `p` is a proposition.
2025-01-02 00:08:36 +00:00
Leonardo de Moura
82bae24e59 fix: another bug in theorem instantiation in grind (#6497)
This PR fixes another theorem instantiation bug in the `grind` tactic.
It also moves new instances to be processed to `Goal`.
2025-01-01 23:09:21 +00:00
Leonardo de Moura
fedaf850bb fix: theorem instantiation in grind (#6492)
This PR fixes a bug in the theorem instantiation procedure in the (WIP)
`grind` tactic. For example, it was missing the following instance in
one of the tests:

```lean
[grind.ematch.instance] Array.get_set_ne: ∀ (hj : i < bs.size), j ≠ i → (bs.set j w ⋯)[i] = bs[i]
```

This PR also renames the `grind` base monad to `GrindCoreM`.
2025-01-01 17:56:27 +00:00
Kyle Miller
6d447156c4 fix: make sure parent structure projections have 'go to definition' information (#6487)
This PR adds source position information for `structure` parent
projections, supporting "go to definition". Closes #3063.
2024-12-31 21:59:38 +00:00
Leonardo de Moura
3427630a14 feat: configuration options for the grind tactic (#6490)
This PR adds basic configuration options for the `grind` tactic.
2024-12-31 21:09:41 +00:00
Leonardo de Moura
5ba476116f fix: E-matching module for grind (#6488)
This PR fixes and refactors the E-matching module for the (WIP) `grind`
tactic.

Next step: top-level search procedure for `grind`.
2024-12-31 20:07:27 +00:00
Leonardo de Moura
8899c7ed8c feat: instantiate ematch theorems in grind (#6485)
This PR implements `Grind.EMatch.instantiateTheorem` in the (WIP)
`grind` tactic.
2024-12-31 19:25:12 +00:00
Kyle Miller
640b356a04 chore: add missing diff-exposing in type/value mismatch errors (#6484)
This PR addresses a few error messages where diffs weren't being
exposed.
2024-12-31 17:47:12 +00:00
Kyle Miller
8f5ce3a356 feat: upstream ToExpr deriving handler from Mathlib (#6473)
This PR adds a deriving handler for the `ToExpr` class. It can handle
mutual and nested inductive types, however it falls back to creating
`partial` instances in such cases. This is upstreamed from the Mathlib
deriving handler written by @kmill, but has fixes to handle autoimplicit
universe level variables.

This is a followup to #6285 (adding the `ToLevel` class). This PR
supersedes #5906.

Co-authored-by: Alex Keizer <alex@keizer.dev>

---------

Co-authored-by: Alex Keizer <alex@keizer.dev>
2024-12-31 15:11:53 +00:00
Leonardo de Moura
2c87905d77 feat: E-matching procedure for the grind tactic (#6481)
This PR implements E-matching for the (WIP) `grind` tactic. We still
need to finalize and internalize the new instances.
2024-12-31 02:46:09 +00:00
Cameron Zwarich
32dc16590b chore: reduce churn in tests/lean/run/meta5.lean (#6480) 2024-12-31 02:32:56 +00:00
Leonardo de Moura
7e8e22e2bd fix: internalize nested ground patterns when activating ematch theorems (#6478)
This PR internalize nested ground patterns when activating ematch
theorems in the (WIP) `grind` tactic.
2024-12-30 16:13:47 +00:00
Leonardo de Moura
9b28c5879a feat: ematch theorem activation for grind (#6475)
This PR adds support for activating relevant theorems for the (WIP)
`grind` tactic. We say a theorem is relevant to a `grind` goal if the
symbols occurring in its patterns also occur in the goal.
2024-12-30 09:57:33 +00:00
Leonardo de Moura
24a8561ec4 feat: check pattern coverage in the grind_pattern command (#6474)
This PR adds pattern validation to the `grind_pattern` command. The new
`checkCoverage` function will also be used to implement the attributes
`@[grind_eq]`, `@[grind_fwd]`, and `@[grind_bwd]`.
2024-12-30 03:40:43 +00:00
Lean stage0 autoupdater
3c326d771c chore: update stage0 2024-12-29 21:10:53 +00:00
Leonardo de Moura
7433e74fc4 feat: theorem patterns for heuristic instantiation in grind (#6472)
This PR implements the command `grind_pattern`. The new command allows
users to associate patterns with theorems. These patterns are used for
performing heuristic instantiation with e-matching. In the future, we
will add the attributes `@[grind_eq]`, `@[grind_fwd]`, and
`@[grind_bwd]` to compute the patterns automatically for theorems.
2024-12-29 20:26:58 +00:00
Eric Wieser
11eea84fd5 doc: explain app_delab (#6450)
This PR adds a docstring to the `@[app_delab]` attribute.

---------

Co-authored-by: Kyle Miller <kmill31415@gmail.com>
2024-12-29 15:06:55 +00:00
Lean stage0 autoupdater
536c6a8ea6 chore: update stage0 2024-12-29 01:19:00 +00:00
Leonardo de Moura
9c0ef2a282 feat: add grind_pattern command (#6470)
This PR introduces a command for specifying patterns used in the
heuristic instantiation of global theorems in the `grind` tactic. Note
that this PR only adds the parser.
2024-12-29 00:44:40 +00:00
Leonardo de Moura
a781f9858c feat: missing data for grind e-match (#6469)
This PR adds support code for implementing e-match in the (WIP) `grind`
tactic.
2024-12-29 00:33:25 +00:00
Leonardo de Moura
5930db946c fix: Int.reduceNeg simproc (#6468)
This PR fixes issue #6467 


closes #6467
2024-12-28 22:58:09 +00:00
Leonardo de Moura
3fc74854d7 fix: check function types when detecting congruences in grind (#6466)
This PR completes the implementation of `addCongrTable` in the (WIP)
`grind` tactic. It also adds a new test to demonstrate why the extra
check is needed. It also updates the field `cgRoot` (congruence root).
2024-12-28 19:53:02 +00:00
Leonardo de Moura
fe45ddd610 feat: projections in grind (#6465)
This PR adds support for projection functions to the (WIP) `grind`
tactic.
2024-12-27 23:50:58 +00:00
Leonardo de Moura
f545df9922 feat: literal values in grind (#6464)
This PR completes support for literal values in the (WIP) `grind`
tactic. `grind` now closes the goal whenever it merges two equivalence
classes with distinct literal values.
2024-12-27 22:18:56 +00:00
Leonardo de Moura
844e82e176 feat: constructors in grind (#6463)
This PR adds support for constructors to the (WIP) `grind` tactic. When
merging equivalence classes, `grind` checks for equalities between
constructors. If they are distinct, it closes the goal; if they are the
same, it applies injectivity.
2024-12-27 21:15:02 +00:00
Leonardo de Moura
2d7d3388e2 fix: missing Not propagation rule in grind (#6461)
This PR adds a new propagation rule for negation to the (WIP) `grind`
tactic.
2024-12-27 17:37:32 +00:00
Henrik Böving
c14e5ae7de chore: implement reduceCond for bv_decide (#6460)
This PR implements the equivalent of `reduceIte` for `cond` in
`bv_decide` as we switched to `bif` for the `if` normal form.
2024-12-27 10:12:52 +00:00
Leonardo de Moura
6a839796fd feat: add grind tactic (#6459)
This PR adds the (WIP) `grind` tactic. It currently generates a warning
message to make it clear that the tactic is not ready for production.
2024-12-27 03:48:01 +00:00
Leonardo de Moura
e76dc20200 feat: use compact congruence proofs in grind if applicable (#6458)
This PR adds support for compact congruence proofs in the (WIP) `grind`
tactic. The `mkCongrProof` function now verifies whether the congruence
proof can be constructed using only `congr`, `congrFun`, and `congrArg`,
avoiding the need to generate the more complex `hcongr` auxiliary
theorems.
2024-12-26 23:58:04 +00:00
Leonardo de Moura
dca874ea57 feat: congruence proofs for grind (#6457)
This PR adds support for generating congruence proofs for congruences
detected by the `grind` tactic.
2024-12-26 22:20:36 +00:00
Leonardo de Moura
c282d558fa fix: fix: bug in mkEqProof within grind (#6456)
This PR fixes another bug in the equality proof generator in the (WIP)
`grind` tactic.
2024-12-26 19:03:35 +00:00
Leonardo de Moura
57050be3ab fix: bug in mkEqProof within grind (#6455)
This PR fixes a bug in the equality proof generator in the (WIP) `grind`
tactic.
2024-12-26 18:25:11 +00:00
Henrik Böving
37b53b70d0 perf: improve bv_decide performance with large literals (#6453)
This PR improves bv_decide's performance in the presence of large
literals.

The core change of this PR is the reformulation of the reflection code
for literals to:
```diff
 def eval (assign : Assignment) : BVExpr w → BitVec w
   | .var idx =>
-    let ⟨bv⟩ := assign.get idx
-    bv.truncate w
+    let packedBv := assign.get idx
+    /-
+    This formulation improves performance, as in a well formed expression the condition always holds
+    so there is no need for the more involved `BitVec.truncate` logic.
+    -/
+    if h : packedBv.w = w then
+      h ▸ packedBv.bv
+    else
+      packedBv.bv.truncate w
```
The remainder is merely further simplifications that make the terms
smaller and easier to deal with in general. This change is motivated by
applying the following diff to the kernel:
```diff
diff --git a/src/kernel/type_checker.cpp b/src/kernel/type_checker.cpp
index b0e6844dca..f13bb96bd4 100644
--- a/src/kernel/type_checker.cpp
+++ b/src/kernel/type_checker.cpp
@@ -518,6 +518,7 @@ optional<constant_info> type_checker::is_delta(expr const & e) const {
 optional<expr> type_checker::unfold_definition_core(expr const & e) {
     if (is_constant(e)) {
         if (auto d = is_delta(e)) {
+//            std::cout << "Working on unfolding: " << d->get_name() << std::endl;
             if (length(const_levels(e)) == d->get_num_lparams()) {
                 if (m_diag) {
                     m_diag->record_unfold(d->get_name());
```
and observing that in the test case from #6043 we see a long series of
```
Working on unfolding: Bool.decEq
Working on unfolding: Bool.decEq.match_1
Working on unfolding: Bool.casesOn
Working on unfolding: Nat.ble
Working on unfolding: Nat.brecOn
Working on unfolding: Nat.beq.match_1
Working on unfolding: Nat.casesOn
Working on unfolding: Nat.casesOn
Working on unfolding: Nat.beq.match_1
Working on unfolding: Nat.casesOn
Working on unfolding: Nat.casesOn
```
the chain begins with `BitVec.truncate`, works through a few
abstractions and then continues like above forever, so I avoid the call
to truncate like this. It is not quite clear to me why removing `ofBool`
helps so much here, maybe some other kernel heuristic kicks in to rescue
us.

Either way this diff is a general improvement for reflection of `BitVec`
constants as we should never have to run `BitVec.truncate` again!

Fixes: #6043
2024-12-26 16:50:00 +00:00
Leonardo de Moura
8a1e50f0b9 feat: equality proof generation for grind (#6452)
This PR adds support for generating (small) proofs for any two
expressions that belong to the same equivalence class in the `grind`
tactic state.
2024-12-26 06:01:45 +00:00
Leonardo de Moura
bdcb7914b5 chore: check whether pointer equality implies structural equality in grind (#6451)
This PR checks whether in the internal state of the `grind` tactic
pointer equality implies structural equality.
2024-12-26 03:50:39 +00:00
Leonardo de Moura
0ebe9e5ba3 feat: support for builtin grind propagators (part 2) (#6449)
This PR completes the implementation of the command
`builtin_grind_propagator`.
2024-12-25 23:54:55 +00:00
Lean stage0 autoupdater
65e8ba0574 chore: update stage0 2024-12-25 23:27:32 +00:00
Leonardo de Moura
3cddae6492 feat: support for builtin grind propagators (#6448)
This PR declares the command `builtin_grind_propagator` for registering
equation propagator for `grind`. It also declares the auxiliary the
attribute.
2024-12-25 22:55:39 +00:00
Leonardo de Moura
977b8e001f refactor: move simplifier support to GrindM (#6447)
This PR refactors `grind` and adds support for invoking the simplifier
using the `GrindM` monad.
2024-12-25 21:01:32 +00:00
Leonardo de Moura
f9f8abe2a3 feat: propagate equality in grind (#6443)
This PR adds support for propagating the truth value of equalities in
the (WIP) `grind` tactic.
2024-12-24 23:54:36 +00:00
Leonardo de Moura
ec80de231e fix: checkParents in grind (#6442)
This PR fixes the `checkParents` sanity check in `grind`.
2024-12-24 22:45:18 +00:00
Leonardo de Moura
630577a9ea feat: truth value propagation for grind (#6441)
This PR adds basic truth value propagation rules to the (WIP) `grind`
tactic.
2024-12-24 21:12:53 +00:00
Leonardo de Moura
cde35bcc0d test: grind (#6440)
This PR adds additional tests for `grind` and fixed minor issues.
2024-12-24 04:33:05 +00:00
Leonardo de Moura
b18f3a3877 feat: detect congruent terms in grind (#6437)
This PR adds support for detecting congruent terms in the (WIP) `grind`
tactic. It also introduces the `grind.debug` option, which, when set to
`true`, checks many invariants after each equivalence class is merged.
This option is intended solely for debugging purposes.
2024-12-24 00:11:36 +00:00
Leonardo de Moura
5240405cf4 feat: congruence table for grind tactic (#6435)
This PR implements the congruence table for the (WIP) `grind` tactic. It
also fixes several bugs, and adds a new preprocessing step.
2024-12-23 02:31:42 +00:00
Leonardo de Moura
eb6c52e7e2 feat: canonicalizer for the grind tactic (#6433)
This PR adds a custom type and instance canonicalizer for the (WIP)
`grind` tactic. The `grind` tactic uses congruence closure but
disregards types, type formers, instances, and proofs. Proofs are
ignored due to proof irrelevance. Types, type formers, and instances are
considered supporting elements and are not factored into congruence
detection. Instead, `grind` only checks whether elements are
structurally equal, which, in the context of the `grind` tactic, is
equivalent to pointer equality. See new tests for examples where the
canonicalizer is important.
2024-12-21 22:32:25 +00:00
Kyle Miller
71942631d7 feat: explanations for cases applied to non-inductive types (#6378)
This PR adds an explanation to the error message when `cases` and
`induction` are applied to a term whose type is not an inductive type.
For `Prop`, these tactics now suggest the `by_cases` tactic. Example:
```
tactic 'cases' failed, major premise type is not an inductive type
  Prop

Explanation: the 'cases' tactic is for constructor-based reasoning as well as for applying
custom cases principles with a 'using' clause or a registered '@[cases_eliminator]' theorem.
The above type neither is an inductive type nor has a registered theorem.

Consider using the 'by_cases' tactic, which does true/false reasoning for propositions.
```

[Zulip
discussion](https://leanprover.zulipchat.com/#narrow/channel/270676-lean4/topic/Improving.20the.20error.20for.20.60cases.20p.60.20when.20.60p.60.20is.20a.20proposition/near/488882682)
2024-12-21 21:38:30 +00:00
Leonardo de Moura
16bc6ebcb6 fix: ensure simp and dsimp do not unfold too much (#6397)
This PR ensures that `simp` and `dsimp` do not unfold definitions that
are not intended to be unfolded by the user. See issue #5755 for an
example affected by this issue.

Closes #5755

---------

Co-authored-by: Kim Morrison <kim@tqft.net>
2024-12-21 04:16:15 +00:00
Leonardo de Moura
9e30ac3265 feat: add Expr.fvarsSubset (#6430)
This PR adds the predicate `Expr.fvarsSet a b`, which returns `true` if
and only if the free variables in `a` are a subset of the free variables
in `b`.
2024-12-20 22:29:47 +00:00
Cameron Zwarich
bf1d253764 feat: add support for extern LCNF decls (#6429)
This PR adds support for extern LCNF decls, which is required for parity
with the existing code generator.
2024-12-20 21:20:56 +00:00
Leonardo de Moura
052f3f54c8 fix: normalize universe levels in grind preprocessor (#6428)
This PR adds a new preprocessing step to the `grind` tactic:
universe-level normalization. The goal is to avoid missing equalities in
the congruence closure module.
2024-12-20 20:41:59 +00:00
Kim Morrison
39eaa214d4 chore: protect some lemmas in List/Array/Vector namespace (#6425) 2024-12-20 11:23:56 +00:00
327 changed files with 9368 additions and 1353 deletions

File diff suppressed because it is too large Load Diff

View File

@@ -76,6 +76,10 @@ We'll use `v4.6.0` as the intended release version as a running example.
- Toolchain bump PR including updated Lake manifest
- Create and push the tag
- There is no `stable` branch; skip this step
- [plausible](https://github.com/leanprover-community/plausible)
- Toolchain bump PR including updated Lake manifest
- Create and push the tag
- There is no `stable` branch; skip this step
- [Mathlib](https://github.com/leanprover-community/mathlib4)
- Dependencies: `Aesop`, `ProofWidgets4`, `lean4checker`, `Batteries`, `doc-gen4`, `import-graph`
- Toolchain bump PR notes:
@@ -93,6 +97,7 @@ We'll use `v4.6.0` as the intended release version as a running example.
- Toolchain bump PR including updated Lake manifest
- Create and push the tag
- Merge the tag into `stable`
- Run `scripts/release_checklist.py v4.6.0` to check that everything is in order.
- The `v4.6.0` section of `RELEASES.md` is out of sync between
`releases/v4.6.0` and `master`. This should be reconciled:
- Replace the `v4.6.0` section on `master` with the `v4.6.0` section on `releases/v4.6.0`

View File

@@ -1,16 +0,0 @@
We replace the inductive predicate `List.lt` with an upstreamed version of `List.Lex` from Mathlib.
(Previously `Lex.lt` was defined in terms of `<`; now it is generalized to take an arbitrary relation.)
This subtely changes the notion of ordering on `List α`.
`List.lt` was a weaker relation: in particular if `l₁ < l₂`, then
`a :: l₁ < b :: l₂` may hold according to `List.lt` even if `a` and `b` are merely incomparable
(either neither `a < b` nor `b < a`), whereas according to `List.Lex` this would require `a = b`.
When `<` is total, in the sense that `¬ · < ·` is antisymmetric, then the two relations coincide.
Mathlib was already overriding the order instances for `List α`,
so this change should not be noticed by anyone already using Mathlib.
We simultaneously add the boolean valued `List.lex` function, parameterised by a `BEq` typeclass
and an arbitrary `lt` function. This will support the flexibility previously provided for `List.lt`,
via a `==` function which is weaker than strict equality.

View File

@@ -43,7 +43,7 @@ echo -n " -DCMAKE_C_COMPILER=$PWD/stage1/bin/clang.exe -DCMAKE_C_COMPILER_WORKS=
echo -n " -DSTAGE0_CMAKE_C_COMPILER=clang -DSTAGE0_CMAKE_CXX_COMPILER=clang++"
echo -n " -DLEAN_EXTRA_CXX_FLAGS='--sysroot $PWD/llvm -idirafter /clang64/include/'"
echo -n " -DLEANC_INTERNAL_FLAGS='--sysroot ROOT -nostdinc -isystem ROOT/include/clang' -DLEANC_CC=ROOT/bin/clang.exe"
echo -n " -DLEANC_INTERNAL_LINKER_FLAGS='-L ROOT/lib -static-libgcc -Wl,-Bstatic -lgmp $(pkg-config --static --libs libuv) -lunwind -Wl,-Bdynamic -fuse-ld=lld'"
echo -n " -DLEANC_INTERNAL_LINKER_FLAGS='-L ROOT/lib -Wl,-Bstatic -lgmp $(pkg-config --static --libs libuv) -lunwind -Wl,-Bdynamic -fuse-ld=lld'"
# when not using the above flags, link GMP dynamically/as usual. Always link ICU dynamically.
echo -n " -DLEAN_EXTRA_LINKER_FLAGS='-lgmp $(pkg-config --libs libuv) -lucrtbase'"
# do not set `LEAN_CC` for tests

132
script/release_checklist.py Executable file
View File

@@ -0,0 +1,132 @@
#!/usr/bin/env python3
import argparse
import yaml
import requests
import base64
import subprocess
import sys
import os
def parse_repos_config(file_path):
with open(file_path, "r") as f:
return yaml.safe_load(f)["repositories"]
def get_github_token():
try:
import subprocess
result = subprocess.run(['gh', 'auth', 'token'], capture_output=True, text=True)
if result.returncode == 0:
return result.stdout.strip()
except FileNotFoundError:
print("Warning: 'gh' CLI not found. Some API calls may be rate-limited.")
return None
def get_branch_content(repo_url, branch, file_path, github_token):
api_url = repo_url.replace("https://github.com/", "https://api.github.com/repos/") + f"/contents/{file_path}?ref={branch}"
headers = {'Authorization': f'token {github_token}'} if github_token else {}
response = requests.get(api_url, headers=headers)
if response.status_code == 200:
content = response.json().get("content", "")
content = content.replace("\n", "")
try:
return base64.b64decode(content).decode('utf-8').strip()
except Exception:
return None
return None
def tag_exists(repo_url, tag_name, github_token):
api_url = repo_url.replace("https://github.com/", "https://api.github.com/repos/") + f"/git/refs/tags/{tag_name}"
headers = {'Authorization': f'token {github_token}'} if github_token else {}
response = requests.get(api_url, headers=headers)
return response.status_code == 200
def is_merged_into_stable(repo_url, tag_name, stable_branch, github_token):
# First get the commit SHA for the tag
api_base = repo_url.replace("https://github.com/", "https://api.github.com/repos/")
headers = {'Authorization': f'token {github_token}'} if github_token else {}
# Get tag's commit SHA
tag_response = requests.get(f"{api_base}/git/refs/tags/{tag_name}", headers=headers)
if tag_response.status_code != 200:
return False
tag_sha = tag_response.json()['object']['sha']
# Get commits on stable branch containing this SHA
commits_response = requests.get(
f"{api_base}/commits?sha={stable_branch}&per_page=100",
headers=headers
)
if commits_response.status_code != 200:
return False
# Check if any commit in stable's history matches our tag's SHA
stable_commits = [commit['sha'] for commit in commits_response.json()]
return tag_sha in stable_commits
def parse_version(version_str):
# Remove 'v' prefix and split into components
# Handle Lean toolchain format (leanprover/lean4:v4.x.y)
if ':' in version_str:
version_str = version_str.split(':')[1]
version = version_str.lstrip('v')
# Handle release candidates by removing -rc part for comparison
version = version.split('-')[0]
return tuple(map(int, version.split('.')))
def is_version_gte(version1, version2):
"""Check if version1 >= version2"""
return parse_version(version1) >= parse_version(version2)
def is_release_candidate(version):
return "-rc" in version
def main():
github_token = get_github_token()
if len(sys.argv) != 2:
print("Usage: python3 release_checklist.py <toolchain>")
sys.exit(1)
toolchain = sys.argv[1]
with open(os.path.join(os.path.dirname(__file__), "release_repos.yml")) as f:
repos = yaml.safe_load(f)["repositories"]
for repo in repos:
name = repo["name"]
url = repo["url"]
branch = repo["branch"]
check_stable = repo["stable-branch"]
check_tag = repo.get("toolchain-tag", True)
print(f"\nRepository: {name}")
# Check if branch is on at least the target toolchain
lean_toolchain_content = get_branch_content(url, branch, "lean-toolchain", github_token)
if lean_toolchain_content is None:
print(f" ❌ No lean-toolchain file found in {branch} branch")
continue
on_target_toolchain = is_version_gte(lean_toolchain_content.strip(), toolchain)
if not on_target_toolchain:
print(f" ❌ Not on target toolchain (needs ≥ {toolchain}, but {branch} is on {lean_toolchain_content.strip()})")
continue
print(f" ✅ On compatible toolchain (>= {toolchain})")
# Only check for tag if toolchain-tag is true
if check_tag:
if not tag_exists(url, toolchain, github_token):
print(f" ❌ Tag {toolchain} does not exist")
continue
print(f" ✅ Tag {toolchain} exists")
# Only check merging into stable if stable-branch is true and not a release candidate
if check_stable and not is_release_candidate(toolchain):
if not is_merged_into_stable(url, toolchain, "stable", github_token):
print(f" ❌ Tag {toolchain} is not merged into stable")
continue
print(f" ✅ Tag {toolchain} is merged into stable")
if __name__ == "__main__":
main()

79
script/release_repos.yml Normal file
View File

@@ -0,0 +1,79 @@
repositories:
- name: Batteries
url: https://github.com/leanprover-community/batteries
toolchain-tag: true
stable-branch: true
branch: main
dependencies: []
- name: lean4checker
url: https://github.com/leanprover/lean4checker
toolchain-tag: true
stable-branch: true
branch: master
dependencies: []
- name: doc-gen4
url: https://github.com/leanprover/doc-gen4
toolchain-tag: true
stable-branch: false
branch: main
dependencies: []
- name: Verso
url: https://github.com/leanprover/verso
toolchain-tag: true
stable-branch: false
branch: main
dependencies: []
- name: ProofWidgets4
url: https://github.com/leanprover-community/ProofWidgets4
toolchain-tag: false
stable-branch: false
branch: main
dependencies:
- Batteries
- name: Aesop
url: https://github.com/leanprover-community/aesop
toolchain-tag: true
stable-branch: true
branch: master
dependencies:
- Batteries
- name: import-graph
url: https://github.com/leanprover-community/import-graph
toolchain-tag: true
stable-branch: false
branch: main
dependencies: []
- name: plausible
url: https://github.com/leanprover-community/plausible
toolchain-tag: true
stable-branch: false
branch: main
dependencies: []
- name: Mathlib
url: https://github.com/leanprover-community/mathlib4
toolchain-tag: true
stable-branch: true
branch: master
dependencies:
- Aesop
- ProofWidgets4
- lean4checker
- Batteries
- doc-gen4
- import-graph
- name: REPL
url: https://github.com/leanprover-community/repl
toolchain-tag: true
stable-branch: true
branch: master
dependencies:
- Mathlib

View File

@@ -37,3 +37,4 @@ import Init.MacroTrace
import Init.Grind
import Init.While
import Init.Syntax
import Init.Internal

View File

@@ -36,6 +36,8 @@ namespace Array
@[simp] theorem mem_toList_iff (a : α) (l : Array α) : a l.toList a l := by
cases l <;> simp
@[simp] theorem length_toList {l : Array α} : l.toList.length = l.size := rfl
/-! ### empty -/
@[simp] theorem empty_eq {xs : Array α} : #[] = xs xs = #[] := by
@@ -954,13 +956,18 @@ theorem size_eq_of_beq [BEq α] {xs ys : Array α} (h : xs == ys) : xs.size = ys
rw [Bool.eq_iff_iff]
simp +contextual
private theorem beq_of_beq_singleton [BEq α] {a b : α} : #[a] == #[b] a == b := by
intro h
have : isEqv #[a] #[b] BEq.beq = true := h
simp [isEqv, isEqvAux] at this
assumption
@[simp] theorem reflBEq_iff [BEq α] : ReflBEq (Array α) ReflBEq α := by
constructor
· intro h
constructor
intro a
suffices (#[a] == #[a]) = true by
simpa only [instBEq, isEqv, isEqvAux, Bool.and_true]
apply beq_of_beq_singleton
simp
· intro h
constructor
@@ -973,11 +980,9 @@ theorem size_eq_of_beq [BEq α] {xs ys : Array α} (h : xs == ys) : xs.size = ys
· intro a b h
apply singleton_inj.1
apply eq_of_beq
simp only [instBEq, isEqv, isEqvAux]
simpa
simpa [instBEq, isEqv, isEqvAux]
· intro a
suffices (#[a] == #[a]) = true by
simpa only [instBEq, isEqv, isEqvAux, Bool.and_true]
apply beq_of_beq_singleton
simp
· intro h
constructor
@@ -998,52 +1003,7 @@ theorem size_eq_of_beq [BEq α] {xs ys : Array α} (h : xs == ys) : xs.size = ys
/-! Content below this point has not yet been aligned with `List`. -/
theorem singleton_eq_toArray_singleton (a : α) : #[a] = [a].toArray := rfl
@[simp] theorem singleton_def (v : α) : singleton v = #[v] := rfl
-- This is a duplicate of `List.toArray_toList`.
-- It's confusing to guess which namespace this theorem should live in,
-- so we provide both.
@[simp] theorem toArray_toList (a : Array α) : a.toList.toArray = a := rfl
@[simp] theorem length_toList {l : Array α} : l.toList.length = l.size := rfl
@[simp] theorem mkEmpty_eq (α n) : @mkEmpty α n = #[] := rfl
@[deprecated size_toArray (since := "2024-12-11")]
theorem size_mk (as : List α) : (Array.mk as).size = as.length := by simp [size]
theorem foldrM_push [Monad m] (f : α β m β) (init : β) (arr : Array α) (a : α) :
(arr.push a).foldrM f init = f a init >>= arr.foldrM f := by
simp only [foldrM_eq_reverse_foldlM_toList, push_toList, List.reverse_append, List.reverse_cons,
List.reverse_nil, List.nil_append, List.singleton_append, List.foldlM_cons, List.foldlM_reverse]
/--
Variant of `foldrM_push` with `h : start = arr.size + 1`
rather than `(arr.push a).size` as the argument.
-/
@[simp] theorem foldrM_push' [Monad m] (f : α β m β) (init : β) (arr : Array α) (a : α)
{start} (h : start = arr.size + 1) :
(arr.push a).foldrM f init start = f a init >>= arr.foldrM f := by
simp [ foldrM_push, h]
theorem foldr_push (f : α β β) (init : β) (arr : Array α) (a : α) :
(arr.push a).foldr f init = arr.foldr f (f a init) := foldrM_push ..
/--
Variant of `foldr_push` with the `h : start = arr.size + 1`
rather than `(arr.push a).size` as the argument.
-/
@[simp] theorem foldr_push' (f : α β β) (init : β) (arr : Array α) (a : α) {start}
(h : start = arr.size + 1) : (arr.push a).foldr f init start = arr.foldr f (f a init) :=
foldrM_push' _ _ _ _ h
/-- A more efficient version of `arr.toList.reverse`. -/
@[inline] def toListRev (arr : Array α) : List α := arr.foldl (fun l t => t :: l) []
@[simp] theorem toListRev_eq (arr : Array α) : arr.toListRev = arr.toList.reverse := by
rw [toListRev, foldl_toList, List.foldr_reverse, List.foldr_cons_nil]
/-! ### map -/
theorem mapM_eq_foldlM [Monad m] [LawfulMonad m] (f : α m β) (arr : Array α) :
arr.mapM f = arr.foldlM (fun bs a => bs.push <$> f a) #[] := by
@@ -1067,6 +1027,11 @@ where
induction l generalizing arr <;> simp [*]
simp [H]
@[simp] theorem _root_.List.map_toArray (f : α β) (l : List α) :
l.toArray.map f = (l.map f).toArray := by
apply ext'
simp
@[simp] theorem size_map (f : α β) (arr : Array α) : (arr.map f).size = arr.size := by
simp only [ length_toList]
simp
@@ -1076,6 +1041,128 @@ where
@[simp] theorem map_empty (f : α β) : map f #[] = #[] := mapM_empty f
@[simp] theorem getElem_map (f : α β) (a : Array α) (i : Nat) (hi : i < (a.map f).size) :
(a.map f)[i] = f (a[i]'(by simpa using hi)) := by
cases a
simp
@[simp] theorem getElem?_map (f : α β) (as : Array α) (i : Nat) :
(as.map f)[i]? = as[i]?.map f := by
simp [getElem?_def]
@[simp] theorem map_id_fun : map (id : α α) = id := by
funext l
induction l <;> simp_all
/-- `map_id_fun'` differs from `map_id_fun` by representing the identity function as a lambda, rather than `id`. -/
@[simp] theorem map_id_fun' : map (fun (a : α) => a) = id := map_id_fun
-- This is not a `@[simp]` lemma because `map_id_fun` will apply.
theorem map_id (l : Array α) : map (id : α α) l = l := by
cases l <;> simp_all
/-- `map_id'` differs from `map_id` by representing the identity function as a lambda, rather than `id`. -/
-- This is not a `@[simp]` lemma because `map_id_fun'` will apply.
theorem map_id' (l : Array α) : map (fun (a : α) => a) l = l := map_id l
/-- Variant of `map_id`, with a side condition that the function is pointwise the identity. -/
theorem map_id'' {f : α α} (h : x, f x = x) (l : Array α) : map f l = l := by
simp [show f = id from funext h]
theorem map_singleton (f : α β) (a : α) : map f #[a] = #[f a] := rfl
@[simp] theorem mem_map {f : α β} {l : Array α} : b l.map f a, a l f a = b := by
simp only [mem_def, toList_map, List.mem_map]
theorem exists_of_mem_map (h : b map f l) : a, a l f a = b := mem_map.1 h
theorem mem_map_of_mem (f : α β) (h : a l) : f a map f l := mem_map.2 _, h, rfl
theorem forall_mem_map {f : α β} {l : Array α} {P : β Prop} :
( (i) (_ : i l.map f), P i) (j) (_ : j l), P (f j) := by
simp
@[simp] theorem map_inj_left {f g : α β} : map f l = map g l a l, f a = g a := by
cases l <;> simp_all
theorem map_congr_left (h : a l, f a = g a) : map f l = map g l :=
map_inj_left.2 h
theorem map_inj : map f = map g f = g := by
constructor
· intro h; ext a; replace h := congrFun h #[a]; simpa using h
· intro h; subst h; rfl
@[simp] theorem map_eq_empty_iff {f : α β} {l : Array α} : map f l = #[] l = #[] := by
cases l
simp
theorem eq_empty_of_map_eq_empty {f : α β} {l : Array α} (h : map f l = #[]) : l = #[] :=
map_eq_empty_iff.mp h
@[simp] theorem map_map {f : α β} {g : β γ} {as : Array α} :
(as.map f).map g = as.map (g f) := by
cases as; simp
@[simp] theorem map_push {f : α β} {as : Array α} {x : α} :
(as.push x).map f = (as.map f).push (f x) := by
ext
· simp
· simp only [getElem_map, getElem_push, size_map]
split <;> rfl
theorem mapM_eq_mapM_toList [Monad m] [LawfulMonad m] (f : α m β) (arr : Array α) :
arr.mapM f = List.toArray <$> (arr.toList.mapM f) := by
rw [mapM_eq_foldlM, foldlM_toList, List.foldrM_reverse]
conv => rhs; rw [ List.reverse_reverse arr.toList]
induction arr.toList.reverse with
| nil => simp
| cons a l ih => simp [ih]
@[simp] theorem toList_mapM [Monad m] [LawfulMonad m] (f : α m β) (arr : Array α) :
toList <$> arr.mapM f = arr.toList.mapM f := by
simp [mapM_eq_mapM_toList]
theorem mapM_map_eq_foldl (as : Array α) (f : α β) (i) :
mapM.map (m := Id) f as i b = as.foldl (start := i) (fun r a => r.push (f a)) b := by
unfold mapM.map
split <;> rename_i h
· simp only [Id.bind_eq]
dsimp [foldl, Id.run, foldlM]
rw [mapM_map_eq_foldl, dif_pos (by omega), foldlM.loop, dif_pos h]
-- Calling `split` here gives a bad goal.
have : size as - i = Nat.succ (size as - i - 1) := by omega
rw [this]
simp [foldl, foldlM, Id.run, Nat.sub_add_eq]
· dsimp [foldl, Id.run, foldlM]
rw [dif_pos (by omega), foldlM.loop, dif_neg h]
rfl
termination_by as.size - i
theorem map_eq_foldl (as : Array α) (f : α β) :
as.map f = as.foldl (fun r a => r.push (f a)) #[] :=
mapM_map_eq_foldl _ _ _
theorem singleton_eq_toArray_singleton (a : α) : #[a] = [a].toArray := rfl
@[simp] theorem singleton_def (v : α) : singleton v = #[v] := rfl
-- This is a duplicate of `List.toArray_toList`.
-- It's confusing to guess which namespace this theorem should live in,
-- so we provide both.
@[simp] theorem toArray_toList (a : Array α) : a.toList.toArray = a := rfl
@[simp] theorem mkEmpty_eq (α n) : @mkEmpty α n = #[] := rfl
@[deprecated size_toArray (since := "2024-12-11")]
theorem size_mk (as : List α) : (Array.mk as).size = as.length := by simp [size]
/-- A more efficient version of `arr.toList.reverse`. -/
@[inline] def toListRev (arr : Array α) : List α := arr.foldl (fun l t => t :: l) []
@[simp] theorem toListRev_eq (arr : Array α) : arr.toListRev = arr.toList.reverse := by
rw [toListRev, foldl_toList, List.foldr_reverse, List.foldr_cons_nil]
@[simp] theorem appendList_nil (arr : Array α) : arr ++ ([] : List α) = arr := Array.ext' (by simp)
@[simp] theorem appendList_cons (arr : Array α) (a : α) (l : List α) :
@@ -1091,7 +1178,6 @@ theorem foldl_toList_eq_map (l : List α) (acc : Array β) (G : α → β) :
(l.foldl (fun acc a => acc.push (G a)) acc).toList = acc.toList ++ l.map G := by
induction l generalizing acc <;> simp [*]
/-! # uset -/
attribute [simp] uset
@@ -1267,18 +1353,16 @@ theorem get_set (a : Array α) (i : Nat) (hi : i < a.size) (j : Nat) (hj : j < a
(h : i j) : (a.set i v)[j]'(by simp [*]) = a[j] := by
simp only [set, getElem_toList, List.getElem_set_ne h]
private theorem fin_cast_val (e : n = n') (i : Fin n) : e i = i.1, e i.2 := by cases e; rfl
theorem swap_def (a : Array α) (i j : Nat) (hi hj) :
a.swap i j hi hj = (a.set i a[j]).set j a[i] (by simpa using hj) := by
simp [swap, fin_cast_val]
simp [swap]
@[simp] theorem toList_swap (a : Array α) (i j : Nat) (hi hj) :
(a.swap i j hi hj).toList = (a.toList.set i a[j]).set j a[i] := by simp [swap_def]
theorem getElem?_swap (a : Array α) (i j : Nat) (hi hj) (k : Nat) : (a.swap i j hi hj)[k]? =
if j = k then some a[i] else if i = k then some a[j] else a[k]? := by
simp [swap_def, get?_set, getElem_fin_eq_getElem_toList]
simp [swap_def, get?_set]
@[simp] theorem swapAt_def (a : Array α) (i : Nat) (v : α) (hi) :
a.swapAt i v hi = (a[i], a.set i v) := rfl
@@ -1393,6 +1477,90 @@ theorem getElem_range {n : Nat} {x : Nat} (h : x < (Array.range n).size) : (Arra
true_and, Nat.not_lt] at h
rw [List.getElem?_eq_none_iff.2 _, List.getElem?_eq_none_iff.2 (a.toList.length_reverse _)]
end Array
open Array
namespace List
@[simp] theorem reverse_toArray (l : List α) : l.toArray.reverse = l.reverse.toArray := by
apply ext'
simp only [toList_reverse]
end List
namespace Array
/-! ### foldlM and foldrM -/
theorem foldlM_append [Monad m] [LawfulMonad m] (f : β α m β) (b) (l l' : Array α) :
(l ++ l').foldlM f b = l.foldlM f b >>= l'.foldlM f := by
cases l
cases l'
rw [List.append_toArray]
simp
/-- Variant of `foldM_append` with `h : stop = (l ++ l').size`. -/
@[simp] theorem foldlM_append' [Monad m] [LawfulMonad m] (f : β α m β) (b) (l l' : Array α)
(h : stop = (l ++ l').size) :
(l ++ l').foldlM f b 0 stop = l.foldlM f b >>= l'.foldlM f := by
subst h
rw [foldlM_append]
theorem foldrM_push [Monad m] (f : α β m β) (init : β) (arr : Array α) (a : α) :
(arr.push a).foldrM f init = f a init >>= arr.foldrM f := by
simp only [foldrM_eq_reverse_foldlM_toList, push_toList, List.reverse_append, List.reverse_cons,
List.reverse_nil, List.nil_append, List.singleton_append, List.foldlM_cons, List.foldlM_reverse]
/--
Variant of `foldrM_push` with `h : start = arr.size + 1`
rather than `(arr.push a).size` as the argument.
-/
@[simp] theorem foldrM_push' [Monad m] (f : α β m β) (init : β) (arr : Array α) (a : α)
{start} (h : start = arr.size + 1) :
(arr.push a).foldrM f init start = f a init >>= arr.foldrM f := by
simp [ foldrM_push, h]
theorem foldl_eq_foldlM (f : β α β) (b) (l : Array α) :
l.foldl f b = l.foldlM (m := Id) f b := by
cases l
simp [List.foldl_eq_foldlM]
theorem foldr_eq_foldrM (f : α β β) (b) (l : Array α) :
l.foldr f b = l.foldrM (m := Id) f b := by
cases l
simp [List.foldr_eq_foldrM]
@[simp] theorem id_run_foldlM (f : β α Id β) (b) (l : Array α) :
Id.run (l.foldlM f b) = l.foldl f b := (foldl_eq_foldlM f b l).symm
@[simp] theorem id_run_foldrM (f : α β Id β) (b) (l : Array α) :
Id.run (l.foldrM f b) = l.foldr f b := (foldr_eq_foldrM f b l).symm
/-! ### foldl and foldr -/
theorem foldr_push (f : α β β) (init : β) (arr : Array α) (a : α) :
(arr.push a).foldr f init = arr.foldr f (f a init) := foldrM_push ..
/--
Variant of `foldr_push` with the `h : start = arr.size + 1`
rather than `(arr.push a).size` as the argument.
-/
@[simp] theorem foldr_push' (f : α β β) (init : β) (arr : Array α) (a : α) {start}
(h : start = arr.size + 1) : (arr.push a).foldr f init start = arr.foldr f (f a init) :=
foldrM_push' _ _ _ _ h
@[simp] theorem foldl_push_eq_append (l l' : Array α) : l.foldl push l' = l' ++ l := by
cases l
cases l'
simp
@[simp] theorem foldr_flip_push_eq_append (l l' : Array α) :
l.foldr (fun x y => push y x) l' = l' ++ l.reverse := by
cases l
cases l'
simp
/-! ### take -/
@[simp] theorem size_take_loop (a : Array α) (n : Nat) : (take.loop n a).size = a.size - n := by
@@ -1505,22 +1673,6 @@ theorem foldr_congr {as bs : Array α} (h₀ : as = bs) {f g : α → β → β}
as.foldr f a start stop = bs.foldr g b start' stop' := by
congr
theorem foldl_eq_foldlM (f : β α β) (b) (l : Array α) :
l.foldl f b = l.foldlM (m := Id) f b := by
cases l
simp [List.foldl_eq_foldlM]
theorem foldr_eq_foldrM (f : α β β) (b) (l : Array α) :
l.foldr f b = l.foldrM (m := Id) f b := by
cases l
simp [List.foldr_eq_foldrM]
@[simp] theorem id_run_foldlM (f : β α Id β) (b) (l : Array α) :
Id.run (l.foldlM f b) = l.foldl f b := (foldl_eq_foldlM f b l).symm
@[simp] theorem id_run_foldrM (f : α β Id β) (b) (l : Array α) :
Id.run (l.foldrM f b) = l.foldr f b := (foldr_eq_foldrM f b l).symm
theorem foldl_hom (f : α₁ α₂) (g₁ : α₁ β α₁) (g₂ : α₂ β α₂) (l : Array β) (init : α₁)
(H : x y, g₂ (f x) y = f (g₁ x y)) : l.foldl g₂ (f init) = f (l.foldl g₁ init) := by
cases l
@@ -1535,45 +1687,13 @@ theorem foldr_hom (f : β₁ → β₂) (g₁ : α → β₁ → β₁) (g₂ :
/-! ### map -/
@[simp] theorem mem_map {f : α β} {l : Array α} : b l.map f a, a l f a = b := by
simp only [mem_def, toList_map, List.mem_map]
theorem exists_of_mem_map (h : b map f l) : a, a l f a = b := mem_map.1 h
theorem mem_map_of_mem (f : α β) (h : a l) : f a map f l := mem_map.2 _, h, rfl
theorem mapM_eq_mapM_toList [Monad m] [LawfulMonad m] (f : α m β) (arr : Array α) :
arr.mapM f = List.toArray <$> (arr.toList.mapM f) := by
rw [mapM_eq_foldlM, foldlM_toList, List.foldrM_reverse]
conv => rhs; rw [ List.reverse_reverse arr.toList]
induction arr.toList.reverse with
| nil => simp
| cons a l ih => simp [ih]
@[simp] theorem toList_mapM [Monad m] [LawfulMonad m] (f : α m β) (arr : Array α) :
toList <$> arr.mapM f = arr.toList.mapM f := by
simp [mapM_eq_mapM_toList]
theorem mapM_map_eq_foldl (as : Array α) (f : α β) (i) :
mapM.map (m := Id) f as i b = as.foldl (start := i) (fun r a => r.push (f a)) b := by
unfold mapM.map
split <;> rename_i h
· simp only [Id.bind_eq]
dsimp [foldl, Id.run, foldlM]
rw [mapM_map_eq_foldl, dif_pos (by omega), foldlM.loop, dif_pos h]
-- Calling `split` here gives a bad goal.
have : size as - i = Nat.succ (size as - i - 1) := by omega
rw [this]
simp [foldl, foldlM, Id.run, Nat.sub_add_eq]
· dsimp [foldl, Id.run, foldlM]
rw [dif_pos (by omega), foldlM.loop, dif_neg h]
rfl
termination_by as.size - i
theorem map_eq_foldl (as : Array α) (f : α β) :
as.map f = as.foldl (fun r a => r.push (f a)) #[] :=
mapM_map_eq_foldl _ _ _
@[simp] theorem map_pop {f : α β} {as : Array α} :
as.pop.map f = (as.map f).pop := by
ext
· simp
· simp only [getElem_map, getElem_pop, size_map]
@[deprecated "Use `toList_map` or `List.map_toArray` to characterize `Array.map`." (since := "2025-01-06")]
theorem map_induction (as : Array α) (f : α β) (motive : Nat Prop) (h0 : motive 0)
(p : Fin as.size β Prop) (hs : i, motive i.1 p i (f as[i]) motive (i+1)) :
motive as.size
@@ -1601,36 +1721,13 @@ theorem map_induction (as : Array α) (f : α → β) (motive : Nat → Prop) (h
simp only [show j = i by omega]
exact (hs _ m).1
set_option linter.deprecated false in
@[deprecated "Use `toList_map` or `List.map_toArray` to characterize `Array.map`." (since := "2025-01-06")]
theorem map_spec (as : Array α) (f : α β) (p : Fin as.size β Prop)
(hs : i, p i (f as[i])) :
eq : (as.map f).size = as.size, i h, p i, h ((as.map f)[i]) := by
simpa using map_induction as f (fun _ => True) trivial p (by simp_all)
@[simp] theorem getElem_map (f : α β) (as : Array α) (i : Nat) (h) :
(as.map f)[i] = f (as[i]'(size_map .. h)) := by
have := map_spec as f (fun i b => b = f (as[i]))
simp only [implies_true, true_implies] at this
obtain eq, w := this
apply w
simp_all
@[simp] theorem getElem?_map (f : α β) (as : Array α) (i : Nat) :
(as.map f)[i]? = as[i]?.map f := by
simp [getElem?_def]
@[simp] theorem map_push {f : α β} {as : Array α} {x : α} :
(as.push x).map f = (as.map f).push (f x) := by
ext
· simp
· simp only [getElem_map, getElem_push, size_map]
split <;> rfl
@[simp] theorem map_pop {f : α β} {as : Array α} :
as.pop.map f = (as.map f).pop := by
ext
· simp
· simp only [getElem_map, getElem_pop, size_map]
/-! ### modify -/
@[simp] theorem size_modify (a : Array α) (i : Nat) (f : α α) : (a.modify i f).size = a.size := by
@@ -2117,10 +2214,6 @@ theorem toListRev_toArray (l : List α) : l.toArray.toListRev = l.reverse := by
rw [size_toArray, mapM'_cons, foldlM_toArray]
simp [ih]
@[simp] theorem map_toArray (f : α β) (l : List α) : l.toArray.map f = (l.map f).toArray := by
apply ext'
simp
theorem uset_toArray (l : List α) (i : USize) (a : α) (h : i.toNat < l.toArray.size) :
l.toArray.uset i a h = (l.set i.toNat a).toArray := by simp
@@ -2129,10 +2222,6 @@ theorem uset_toArray (l : List α) (i : USize) (a : α) (h : i.toNat < l.toArray
apply ext'
simp
@[simp] theorem reverse_toArray (l : List α) : l.toArray.reverse = l.reverse.toArray := by
apply ext'
simp
@[simp] theorem modify_toArray (f : α α) (l : List α) :
l.toArray.modify i f = (l.modify f i).toArray := by
apply ext'
@@ -2226,29 +2315,6 @@ namespace Array
/-! ### map -/
@[simp] theorem map_map {f : α β} {g : β γ} {as : Array α} :
(as.map f).map g = as.map (g f) := by
cases as; simp
@[simp] theorem map_id_fun : map (id : α α) = id := by
funext l
induction l <;> simp_all
/-- `map_id_fun'` differs from `map_id_fun` by representing the identity function as a lambda, rather than `id`. -/
@[simp] theorem map_id_fun' : map (fun (a : α) => a) = id := map_id_fun
-- This is not a `@[simp]` lemma because `map_id_fun` will apply.
theorem map_id (as : Array α) : map (id : α α) as = as := by
cases as <;> simp_all
/-- `map_id'` differs from `map_id` by representing the identity function as a lambda, rather than `id`. -/
-- This is not a `@[simp]` lemma because `map_id_fun'` will apply.
theorem map_id' (as : Array α) : map (fun (a : α) => a) as = as := map_id as
/-- Variant of `map_id`, with a side condition that the function is pointwise the identity. -/
theorem map_id'' {f : α α} (h : x, f x = x) (as : Array α) : map f as = as := by
simp [show f = id from funext h]
theorem array_array_induction (P : Array (Array α) Prop) (h : (xss : List (List α)), P (xss.map List.toArray).toArray)
(ass : Array (Array α)) : P ass := by
specialize h (ass.toList.map toList)

View File

@@ -17,8 +17,8 @@ namespace Array
@[simp] theorem lt_toList [LT α] (l₁ l₂ : Array α) : l₁.toList < l₂.toList l₁ < l₂ := Iff.rfl
@[simp] theorem le_toList [LT α] (l₁ l₂ : Array α) : l₁.toList l₂.toList l₁ l₂ := Iff.rfl
theorem not_lt_iff_ge [LT α] (l₁ l₂ : List α) : ¬ l₁ < l₂ l₂ l₁ := Iff.rfl
theorem not_le_iff_gt [DecidableEq α] [LT α] [DecidableLT α] (l₁ l₂ : List α) :
protected theorem not_lt_iff_ge [LT α] (l₁ l₂ : List α) : ¬ l₁ < l₂ l₂ l₁ := Iff.rfl
protected theorem not_le_iff_gt [DecidableEq α] [LT α] [DecidableLT α] (l₁ l₂ : List α) :
¬ l₁ l₂ l₂ < l₁ :=
Decidable.not_not
@@ -135,7 +135,7 @@ protected theorem le_of_lt [DecidableEq α] [LT α] [DecidableLT α]
{l₁ l₂ : Array α} (h : l₁ < l₂) : l₁ l₂ :=
List.le_of_lt h
theorem le_iff_lt_or_eq [DecidableEq α] [LT α] [DecidableLT α]
protected theorem le_iff_lt_or_eq [DecidableEq α] [LT α] [DecidableLT α]
[Std.Irrefl (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)]
[Std.Total (¬ · < · : α α Prop)]
@@ -211,7 +211,7 @@ theorem lex_eq_false_iff_exists [BEq α] [PartialEquivBEq α] (lt : αα
cases l₂
simp_all [List.lex_eq_false_iff_exists]
theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : Array α} :
protected theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : Array α} :
l₁ < l₂
(l₁ = l₂.take l₁.size l₁.size < l₂.size)
( (i : Nat) (h₁ : i < l₁.size) (h₂ : i < l₂.size),
@@ -221,7 +221,7 @@ theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : Arr
cases l₂
simp [List.lt_iff_exists]
theorem le_iff_exists [DecidableEq α] [LT α] [DecidableLT α]
protected theorem le_iff_exists [DecidableEq α] [LT α] [DecidableLT α]
[Std.Irrefl (· < · : α α Prop)]
[Std.Asymm (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)] {l₁ l₂ : Array α} :
@@ -258,14 +258,14 @@ theorem le_append_left [LT α] [Std.Irrefl (· < · : αα → Prop)]
cases l₂
simpa using List.le_append_left
theorem map_lt [LT α] [LT β]
protected theorem map_lt [LT α] [LT β]
{l₁ l₂ : Array α} {f : α β} (w : x y, x < y f x < f y) (h : l₁ < l₂) :
map f l₁ < map f l₂ := by
cases l₁
cases l₂
simpa using List.map_lt w h
theorem map_le [DecidableEq α] [LT α] [DecidableLT α] [DecidableEq β] [LT β] [DecidableLT β]
protected theorem map_le [DecidableEq α] [LT α] [DecidableLT α] [DecidableEq β] [LT β] [DecidableLT β]
[Std.Irrefl (· < · : α α Prop)]
[Std.Asymm (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)]

View File

@@ -1098,6 +1098,32 @@ theorem bmod_def (x : Int) (m : Nat) : bmod x m =
(x % m) - m :=
rfl
theorem bdiv_add_bmod (x : Int) (m : Nat) : m * bdiv x m + bmod x m = x := by
unfold bdiv bmod
split
· simp_all only [Nat.cast_ofNat_Int, Int.mul_zero, emod_zero, Int.zero_add, Int.sub_zero,
ite_self]
· dsimp only
split
· exact ediv_add_emod x m
· rw [Int.mul_add, Int.mul_one, Int.add_assoc, Int.add_comm m, Int.sub_add_cancel]
exact ediv_add_emod x m
theorem bmod_add_bdiv (x : Int) (m : Nat) : bmod x m + m * bdiv x m = x := by
rw [Int.add_comm]; exact bdiv_add_bmod x m
theorem bdiv_add_bmod' (x : Int) (m : Nat) : bdiv x m * m + bmod x m = x := by
rw [Int.mul_comm]; exact bdiv_add_bmod x m
theorem bmod_add_bdiv' (x : Int) (m : Nat) : bmod x m + bdiv x m * m = x := by
rw [Int.add_comm]; exact bdiv_add_bmod' x m
theorem bmod_eq_self_sub_mul_bdiv (x : Int) (m : Nat) : bmod x m = x - m * bdiv x m := by
rw [ Int.add_sub_cancel (bmod x m), bmod_add_bdiv]
theorem bmod_eq_self_sub_bdiv_mul (x : Int) (m : Nat) : bmod x m = x - bdiv x m * m := by
rw [ Int.add_sub_cancel (bmod x m), bmod_add_bdiv']
theorem bmod_pos (x : Int) (m : Nat) (p : x % m < (m + 1) / 2) : bmod x m = x % m := by
simp [bmod_def, p]

View File

@@ -757,207 +757,6 @@ theorem length_eq_of_beq [BEq α] {l₁ l₂ : List α} (h : l₁ == l₂) : l
| nil => simp
| cons b l₂ => simp [isEqv, ih]
/-! ### foldlM and foldrM -/
@[simp] theorem foldlM_reverse [Monad m] (l : List α) (f : β α m β) (b) :
l.reverse.foldlM f b = l.foldrM (fun x y => f y x) b := rfl
@[simp] theorem foldlM_append [Monad m] [LawfulMonad m] (f : β α m β) (b) (l l' : List α) :
(l ++ l').foldlM f b = l.foldlM f b >>= l'.foldlM f := by
induction l generalizing b <;> simp [*]
@[simp] theorem foldrM_cons [Monad m] [LawfulMonad m] (a : α) (l) (f : α β m β) (b) :
(a :: l).foldrM f b = l.foldrM f b >>= f a := by
simp only [foldrM]
induction l <;> simp_all
theorem foldl_eq_foldlM (f : β α β) (b) (l : List α) :
l.foldl f b = l.foldlM (m := Id) f b := by
induction l generalizing b <;> simp [*, foldl]
theorem foldr_eq_foldrM (f : α β β) (b) (l : List α) :
l.foldr f b = l.foldrM (m := Id) f b := by
induction l <;> simp [*, foldr]
@[simp] theorem id_run_foldlM (f : β α Id β) (b) (l : List α) :
Id.run (l.foldlM f b) = l.foldl f b := (foldl_eq_foldlM f b l).symm
@[simp] theorem id_run_foldrM (f : α β Id β) (b) (l : List α) :
Id.run (l.foldrM f b) = l.foldr f b := (foldr_eq_foldrM f b l).symm
/-! ### foldl and foldr -/
@[simp] theorem foldr_cons_eq_append (l : List α) : l.foldr cons l' = l ++ l' := by
induction l <;> simp [*]
@[deprecated foldr_cons_eq_append (since := "2024-08-22")] abbrev foldr_self_append := @foldr_cons_eq_append
@[simp] theorem foldl_flip_cons_eq_append (l : List α) : l.foldl (fun x y => y :: x) l' = l.reverse ++ l' := by
induction l generalizing l' <;> simp [*]
theorem foldr_cons_nil (l : List α) : l.foldr cons [] = l := by simp
@[deprecated foldr_cons_nil (since := "2024-09-04")] abbrev foldr_self := @foldr_cons_nil
theorem foldl_map (f : β₁ β₂) (g : α β₂ α) (l : List β₁) (init : α) :
(l.map f).foldl g init = l.foldl (fun x y => g x (f y)) init := by
induction l generalizing init <;> simp [*]
theorem foldr_map (f : α₁ α₂) (g : α₂ β β) (l : List α₁) (init : β) :
(l.map f).foldr g init = l.foldr (fun x y => g (f x) y) init := by
induction l generalizing init <;> simp [*]
theorem foldl_filterMap (f : α Option β) (g : γ β γ) (l : List α) (init : γ) :
(l.filterMap f).foldl g init = l.foldl (fun x y => match f y with | some b => g x b | none => x) init := by
induction l generalizing init with
| nil => rfl
| cons a l ih =>
simp only [filterMap_cons, foldl_cons]
cases f a <;> simp [ih]
theorem foldr_filterMap (f : α Option β) (g : β γ γ) (l : List α) (init : γ) :
(l.filterMap f).foldr g init = l.foldr (fun x y => match f x with | some b => g b y | none => y) init := by
induction l generalizing init with
| nil => rfl
| cons a l ih =>
simp only [filterMap_cons, foldr_cons]
cases f a <;> simp [ih]
theorem foldl_map' (g : α β) (f : α α α) (f' : β β β) (a : α) (l : List α)
(h : x y, f' (g x) (g y) = g (f x y)) :
(l.map g).foldl f' (g a) = g (l.foldl f a) := by
induction l generalizing a
· simp
· simp [*, h]
theorem foldr_map' (g : α β) (f : α α α) (f' : β β β) (a : α) (l : List α)
(h : x y, f' (g x) (g y) = g (f x y)) :
(l.map g).foldr f' (g a) = g (l.foldr f a) := by
induction l generalizing a
· simp
· simp [*, h]
theorem foldl_assoc {op : α α α} [ha : Std.Associative op] :
{l : List α} {a₁ a₂}, l.foldl op (op a₁ a₂) = op a₁ (l.foldl op a₂)
| [], a₁, a₂ => rfl
| a :: l, a₁, a₂ => by
simp only [foldl_cons, ha.assoc]
rw [foldl_assoc]
theorem foldr_assoc {op : α α α} [ha : Std.Associative op] :
{l : List α} {a₁ a₂}, l.foldr op (op a₁ a₂) = op (l.foldr op a₁) a₂
| [], a₁, a₂ => rfl
| a :: l, a₁, a₂ => by
simp only [foldr_cons, ha.assoc]
rw [foldr_assoc]
theorem foldl_hom (f : α₁ α₂) (g₁ : α₁ β α₁) (g₂ : α₂ β α₂) (l : List β) (init : α₁)
(H : x y, g₂ (f x) y = f (g₁ x y)) : l.foldl g₂ (f init) = f (l.foldl g₁ init) := by
induction l generalizing init <;> simp [*, H]
theorem foldr_hom (f : β₁ β₂) (g₁ : α β₁ β₁) (g₂ : α β₂ β₂) (l : List α) (init : β₁)
(H : x y, g₂ x (f y) = f (g₁ x y)) : l.foldr g₂ (f init) = f (l.foldr g₁ init) := by
induction l <;> simp [*, H]
/--
Prove a proposition about the result of `List.foldl`,
by proving it for the initial data,
and the implication that the operation applied to any element of the list preserves the property.
The motive can take values in `Sort _`, so this may be used to construct data,
as well as to prove propositions.
-/
def foldlRecOn {motive : β Sort _} : (l : List α) (op : β α β) (b : β) (_ : motive b)
(_ : (b : β) (_ : motive b) (a : α) (_ : a l), motive (op b a)), motive (List.foldl op b l)
| [], _, _, hb, _ => hb
| hd :: tl, op, b, hb, hl =>
foldlRecOn tl op (op b hd) (hl b hb hd (mem_cons_self hd tl))
fun y hy x hx => hl y hy x (mem_cons_of_mem hd hx)
@[simp] theorem foldlRecOn_nil {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a []), motive (op b a)) :
foldlRecOn [] op b hb hl = hb := rfl
@[simp] theorem foldlRecOn_cons {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a x :: l), motive (op b a)) :
foldlRecOn (x :: l) op b hb hl =
foldlRecOn l op (op b x) (hl b hb x (mem_cons_self x l))
(fun b c a m => hl b c a (mem_cons_of_mem x m)) :=
rfl
/--
Prove a proposition about the result of `List.foldr`,
by proving it for the initial data,
and the implication that the operation applied to any element of the list preserves the property.
The motive can take values in `Sort _`, so this may be used to construct data,
as well as to prove propositions.
-/
def foldrRecOn {motive : β Sort _} : (l : List α) (op : α β β) (b : β) (_ : motive b)
(_ : (b : β) (_ : motive b) (a : α) (_ : a l), motive (op a b)), motive (List.foldr op b l)
| nil, _, _, hb, _ => hb
| x :: l, op, b, hb, hl =>
hl (foldr op b l)
(foldrRecOn l op b hb fun b c a m => hl b c a (mem_cons_of_mem x m)) x (mem_cons_self x l)
@[simp] theorem foldrRecOn_nil {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a []), motive (op a b)) :
foldrRecOn [] op b hb hl = hb := rfl
@[simp] theorem foldrRecOn_cons {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a x :: l), motive (op a b)) :
foldrRecOn (x :: l) op b hb hl =
hl _ (foldrRecOn l op b hb fun b c a m => hl b c a (mem_cons_of_mem x m))
x (mem_cons_self x l) :=
rfl
/--
We can prove that two folds over the same list are related (by some arbitrary relation)
if we know that the initial elements are related and the folding function, for each element of the list,
preserves the relation.
-/
theorem foldl_rel {l : List α} {f g : β α β} {a b : β} (r : β β Prop)
(h : r a b) (h' : (a : α), a l (c c' : β), r c c' r (f c a) (g c' a)) :
r (l.foldl (fun acc a => f acc a) a) (l.foldl (fun acc a => g acc a) b) := by
induction l generalizing a b with
| nil => simp_all
| cons a l ih =>
simp only [foldl_cons]
apply ih
· simp_all
· exact fun a m c c' h => h' _ (by simp_all) _ _ h
/--
We can prove that two folds over the same list are related (by some arbitrary relation)
if we know that the initial elements are related and the folding function, for each element of the list,
preserves the relation.
-/
theorem foldr_rel {l : List α} {f g : α β β} {a b : β} (r : β β Prop)
(h : r a b) (h' : (a : α), a l (c c' : β), r c c' r (f a c) (g a c')) :
r (l.foldr (fun a acc => f a acc) a) (l.foldr (fun a acc => g a acc) b) := by
induction l generalizing a b with
| nil => simp_all
| cons a l ih =>
simp only [foldr_cons]
apply h'
· simp
· exact ih h fun a m c c' h => h' _ (by simp_all) _ _ h
@[simp] theorem foldl_add_const (l : List α) (a b : Nat) :
l.foldl (fun x _ => x + a) b = b + a * l.length := by
induction l generalizing b with
| nil => simp
| cons y l ih =>
simp only [foldl_cons, ih, length_cons, Nat.mul_add, Nat.mul_one, Nat.add_assoc,
Nat.add_comm a]
@[simp] theorem foldr_add_const (l : List α) (a b : Nat) :
l.foldr (fun _ x => x + a) b = b + a * l.length := by
induction l generalizing b with
| nil => simp
| cons y l ih =>
simp only [foldr_cons, ih, length_cons, Nat.mul_add, Nat.mul_one, Nat.add_assoc]
/-! ### getLast -/
theorem getLast_eq_getElem : (l : List α) (h : l []),
@@ -1216,27 +1015,6 @@ theorem getLast?_tail (l : List α) : (tail l).getLast? = if l.length = 1 then n
/-! ### map -/
@[simp] theorem map_id_fun : map (id : α α) = id := by
funext l
induction l <;> simp_all
/-- `map_id_fun'` differs from `map_id_fun` by representing the identity function as a lambda, rather than `id`. -/
@[simp] theorem map_id_fun' : map (fun (a : α) => a) = id := map_id_fun
-- This is not a `@[simp]` lemma because `map_id_fun` will apply.
theorem map_id (l : List α) : map (id : α α) l = l := by
induction l <;> simp_all
/-- `map_id'` differs from `map_id` by representing the identity function as a lambda, rather than `id`. -/
-- This is not a `@[simp]` lemma because `map_id_fun'` will apply.
theorem map_id' (l : List α) : map (fun (a : α) => a) l = l := map_id l
/-- Variant of `map_id`, with a side condition that the function is pointwise the identity. -/
theorem map_id'' {f : α α} (h : x, f x = x) (l : List α) : map f l = l := by
simp [show f = id from funext h]
theorem map_singleton (f : α β) (a : α) : map f [a] = [f a] := rfl
@[simp] theorem length_map (as : List α) (f : α β) : (as.map f).length = as.length := by
induction as with
| nil => simp [List.map]
@@ -1262,6 +1040,27 @@ theorem get_map (f : α → β) {l i} :
get (map f l) i = f (get l i, length_map l f i.2) := by
simp
@[simp] theorem map_id_fun : map (id : α α) = id := by
funext l
induction l <;> simp_all
/-- `map_id_fun'` differs from `map_id_fun` by representing the identity function as a lambda, rather than `id`. -/
@[simp] theorem map_id_fun' : map (fun (a : α) => a) = id := map_id_fun
-- This is not a `@[simp]` lemma because `map_id_fun` will apply.
theorem map_id (l : List α) : map (id : α α) l = l := by
induction l <;> simp_all
/-- `map_id'` differs from `map_id` by representing the identity function as a lambda, rather than `id`. -/
-- This is not a `@[simp]` lemma because `map_id_fun'` will apply.
theorem map_id' (l : List α) : map (fun (a : α) => a) l = l := map_id l
/-- Variant of `map_id`, with a side condition that the function is pointwise the identity. -/
theorem map_id'' {f : α α} (h : x, f x = x) (l : List α) : map f l = l := by
simp [show f = id from funext h]
theorem map_singleton (f : α β) (a : α) : map f [a] = [f a] := rfl
@[simp] theorem mem_map {f : α β} : {l : List α}, b l.map f a, a l f a = b
| [] => by simp
| _ :: l => by simp [mem_map (l := l), eq_comm (a := b)]
@@ -1961,16 +1760,6 @@ theorem set_append {s t : List α} :
(s ++ t).set i x = s ++ t.set (i - s.length) x := by
rw [set_append, if_neg (by simp_all)]
@[simp] theorem foldrM_append [Monad m] [LawfulMonad m] (f : α β m β) (b) (l l' : List α) :
(l ++ l').foldrM f b = l'.foldrM f b >>= l.foldrM f := by
induction l <;> simp [*]
@[simp] theorem foldl_append {β : Type _} (f : β α β) (b) (l l' : List α) :
(l ++ l').foldl f b = l'.foldl f (l.foldl f b) := by simp [foldl_eq_foldlM]
@[simp] theorem foldr_append (f : α β β) (b) (l l' : List α) :
(l ++ l').foldr f b = l.foldr f (l'.foldr f b) := by simp [foldr_eq_foldrM]
theorem filterMap_eq_append_iff {f : α Option β} :
filterMap f l = L₁ ++ L₂ l₁ l₂, l = l₁ ++ l₂ filterMap f l₁ = L₁ filterMap f l₂ = L₂ := by
constructor
@@ -2119,14 +1908,6 @@ theorem head?_flatten {L : List (List α)} : (flatten L).head? = L.findSome? fun
-- `getLast?_flatten` is proved later, after the `reverse` section.
-- `head_flatten` and `getLast_flatten` are proved in `Init.Data.List.Find`.
theorem foldl_flatten (f : β α β) (b : β) (L : List (List α)) :
(flatten L).foldl f b = L.foldl (fun b l => l.foldl f b) b := by
induction L generalizing b <;> simp_all
theorem foldr_flatten (f : α β β) (b : β) (L : List (List α)) :
(flatten L).foldr f b = L.foldr (fun l b => l.foldr f b) b := by
induction L <;> simp_all
@[simp] theorem map_flatten (f : α β) (L : List (List α)) : map f (flatten L) = flatten (map (map f) L) := by
induction L <;> simp_all
@@ -2699,10 +2480,114 @@ theorem flatMap_reverse {β} (l : List α) (f : α → List β) : (l.reverse.fla
@[simp] theorem reverseAux_eq (as bs : List α) : reverseAux as bs = reverse as ++ bs :=
reverseAux_eq_append ..
@[simp] theorem reverse_replicate (n) (a : α) : reverse (replicate n a) = replicate n a :=
eq_replicate_iff.2
by rw [length_reverse, length_replicate],
fun _ h => eq_of_mem_replicate (mem_reverse.1 h)
/-! ### foldlM and foldrM -/
@[simp] theorem foldlM_append [Monad m] [LawfulMonad m] (f : β α m β) (b) (l l' : List α) :
(l ++ l').foldlM f b = l.foldlM f b >>= l'.foldlM f := by
induction l generalizing b <;> simp [*]
@[simp] theorem foldrM_cons [Monad m] [LawfulMonad m] (a : α) (l) (f : α β m β) (b) :
(a :: l).foldrM f b = l.foldrM f b >>= f a := by
simp only [foldrM]
induction l <;> simp_all
theorem foldl_eq_foldlM (f : β α β) (b) (l : List α) :
l.foldl f b = l.foldlM (m := Id) f b := by
induction l generalizing b <;> simp [*, foldl]
theorem foldr_eq_foldrM (f : α β β) (b) (l : List α) :
l.foldr f b = l.foldrM (m := Id) f b := by
induction l <;> simp [*, foldr]
@[simp] theorem id_run_foldlM (f : β α Id β) (b) (l : List α) :
Id.run (l.foldlM f b) = l.foldl f b := (foldl_eq_foldlM f b l).symm
@[simp] theorem id_run_foldrM (f : α β Id β) (b) (l : List α) :
Id.run (l.foldrM f b) = l.foldr f b := (foldr_eq_foldrM f b l).symm
@[simp] theorem foldlM_reverse [Monad m] (l : List α) (f : β α m β) (b) :
l.reverse.foldlM f b = l.foldrM (fun x y => f y x) b := rfl
@[simp] theorem foldrM_reverse [Monad m] (l : List α) (f : α β m β) (b) :
l.reverse.foldrM f b = l.foldlM (fun x y => f y x) b :=
(foldlM_reverse ..).symm.trans <| by simp
/-! ### foldl and foldr -/
@[simp] theorem foldr_cons_eq_append (l : List α) : l.foldr cons l' = l ++ l' := by
induction l <;> simp [*]
@[deprecated foldr_cons_eq_append (since := "2024-08-22")] abbrev foldr_self_append := @foldr_cons_eq_append
@[simp] theorem foldl_flip_cons_eq_append (l : List α) : l.foldl (fun x y => y :: x) l' = l.reverse ++ l' := by
induction l generalizing l' <;> simp [*]
theorem foldr_cons_nil (l : List α) : l.foldr cons [] = l := by simp
@[deprecated foldr_cons_nil (since := "2024-09-04")] abbrev foldr_self := @foldr_cons_nil
theorem foldl_map (f : β₁ β₂) (g : α β₂ α) (l : List β₁) (init : α) :
(l.map f).foldl g init = l.foldl (fun x y => g x (f y)) init := by
induction l generalizing init <;> simp [*]
theorem foldr_map (f : α₁ α₂) (g : α₂ β β) (l : List α₁) (init : β) :
(l.map f).foldr g init = l.foldr (fun x y => g (f x) y) init := by
induction l generalizing init <;> simp [*]
theorem foldl_filterMap (f : α Option β) (g : γ β γ) (l : List α) (init : γ) :
(l.filterMap f).foldl g init = l.foldl (fun x y => match f y with | some b => g x b | none => x) init := by
induction l generalizing init with
| nil => rfl
| cons a l ih =>
simp only [filterMap_cons, foldl_cons]
cases f a <;> simp [ih]
theorem foldr_filterMap (f : α Option β) (g : β γ γ) (l : List α) (init : γ) :
(l.filterMap f).foldr g init = l.foldr (fun x y => match f x with | some b => g b y | none => y) init := by
induction l generalizing init with
| nil => rfl
| cons a l ih =>
simp only [filterMap_cons, foldr_cons]
cases f a <;> simp [ih]
theorem foldl_map' (g : α β) (f : α α α) (f' : β β β) (a : α) (l : List α)
(h : x y, f' (g x) (g y) = g (f x y)) :
(l.map g).foldl f' (g a) = g (l.foldl f a) := by
induction l generalizing a
· simp
· simp [*, h]
theorem foldr_map' (g : α β) (f : α α α) (f' : β β β) (a : α) (l : List α)
(h : x y, f' (g x) (g y) = g (f x y)) :
(l.map g).foldr f' (g a) = g (l.foldr f a) := by
induction l generalizing a
· simp
· simp [*, h]
@[simp] theorem foldrM_append [Monad m] [LawfulMonad m] (f : α β m β) (b) (l l' : List α) :
(l ++ l').foldrM f b = l'.foldrM f b >>= l.foldrM f := by
induction l <;> simp [*]
@[simp] theorem foldl_append {β : Type _} (f : β α β) (b) (l l' : List α) :
(l ++ l').foldl f b = l'.foldl f (l.foldl f b) := by simp [foldl_eq_foldlM]
@[simp] theorem foldr_append (f : α β β) (b) (l l' : List α) :
(l ++ l').foldr f b = l.foldr f (l'.foldr f b) := by simp [foldr_eq_foldrM]
theorem foldl_flatten (f : β α β) (b : β) (L : List (List α)) :
(flatten L).foldl f b = L.foldl (fun b l => l.foldl f b) b := by
induction L generalizing b <;> simp_all
theorem foldr_flatten (f : α β β) (b : β) (L : List (List α)) :
(flatten L).foldr f b = L.foldr (fun l b => l.foldr f b) b := by
induction L <;> simp_all
@[simp] theorem foldl_reverse (l : List α) (f : β α β) (b) :
l.reverse.foldl f b = l.foldr (fun x y => f y x) b := by simp [foldl_eq_foldlM, foldr_eq_foldrM]
@@ -2716,10 +2601,127 @@ theorem foldl_eq_foldr_reverse (l : List α) (f : β → α → β) (b) :
theorem foldr_eq_foldl_reverse (l : List α) (f : α β β) (b) :
l.foldr f b = l.reverse.foldl (fun x y => f y x) b := by simp
@[simp] theorem reverse_replicate (n) (a : α) : reverse (replicate n a) = replicate n a :=
eq_replicate_iff.2
by rw [length_reverse, length_replicate],
fun _ h => eq_of_mem_replicate (mem_reverse.1 h)
theorem foldl_assoc {op : α α α} [ha : Std.Associative op] :
{l : List α} {a₁ a₂}, l.foldl op (op a₁ a₂) = op a₁ (l.foldl op a₂)
| [], a₁, a₂ => rfl
| a :: l, a₁, a₂ => by
simp only [foldl_cons, ha.assoc]
rw [foldl_assoc]
theorem foldr_assoc {op : α α α} [ha : Std.Associative op] :
{l : List α} {a₁ a₂}, l.foldr op (op a₁ a₂) = op (l.foldr op a₁) a₂
| [], a₁, a₂ => rfl
| a :: l, a₁, a₂ => by
simp only [foldr_cons, ha.assoc]
rw [foldr_assoc]
theorem foldl_hom (f : α₁ α₂) (g₁ : α₁ β α₁) (g₂ : α₂ β α₂) (l : List β) (init : α₁)
(H : x y, g₂ (f x) y = f (g₁ x y)) : l.foldl g₂ (f init) = f (l.foldl g₁ init) := by
induction l generalizing init <;> simp [*, H]
theorem foldr_hom (f : β₁ β₂) (g₁ : α β₁ β₁) (g₂ : α β₂ β₂) (l : List α) (init : β₁)
(H : x y, g₂ x (f y) = f (g₁ x y)) : l.foldr g₂ (f init) = f (l.foldr g₁ init) := by
induction l <;> simp [*, H]
/--
Prove a proposition about the result of `List.foldl`,
by proving it for the initial data,
and the implication that the operation applied to any element of the list preserves the property.
The motive can take values in `Sort _`, so this may be used to construct data,
as well as to prove propositions.
-/
def foldlRecOn {motive : β Sort _} : (l : List α) (op : β α β) (b : β) (_ : motive b)
(_ : (b : β) (_ : motive b) (a : α) (_ : a l), motive (op b a)), motive (List.foldl op b l)
| [], _, _, hb, _ => hb
| hd :: tl, op, b, hb, hl =>
foldlRecOn tl op (op b hd) (hl b hb hd (mem_cons_self hd tl))
fun y hy x hx => hl y hy x (mem_cons_of_mem hd hx)
@[simp] theorem foldlRecOn_nil {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a []), motive (op b a)) :
foldlRecOn [] op b hb hl = hb := rfl
@[simp] theorem foldlRecOn_cons {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a x :: l), motive (op b a)) :
foldlRecOn (x :: l) op b hb hl =
foldlRecOn l op (op b x) (hl b hb x (mem_cons_self x l))
(fun b c a m => hl b c a (mem_cons_of_mem x m)) :=
rfl
/--
Prove a proposition about the result of `List.foldr`,
by proving it for the initial data,
and the implication that the operation applied to any element of the list preserves the property.
The motive can take values in `Sort _`, so this may be used to construct data,
as well as to prove propositions.
-/
def foldrRecOn {motive : β Sort _} : (l : List α) (op : α β β) (b : β) (_ : motive b)
(_ : (b : β) (_ : motive b) (a : α) (_ : a l), motive (op a b)), motive (List.foldr op b l)
| nil, _, _, hb, _ => hb
| x :: l, op, b, hb, hl =>
hl (foldr op b l)
(foldrRecOn l op b hb fun b c a m => hl b c a (mem_cons_of_mem x m)) x (mem_cons_self x l)
@[simp] theorem foldrRecOn_nil {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a []), motive (op a b)) :
foldrRecOn [] op b hb hl = hb := rfl
@[simp] theorem foldrRecOn_cons {motive : β Sort _} (hb : motive b)
(hl : (b : β) (_ : motive b) (a : α) (_ : a x :: l), motive (op a b)) :
foldrRecOn (x :: l) op b hb hl =
hl _ (foldrRecOn l op b hb fun b c a m => hl b c a (mem_cons_of_mem x m))
x (mem_cons_self x l) :=
rfl
/--
We can prove that two folds over the same list are related (by some arbitrary relation)
if we know that the initial elements are related and the folding function, for each element of the list,
preserves the relation.
-/
theorem foldl_rel {l : List α} {f g : β α β} {a b : β} (r : β β Prop)
(h : r a b) (h' : (a : α), a l (c c' : β), r c c' r (f c a) (g c' a)) :
r (l.foldl (fun acc a => f acc a) a) (l.foldl (fun acc a => g acc a) b) := by
induction l generalizing a b with
| nil => simp_all
| cons a l ih =>
simp only [foldl_cons]
apply ih
· simp_all
· exact fun a m c c' h => h' _ (by simp_all) _ _ h
/--
We can prove that two folds over the same list are related (by some arbitrary relation)
if we know that the initial elements are related and the folding function, for each element of the list,
preserves the relation.
-/
theorem foldr_rel {l : List α} {f g : α β β} {a b : β} (r : β β Prop)
(h : r a b) (h' : (a : α), a l (c c' : β), r c c' r (f a c) (g a c')) :
r (l.foldr (fun a acc => f a acc) a) (l.foldr (fun a acc => g a acc) b) := by
induction l generalizing a b with
| nil => simp_all
| cons a l ih =>
simp only [foldr_cons]
apply h'
· simp
· exact ih h fun a m c c' h => h' _ (by simp_all) _ _ h
@[simp] theorem foldl_add_const (l : List α) (a b : Nat) :
l.foldl (fun x _ => x + a) b = b + a * l.length := by
induction l generalizing b with
| nil => simp
| cons y l ih =>
simp only [foldl_cons, ih, length_cons, Nat.mul_add, Nat.mul_one, Nat.add_assoc,
Nat.add_comm a]
@[simp] theorem foldr_add_const (l : List α) (a b : Nat) :
l.foldr (fun _ x => x + a) b = b + a * l.length := by
induction l generalizing b with
| nil => simp
| cons y l ih =>
simp only [foldr_cons, ih, length_cons, Nat.mul_add, Nat.mul_one, Nat.add_assoc]
/-! #### Further results about `getLast` and `getLast?` -/

View File

@@ -14,8 +14,8 @@ namespace List
@[simp] theorem lex_lt [LT α] (l₁ l₂ : List α) : Lex (· < ·) l₁ l₂ l₁ < l₂ := Iff.rfl
@[simp] theorem not_lex_lt [LT α] (l₁ l₂ : List α) : ¬ Lex (· < ·) l₁ l₂ l₂ l₁ := Iff.rfl
theorem not_lt_iff_ge [LT α] (l₁ l₂ : List α) : ¬ l₁ < l₂ l₂ l₁ := Iff.rfl
theorem not_le_iff_gt [DecidableEq α] [LT α] [DecidableLT α] (l₁ l₂ : List α) :
protected theorem not_lt_iff_ge [LT α] (l₁ l₂ : List α) : ¬ l₁ < l₂ l₂ l₁ := Iff.rfl
protected theorem not_le_iff_gt [DecidableEq α] [LT α] [DecidableLT α] (l₁ l₂ : List α) :
¬ l₁ l₂ l₂ < l₁ :=
Decidable.not_not
@@ -260,7 +260,7 @@ protected theorem le_of_lt [DecidableEq α] [LT α] [DecidableLT α]
· exfalso
exact h' h
theorem le_iff_lt_or_eq [DecidableEq α] [LT α] [DecidableLT α]
protected theorem le_iff_lt_or_eq [DecidableEq α] [LT α] [DecidableLT α]
[Std.Irrefl (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)]
[Std.Total (¬ · < · : α α Prop)]
@@ -333,7 +333,7 @@ theorem lex_eq_true_iff_exists [BEq α] (lt : αα → Bool) :
cases l₂ with
| nil => simp [lex]
| cons b l₂ =>
simp only [lex_cons_cons, Bool.or_eq_true, Bool.and_eq_true, ih, isEqv, length_cons]
simp [lex_cons_cons, Bool.or_eq_true, Bool.and_eq_true, ih, isEqv, length_cons]
constructor
· rintro (hab | hab, h₁, h₂ | i, h₁, h₂, w₁, w₂)
· exact .inr 0, by simp [hab]
@@ -397,7 +397,7 @@ theorem lex_eq_false_iff_exists [BEq α] [PartialEquivBEq α] (lt : αα
cases l₂ with
| nil => simp [lex]
| cons b l₂ =>
simp only [lex_cons_cons, Bool.or_eq_false_iff, Bool.and_eq_false_imp, ih, isEqv,
simp [lex_cons_cons, Bool.or_eq_false_iff, Bool.and_eq_false_imp, ih, isEqv,
Bool.and_eq_true, length_cons]
constructor
· rintro hab, h
@@ -435,7 +435,7 @@ theorem lex_eq_false_iff_exists [BEq α] [PartialEquivBEq α] (lt : αα
simpa using w₁ (j + 1) (by simpa)
· simpa using w₂
theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : List α} :
protected theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : List α} :
l₁ < l₂
(l₁ = l₂.take l₁.length l₁.length < l₂.length)
( (i : Nat) (h₁ : i < l₁.length) (h₂ : i < l₂.length),
@@ -444,7 +444,7 @@ theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : Lis
rw [ lex_eq_true_iff_lt, lex_eq_true_iff_exists]
simp
theorem le_iff_exists [DecidableEq α] [LT α] [DecidableLT α]
protected theorem le_iff_exists [DecidableEq α] [LT α] [DecidableLT α]
[Std.Irrefl (· < · : α α Prop)]
[Std.Asymm (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)] {l₁ l₂ : List α} :
@@ -489,7 +489,7 @@ theorem IsPrefix.le [LT α] [Std.Irrefl (· < · : αα → Prop)]
rcases h with _, rfl
apply le_append_left
theorem map_lt [LT α] [LT β]
protected theorem map_lt [LT α] [LT β]
{l₁ l₂ : List α} {f : α β} (w : x y, x < y f x < f y) (h : l₁ < l₂) :
map f l₁ < map f l₂ := by
match l₁, l₂, h with
@@ -497,11 +497,11 @@ theorem map_lt [LT α] [LT β]
| nil, cons b l₂, h => simp
| cons a l₁, nil, h => simp at h
| cons a l₁, cons _ l₂, .cons h =>
simp [cons_lt_cons_iff, map_lt w (by simpa using h)]
simp [cons_lt_cons_iff, List.map_lt w (by simpa using h)]
| cons a l₁, cons b l₂, .rel h =>
simp [cons_lt_cons_iff, w, h]
theorem map_le [DecidableEq α] [LT α] [DecidableLT α] [DecidableEq β] [LT β] [DecidableLT β]
protected theorem map_le [DecidableEq α] [LT α] [DecidableLT α] [DecidableEq β] [LT β] [DecidableLT β]
[Std.Irrefl (· < · : α α Prop)]
[Std.Asymm (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)]
@@ -510,7 +510,7 @@ theorem map_le [DecidableEq α] [LT α] [DecidableLT α] [DecidableEq β] [LT β
[Std.Antisymm (¬ · < · : β β Prop)]
{l₁ l₂ : List α} {f : α β} (w : x y, x < y f x < f y) (h : l₁ l₂) :
map f l₁ map f l₂ := by
rw [le_iff_exists] at h
rw [List.le_iff_exists] at h
obtain (h | i, h₁, h₂, w₁, w₂) := h
· left
rw [h]

View File

@@ -510,4 +510,18 @@ theorem Perm.eraseP (f : α → Bool) {l₁ l₂ : List α}
refine (IH₁ H).trans (IH₂ ((p₁.pairwise_iff ?_).1 H))
exact fun h h₁ h₂ => h h₂ h₁
theorem perm_insertIdx {α} (x : α) (l : List α) {n} (h : n l.length) :
insertIdx n x l ~ x :: l := by
induction l generalizing n with
| nil =>
cases n with
| zero => rfl
| succ => cases h
| cons _ _ ih =>
cases n with
| zero => simp [insertIdx]
| succ =>
simp only [insertIdx, modifyTailIdx]
refine .trans (.cons _ (ih (Nat.le_of_succ_le_succ h))) (.swap ..)
end List

View File

@@ -253,6 +253,10 @@ theorem merge_perm_append : ∀ {xs ys : List α}, merge xs ys le ~ xs ++ ys
· exact (merge_perm_append.cons y).trans
((Perm.swap x y _).trans (perm_middle.symm.cons x))
theorem Perm.merge (s₁ s₂ : α α Bool) (hl : l₁ ~ l₂) (hr : r₁ ~ r₂) :
merge l₁ r₁ s₁ ~ merge l₂ r₂ s₂ :=
Perm.trans (merge_perm_append ..) <| Perm.trans (Perm.append hl hr) <| Perm.symm (merge_perm_append ..)
/-! ### mergeSort -/
@[simp] theorem mergeSort_nil : [].mergeSort r = [] := by rw [List.mergeSort]

View File

@@ -259,7 +259,7 @@ theorem zip_map (f : αγ) (g : β → δ) :
| [], _ => rfl
| _, [] => by simp only [map, zip_nil_right]
| _ :: _, _ :: _ => by
simp only [map, zip_cons_cons, zip_map, Prod.map]; constructor
simp only [map, zip_cons_cons, zip_map, Prod.map]; try constructor -- TODO: remove try constructor after update stage0
theorem zip_map_left (f : α γ) (l₁ : List α) (l₂ : List β) :
zip (l₁.map f) l₂ = (zip l₁ l₂).map (Prod.map f id) := by rw [ zip_map, map_id]

View File

@@ -136,6 +136,18 @@ This will perform the update destructively provided that the vector has a refere
@[inline] def set! (v : Vector α n) (i : Nat) (x : α) : Vector α n :=
v.toArray.set! i x, by simp
@[inline] def foldlM [Monad m] (f : β α m β) (b : β) (v : Vector α n) : m β :=
v.toArray.foldlM f b
@[inline] def foldrM [Monad m] (f : α β m β) (b : β) (v : Vector α n) : m β :=
v.toArray.foldrM f b
@[inline] def foldl (f : β α β) (b : β) (v : Vector α n) : β :=
v.toArray.foldl f b
@[inline] def foldr (f : α β β) (b : β) (v : Vector α n) : β :=
v.toArray.foldr f b
/-- Append two vectors. -/
@[inline] def append (v : Vector α n) (w : Vector α m) : Vector α (n + m) :=
v.toArray ++ w.toArray, by simp

View File

@@ -66,6 +66,18 @@ theorem toArray_mk (a : Array α) (h : a.size = n) : (Vector.mk a h).toArray = a
@[simp] theorem back?_mk (a : Array α) (h : a.size = n) :
(Vector.mk a h).back? = a.back? := rfl
@[simp] theorem foldlM_mk [Monad m] (f : β α m β) (b : β) (a : Array α) (h : a.size = n) :
(Vector.mk a h).foldlM f b = a.foldlM f b := rfl
@[simp] theorem foldrM_mk [Monad m] (f : α β m β) (b : β) (a : Array α) (h : a.size = n) :
(Vector.mk a h).foldrM f b = a.foldrM f b := rfl
@[simp] theorem foldl_mk (f : β α β) (b : β) (a : Array α) (h : a.size = n) :
(Vector.mk a h).foldl f b = a.foldl f b := rfl
@[simp] theorem foldr_mk (f : α β β) (b : β) (a : Array α) (h : a.size = n) :
(Vector.mk a h).foldr f b = a.foldr f b := rfl
@[simp] theorem drop_mk (a : Array α) (h : a.size = n) (m) :
(Vector.mk a h).drop m = Vector.mk (a.extract m a.size) (by simp [h]) := rfl
@@ -1025,6 +1037,13 @@ theorem mem_setIfInBounds (v : Vector α n) (i : Nat) (hi : i < n) (a : α) :
/-! Content below this point has not yet been aligned with `List` and `Array`. -/
/-! ### map -/
@[simp] theorem getElem_map (f : α β) (a : Vector α n) (i : Nat) (hi : i < n) :
(a.map f)[i] = f a[i] := by
cases a
simp
@[simp] theorem getElem_ofFn {α n} (f : Fin n α) (i : Nat) (h : i < n) :
(Vector.ofFn f)[i] = f i, by simpa using h := by
simp [ofFn]
@@ -1088,13 +1107,6 @@ theorem getElem_append_right {a : Vector α n} {b : Vector α m} {i : Nat} (h :
cases a
simp
/-! ### map -/
@[simp] theorem getElem_map (f : α β) (a : Vector α n) (i : Nat) (hi : i < n) :
(a.map f)[i] = f a[i] := by
cases a
simp
/-! ### zipWith -/
@[simp] theorem getElem_zipWith (f : α β γ) (a : Vector α n) (b : Vector β n) (i : Nat)
@@ -1103,6 +1115,37 @@ theorem getElem_append_right {a : Vector α n} {b : Vector α m} {i : Nat} (h :
cases b
simp
/-! ### foldlM and foldrM -/
@[simp] theorem foldlM_append [Monad m] [LawfulMonad m] (f : β α m β) (b) (l : Vector α n) (l' : Vector α n') :
(l ++ l').foldlM f b = l.foldlM f b >>= l'.foldlM f := by
cases l
cases l'
simp
@[simp] theorem foldrM_push [Monad m] (f : α β m β) (init : β) (l : Vector α n) (a : α) :
(l.push a).foldrM f init = f a init >>= l.foldrM f := by
cases l
simp
theorem foldl_eq_foldlM (f : β α β) (b) (l : Vector α n) :
l.foldl f b = l.foldlM (m := Id) f b := by
cases l
simp [Array.foldl_eq_foldlM]
theorem foldr_eq_foldrM (f : α β β) (b) (l : Vector α n) :
l.foldr f b = l.foldrM (m := Id) f b := by
cases l
simp [Array.foldr_eq_foldrM]
@[simp] theorem id_run_foldlM (f : β α Id β) (b) (l : Vector α n) :
Id.run (l.foldlM f b) = l.foldl f b := (foldl_eq_foldlM f b l).symm
@[simp] theorem id_run_foldrM (f : α β Id β) (b) (l : Vector α n) :
Id.run (l.foldrM f b) = l.foldr f b := (foldr_eq_foldrM f b l).symm
/-! ### foldl and foldr -/
/-! ### take -/
@[simp] theorem take_size (a : Vector α n) : a.take n = a.cast (by simp) := by

View File

@@ -19,6 +19,11 @@ namespace Vector
@[simp] theorem lt_toList [LT α] (l₁ l₂ : Vector α n) : l₁.toList < l₂.toList l₁ < l₂ := Iff.rfl
@[simp] theorem le_toList [LT α] (l₁ l₂ : Vector α n) : l₁.toList l₂.toList l₁ l₂ := Iff.rfl
protected theorem not_lt_iff_ge [LT α] (l₁ l₂ : Vector α n) : ¬ l₁ < l₂ l₂ l₁ := Iff.rfl
protected theorem not_le_iff_gt [DecidableEq α] [LT α] [DecidableLT α] (l₁ l₂ : Vector α n) :
¬ l₁ l₂ l₂ < l₁ :=
Decidable.not_not
@[simp] theorem mk_lt_mk [LT α] :
Vector.mk (α := α) (n := n) data₁ size₁ < Vector.mk data₂ size₂ data₁ < data₂ := Iff.rfl
@@ -133,7 +138,7 @@ protected theorem le_of_lt [DecidableEq α] [LT α] [DecidableLT α]
{l₁ l₂ : Vector α n} (h : l₁ < l₂) : l₁ l₂ :=
Array.le_of_lt h
theorem le_iff_lt_or_eq [DecidableEq α] [LT α] [DecidableLT α]
protected theorem le_iff_lt_or_eq [DecidableEq α] [LT α] [DecidableLT α]
[Std.Irrefl (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)]
[Std.Total (¬ · < · : α α Prop)]
@@ -200,14 +205,14 @@ theorem lex_eq_false_iff_exists [BEq α] [PartialEquivBEq α] (lt : αα
rcases l₂ with l₂, n₂
simp_all [Array.lex_eq_false_iff_exists, n₂]
theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : Vector α n} :
protected theorem lt_iff_exists [DecidableEq α] [LT α] [DecidableLT α] {l₁ l₂ : Vector α n} :
l₁ < l₂
( (i : Nat) (h : i < n), ( j, (hj : j < i) l₁[j] = l₂[j]) l₁[i] < l₂[i]) := by
cases l₁
cases l₂
simp_all [Array.lt_iff_exists]
theorem le_iff_exists [DecidableEq α] [LT α] [DecidableLT α]
protected theorem le_iff_exists [DecidableEq α] [LT α] [DecidableLT α]
[Std.Irrefl (· < · : α α Prop)]
[Std.Asymm (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)] {l₁ l₂ : Vector α n} :
@@ -230,12 +235,12 @@ theorem append_left_le [DecidableEq α] [LT α] [DecidableLT α]
l₁ ++ l₂ l₁ ++ l₃ := by
simpa using Array.append_left_le h
theorem map_lt [LT α] [LT β]
protected theorem map_lt [LT α] [LT β]
{l₁ l₂ : Vector α n} {f : α β} (w : x y, x < y f x < f y) (h : l₁ < l₂) :
map f l₁ < map f l₂ := by
simpa using Array.map_lt w h
theorem map_le [DecidableEq α] [LT α] [DecidableLT α] [DecidableEq β] [LT β] [DecidableLT β]
protected theorem map_le [DecidableEq α] [LT α] [DecidableLT α] [DecidableEq β] [LT β] [DecidableLT β]
[Std.Irrefl (· < · : α α Prop)]
[Std.Asymm (· < · : α α Prop)]
[Std.Antisymm (¬ · < · : α α Prop)]

View File

@@ -8,3 +8,5 @@ import Init.Grind.Norm
import Init.Grind.Tactics
import Init.Grind.Lemmas
import Init.Grind.Cases
import Init.Grind.Propagator
import Init.Grind.Util

View File

@@ -5,10 +5,65 @@ Authors: Leonardo de Moura
-/
prelude
import Init.Core
import Init.SimpLemmas
import Init.Classical
import Init.ByCases
import Init.Grind.Util
namespace Lean.Grind
theorem intro_with_eq (p p' q : Prop) (he : p = p') (h : p' q) : p q :=
fun hp => h (he.mp hp)
/-! And -/
theorem and_eq_of_eq_true_left {a b : Prop} (h : a = True) : (a b) = b := by simp [h]
theorem and_eq_of_eq_true_right {a b : Prop} (h : b = True) : (a b) = a := by simp [h]
theorem and_eq_of_eq_false_left {a b : Prop} (h : a = False) : (a b) = False := by simp [h]
theorem and_eq_of_eq_false_right {a b : Prop} (h : b = False) : (a b) = False := by simp [h]
theorem eq_true_of_and_eq_true_left {a b : Prop} (h : (a b) = True) : a = True := by simp_all
theorem eq_true_of_and_eq_true_right {a b : Prop} (h : (a b) = True) : b = True := by simp_all
/-! Or -/
theorem or_eq_of_eq_true_left {a b : Prop} (h : a = True) : (a b) = True := by simp [h]
theorem or_eq_of_eq_true_right {a b : Prop} (h : b = True) : (a b) = True := by simp [h]
theorem or_eq_of_eq_false_left {a b : Prop} (h : a = False) : (a b) = b := by simp [h]
theorem or_eq_of_eq_false_right {a b : Prop} (h : b = False) : (a b) = a := by simp [h]
theorem eq_false_of_or_eq_false_left {a b : Prop} (h : (a b) = False) : a = False := by simp_all
theorem eq_false_of_or_eq_false_right {a b : Prop} (h : (a b) = False) : b = False := by simp_all
/-! Not -/
theorem not_eq_of_eq_true {a : Prop} (h : a = True) : (Not a) = False := by simp [h]
theorem not_eq_of_eq_false {a : Prop} (h : a = False) : (Not a) = True := by simp [h]
theorem eq_false_of_not_eq_true {a : Prop} (h : (Not a) = True) : a = False := by simp_all
theorem eq_true_of_not_eq_false {a : Prop} (h : (Not a) = False) : a = True := by simp_all
theorem false_of_not_eq_self {a : Prop} (h : (Not a) = a) : False := by
by_cases a <;> simp_all
/-! Eq -/
theorem eq_eq_of_eq_true_left {a b : Prop} (h : a = True) : (a = b) = b := by simp [h]
theorem eq_eq_of_eq_true_right {a b : Prop} (h : b = True) : (a = b) = a := by simp [h]
theorem eq_congr {α : Sort u} {a₁ b₁ a₂ b₂ : α} (h₁ : a₁ = a₂) (h₂ : b₁ = b₂) : (a₁ = b₁) = (a₂ = b₂) := by simp [*]
theorem eq_congr' {α : Sort u} {a₁ b₁ a₂ b₂ : α} (h₁ : a₁ = b₂) (h₂ : b₁ = a₂) : (a₁ = b₁) = (a₂ = b₂) := by rw [h₁, h₂, Eq.comm (a := a₂)]
/-! Forall -/
theorem forall_propagator (p : Prop) (q : p Prop) (q' : Prop) (h₁ : p = True) (h₂ : q (of_eq_true h₁) = q') : ( hp : p, q hp) = q' := by
apply propext; apply Iff.intro
· intro h'; exact Eq.mp h₂ (h' (of_eq_true h₁))
· intro h'; intros; exact Eq.mpr h₂ h'
/-! dite -/
theorem dite_cond_eq_true' {α : Sort u} {c : Prop} {_ : Decidable c} {a : c α} {b : ¬ c α} {r : α} (h₁ : c = True) (h₂ : a (of_eq_true h₁) = r) : (dite c a b) = r := by simp [h₁, h₂]
theorem dite_cond_eq_false' {α : Sort u} {c : Prop} {_ : Decidable c} {a : c α} {b : ¬ c α} {r : α} (h₁ : c = False) (h₂ : b (of_eq_false h₁) = r) : (dite c a b) = r := by simp [h₁, h₂]
end Lean.Grind

View File

@@ -5,6 +5,7 @@ Authors: Leonardo de Moura
-/
prelude
import Init.SimpLemmas
import Init.PropLemmas
import Init.Classical
import Init.ByCases
@@ -41,7 +42,7 @@ attribute [grind_norm] not_true
attribute [grind_norm] not_false_eq_true
-- Implication as a clause
@[grind_norm] theorem imp_eq (p q : Prop) : (p q) = (¬ p q) := by
@[grind_norm] theorem imp_eq (p q : Prop) : (p q) = (¬ p q) := by
by_cases p <;> by_cases q <;> simp [*]
-- And
@@ -58,13 +59,19 @@ attribute [grind_norm] ite_true ite_false
@[grind_norm] theorem not_ite {_ : Decidable p} (q r : Prop) : (¬ite p q r) = ite p (¬q) (¬r) := by
by_cases p <;> simp [*]
@[grind_norm] theorem ite_true_false {_ : Decidable p} : (ite p True False) = p := by
by_cases p <;> simp
@[grind_norm] theorem ite_false_true {_ : Decidable p} : (ite p False True) = ¬p := by
by_cases p <;> simp
-- Forall
@[grind_norm] theorem not_forall (p : α Prop) : (¬ x, p x) = x, ¬p x := by simp
attribute [grind_norm] forall_and
-- Exists
@[grind_norm] theorem not_exists (p : α Prop) : (¬ x, p x) = x, ¬p x := by simp
attribute [grind_norm] exists_const exists_or
attribute [grind_norm] exists_const exists_or exists_prop exists_and_left exists_and_right
-- Bool cond
@[grind_norm] theorem cond_eq_ite (c : Bool) (a b : α) : cond c a b = ite c a b := by
@@ -107,4 +114,7 @@ attribute [grind_norm] Nat.le_zero_eq
-- GT GE
attribute [grind_norm] GT.gt GE.ge
-- Succ
attribute [grind_norm] Nat.succ_eq_add_one
end Lean.Grind

View File

@@ -0,0 +1,27 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.NotationExtra
namespace Lean.Parser
/-- A user-defined propagator for the `grind` tactic. -/
-- TODO: not implemented yet
syntax (docComment)? "grind_propagator " (Tactic.simpPre <|> Tactic.simpPost) ident " (" ident ")" " := " term : command
/-- A builtin propagator for the `grind` tactic. -/
syntax (docComment)? "builtin_grind_propagator " ident (Tactic.simpPre <|> Tactic.simpPost) ident " := " term : command
/-- Auxiliary attribute for builtin `grind` propagators. -/
syntax (name := grindPropagatorBuiltinAttr) "builtin_grind_propagator" (Tactic.simpPre <|> Tactic.simpPost) ident : attr
macro_rules
| `($[$doc?:docComment]? builtin_grind_propagator $propagatorName:ident $direction $op:ident := $body) => do
let propagatorType := `Lean.Meta.Grind.Propagator
`($[$doc?:docComment]? def $propagatorName:ident : $(mkIdent propagatorType) := $body
attribute [builtin_grind_propagator $direction $op] $propagatorName)
end Lean.Parser

View File

@@ -6,20 +6,53 @@ Authors: Leonardo de Moura
prelude
import Init.Tactics
namespace Lean.Parser.Attr
syntax grindEq := "="
syntax grindEqBoth := "_=_"
syntax grindEqRhs := "=_"
syntax grindBwd := ""
syntax grindFwd := ""
syntax (name := grind) "grind" (grindEq <|> grindBwd <|> grindFwd <|> grindEqBoth <|> grindEqRhs)? : attr
end Lean.Parser.Attr
namespace Lean.Grind
/--
The configuration for `grind`.
Passed to `grind` using, for example, the `grind (config := { eager := true })` syntax.
-/
structure Config where
/--
When `eager` is true (default: `false`), `grind` eagerly splits `if-then-else` and `match`
expressions.
-/
/-- When `eager` is true (default: `false`), `grind` eagerly splits `if-then-else` and `match` expressions during internalization. -/
eager : Bool := false
/-- Maximum number of branches (i.e., case-splits) in the proof search tree. -/
splits : Nat := 100
/--
Maximum number of E-matching (aka heuristic theorem instantiation)
in a proof search tree branch.
-/
ematch : Nat := 5
/--
Maximum term generation.
The input goal terms have generation 0. When we instantiate a theorem using a term from generation `n`,
the new terms have generation `n+1`. Thus, this parameter limits the length of an instantiation chain. -/
gen : Nat := 5
/-- Maximum number of theorem instances generated using E-matching in a proof search tree branch. -/
instances : Nat := 1000
/-- If `matchEqs` is `true`, `grind` uses `match`-equations as E-matching theorems. -/
matchEqs : Bool := true
deriving Inhabited, BEq
end Lean.Grind
namespace Lean.Parser.Tactic
/-!
`grind` tactic and related tactics.
-/
end Lean.Grind
-- TODO: parameters
syntax (name := grind) "grind" optConfig ("on_failure " term)? : tactic
end Lean.Parser.Tactic

29
src/Init/Grind/Util.lean Normal file
View File

@@ -0,0 +1,29 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Core
namespace Lean.Grind
/-- A helper gadget for annotating nested proofs in goals. -/
def nestedProof (p : Prop) (h : p) : p := h
/--
Gadget for marking terms that should not be normalized by `grind`s simplifier.
`grind` uses a simproc to implement this feature.
We use it when adding instances of `match`-equations to prevent them from being simplified to true.
-/
def doNotSimp {α : Sort u} (a : α) : α := a
/-- Gadget for representing offsets `t+k` in patterns. -/
def offset (a b : Nat) : Nat := a + b
set_option pp.proofs true
theorem nestedProof_congr (p q : Prop) (h : p = q) (hp : p) (hq : q) : HEq (nestedProof p hp) (nestedProof q hq) := by
subst h; apply HEq.refl
end Lean.Grind

13
src/Init/Internal.lean Normal file
View File

@@ -0,0 +1,13 @@
/-
Copyright (c) 2024 Lean FRO, LLC. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Joachim Breitner
-/
prelude
import Init.Internal.Order
/-!
This directory is used for components of the standard library that are either considered
implementation details or not yet ready for public consumption, and that should be available
without explicit import (in contrast to `Std.Internal`)
-/

View File

@@ -0,0 +1,8 @@
/-
Copyright (c) 2024 Lean FRO, LLC. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Joachim Breitner
-/
prelude
import Init.Internal.Order.Basic
import Init.Internal.Order.Tactic

View File

@@ -0,0 +1,693 @@
/-
Copyright (c) 2024 Lean FRO, LLC. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Joachim Breitner
-/
prelude
import Init.ByCases
import Init.RCases
/-!
This module contains some basic definitions and results from domain theory, intended to be used as
the underlying construction of the `partial_fixpoint` feature. It is not meant to be used as a
general purpose library for domain theory, but can be of interest to users who want to extend
the `partial_fixpoint` machinery (e.g. mark more functions as monotone or register more monads).
This follows the corresponding
[Isabelle development](https://isabelle.in.tum.de/library/HOL/HOL/Partial_Function.html), as also
described in [Alexander Krauss: Recursive Definitions of Monadic Functions](https://www21.in.tum.de/~krauss/papers/mrec.pdf).
-/
universe u v w
namespace Lean.Order
/--
A partial order is a reflexive, transitive and antisymmetric relation.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
class PartialOrder (α : Sort u) where
/--
A “less-or-equal-to” or “approximates” relation.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
rel : α α Prop
rel_refl : {x}, rel x x
rel_trans : {x y z}, rel x y rel y z rel x z
rel_antisymm : {x y}, rel x y rel y x x = y
@[inherit_doc] scoped infix:50 "" => PartialOrder.rel
section PartialOrder
variable {α : Sort u} [PartialOrder α]
theorem PartialOrder.rel_of_eq {x y : α} (h : x = y) : x y := by cases h; apply rel_refl
/--
A chain is a totally ordered set (representing a set as a predicate).
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
def chain (c : α Prop) : Prop := x y , c x c y x y y x
end PartialOrder
section CCPO
/--
A chain-complete partial order (CCPO) is a partial order where every chain a least upper bound.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
class CCPO (α : Sort u) extends PartialOrder α where
/--
The least upper bound of a chain.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
csup : (α Prop) α
csup_spec {c : α Prop} (hc : chain c) : csup c x ( y, c y y x)
open PartialOrder CCPO
variable {α : Sort u} [CCPO α]
theorem csup_le {c : α Prop} (hchain : chain c) : ( y, c y y x) csup c x :=
(csup_spec hchain).mpr
theorem le_csup {c : α Prop} (hchain : chain c) {y : α} (hy : c y) : y csup c :=
(csup_spec hchain).mp rel_refl y hy
/--
The bottom element is the least upper bound of the empty chain.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
def bot : α := csup (fun _ => False)
scoped notation "" => bot
theorem bot_le (x : α) : x := by
apply csup_le
· intro x y hx hy; contradiction
· intro x hx; contradiction
end CCPO
section monotone
variable {α : Sort u} [PartialOrder α]
variable {β : Sort v} [PartialOrder β]
/--
A function is monotone if if it maps related elements to releated elements.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
def monotone (f : α β) : Prop := x y, x y f x f y
theorem monotone_const (c : β) : monotone (fun (_ : α) => c) :=
fun _ _ _ => PartialOrder.rel_refl
theorem monotone_id : monotone (fun (x : α) => x) :=
fun _ _ hxy => hxy
theorem monotone_compose
{γ : Sort w} [PartialOrder γ]
{f : α β} {g : β γ}
(hf : monotone f) (hg : monotone g) :
monotone (fun x => g (f x)) := fun _ _ hxy => hg _ _ (hf _ _ hxy)
end monotone
section admissibility
variable {α : Sort u} [CCPO α]
open PartialOrder CCPO
/--
A predicate is admissable if it can be transferred from the elements of a chain to the chains least
upper bound. Such predicates can be used in fixpoint induction.
This definition implies `P ⊥`. Sometimes (e.g. in Isabelle) the empty chain is excluded
from this definition, and `P ⊥` is a separate condition of the induction predicate.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
def admissible (P : α Prop) :=
(c : α Prop), chain c ( x, c x P x) P (csup c)
theorem admissible_const_true : admissible (fun (_ : α) => True) :=
fun _ _ _ => trivial
theorem admissible_and (P Q : α Prop)
(hadm₁ : admissible P) (hadm₂ : admissible Q) : admissible (fun x => P x Q x) :=
fun c hchain h =>
hadm₁ c hchain fun x hx => (h x hx).1,
hadm₂ c hchain fun x hx => (h x hx).2
theorem chain_conj (c P : α Prop) (hchain : chain c) : chain (fun x => c x P x) := by
intro x y hcx, _ hcy, _
exact hchain x y hcx hcy
theorem csup_conj (c P : α Prop) (hchain : chain c) (h : x, c x y, c y x y P y) :
csup c = csup (fun x => c x P x) := by
apply rel_antisymm
· apply csup_le hchain
intro x hcx
obtain y, hcy, hxy, hPy := h x hcx
apply rel_trans hxy; clear x hcx hxy
apply le_csup (chain_conj _ _ hchain) hcy, hPy
· apply csup_le (chain_conj _ _ hchain)
intro x hcx, hPx
apply le_csup hchain hcx
theorem admissible_or (P Q : α Prop)
(hadm₁ : admissible P) (hadm₂ : admissible Q) : admissible (fun x => P x Q x) := by
intro c hchain h
have : ( x, c x y, c y x y P y) ( x, c x y, c y x y Q y) := by
open Classical in
apply Decidable.or_iff_not_imp_left.mpr
intro h'
simp only [not_forall, not_imp, not_exists, not_and] at h'
obtain x, hcx, hx := h'
intro y hcy
cases hchain x y hcx hcy with
| inl hxy =>
refine y, hcy, rel_refl, ?_
cases h y hcy with
| inl hPy => exfalso; apply hx y hcy hxy hPy
| inr hQy => assumption
| inr hyx =>
refine x, hcx, hyx , ?_
cases h x hcx with
| inl hPx => exfalso; apply hx x hcx rel_refl hPx
| inr hQx => assumption
cases this with
| inl hP =>
left
rw [csup_conj (h := hP) (hchain := hchain)]
apply hadm₁ _ (chain_conj _ _ hchain)
intro x hcx, hPx
exact hPx
| inr hQ =>
right
rw [csup_conj (h := hQ) (hchain := hchain)]
apply hadm₂ _ (chain_conj _ _ hchain)
intro x hcx, hQx
exact hQx
def admissible_pi (P : α β Prop)
(hadm₁ : y, admissible (fun x => P x y)) : admissible (fun x => y, P x y) :=
fun c hchain h y => hadm₁ y c hchain fun x hx => h x hx y
end admissibility
section fix
open PartialOrder CCPO
variable {α : Sort u} [CCPO α]
variable {c : α Prop} (hchain : chain c)
/--
The transfinite iteration of a function `f` is a set that is `⊥ ` and is closed under application
of `f` and `csup`.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
inductive iterates (f : α α) : α Prop where
| step : iterates f x iterates f (f x)
| sup {c : α Prop} (hc : chain c) (hi : x, c x iterates f x) : iterates f (csup c)
theorem chain_iterates {f : α α} (hf : monotone f) : chain (iterates f) := by
intros x y hx hy
induction hx generalizing y
case step x hx ih =>
induction hy
case step y hy _ =>
cases ih y hy
· left; apply hf; assumption
· right; apply hf; assumption
case sup c hchain hi ih2 =>
show f x csup c csup c f x
by_cases h : z, c z f x z
· left
obtain z, hz, hfz := h
apply rel_trans hfz
apply le_csup hchain hz
· right
apply csup_le hchain _
intro z hz
rw [not_exists] at h
specialize h z
rw [not_and] at h
specialize h hz
cases ih2 z hz
next => contradiction
next => assumption
case sup c hchain hi ih =>
show rel (csup c) y rel y (csup c)
by_cases h : z, c z rel y z
· right
obtain z, hz, hfz := h
apply rel_trans hfz
apply le_csup hchain hz
· left
apply csup_le hchain _
intro z hz
rw [not_exists] at h
specialize h z
rw [not_and] at h
specialize h hz
cases ih z hz y hy
next => assumption
next => contradiction
theorem rel_f_of_iterates {f : α α} (hf : monotone f) {x : α} (hx : iterates f x) : x f x := by
induction hx
case step ih =>
apply hf
assumption
case sup c hchain hi ih =>
apply csup_le hchain
intro y hy
apply rel_trans (ih y hy)
apply hf
apply le_csup hchain hy
set_option linter.unusedVariables false in
/--
The least fixpoint of a monotone function is the least upper bound of its transfinite iteration.
The `monotone f` assumption is not strictly necessarily for the definition, but without this the
definition is not very meaningful and it simplifies applying theorems like `fix_eq` if every use of
`fix` already has the monotonicty requirement.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
def fix (f : α α) (hmono : monotone f) := csup (iterates f)
/--
The main fixpoint theorem for fixedpoints of monotone functions in chain-complete partial orders.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
theorem fix_eq {f : α α} (hf : monotone f) : fix f hf = f (fix f hf) := by
apply rel_antisymm
· apply rel_f_of_iterates hf
apply iterates.sup (chain_iterates hf)
exact fun _ h => h
· apply le_csup (chain_iterates hf)
apply iterates.step
apply iterates.sup (chain_iterates hf)
intro y hy
exact hy
/--
The fixpoint induction theme: An admissible predicate holds for a least fixpoint if it is preserved
by the fixpoint's function.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
theorem fix_induct {f : α α} (hf : monotone f)
(motive : α Prop) (hadm: admissible motive)
(h : x, motive x motive (f x)) : motive (fix f hf) := by
apply hadm _ (chain_iterates hf)
intro x hiterates
induction hiterates with
| @step x hiter ih => apply h x ih
| @sup c hchain hiter ih => apply hadm c hchain ih
end fix
section fun_order
open PartialOrder
variable {α : Sort u}
variable {β : α Sort v}
variable {γ : Sort w}
instance instOrderPi [ x, PartialOrder (β x)] : PartialOrder ( x, β x) where
rel f g := x, f x g x
rel_refl _ := rel_refl
rel_trans hf hg x := rel_trans (hf x) (hg x)
rel_antisymm hf hg := funext (fun x => rel_antisymm (hf x) (hg x))
theorem monotone_of_monotone_apply [PartialOrder γ] [ x, PartialOrder (β x)] (f : γ ( x, β x))
(h : y, monotone (fun x => f x y)) : monotone f :=
fun x y hxy z => h z x y hxy
theorem monotone_apply [PartialOrder γ] [ x, PartialOrder (β x)] (a : α) (f : γ x, β x)
(h : monotone f) :
monotone (fun x => f x a) := fun _ _ hfg => h _ _ hfg a
theorem chain_apply [ x, PartialOrder (β x)] {c : ( x, β x) Prop} (hc : chain c) (x : α) :
chain (fun y => f, c f f x = y) := by
intro _ _ f, hf, hfeq g, hg, hgeq
subst hfeq; subst hgeq
cases hc f g hf hg
next h => left; apply h x
next h => right; apply h x
def fun_csup [ x, CCPO (β x)] (c : ( x, β x) Prop) (x : α) :=
CCPO.csup (fun y => f, c f f x = y)
instance instCCPOPi [ x, CCPO (β x)] : CCPO ( x, β x) where
csup := fun_csup
csup_spec := by
intro f c hc
constructor
next =>
intro hf g hg x
apply rel_trans _ (hf x); clear hf
apply le_csup (chain_apply hc x)
exact g, hg, rfl
next =>
intro h x
apply csup_le (chain_apply hc x)
intro y z, hz, hyz
subst y
apply h z hz
def admissible_apply [ x, CCPO (β x)] (P : x, β x Prop) (x : α)
(hadm : admissible (P x)) : admissible (fun (f : x, β x) => P x (f x)) := by
intro c hchain h
apply hadm _ (chain_apply hchain x)
rintro _ f, hcf, rfl
apply h _ hcf
def admissible_pi_apply [ x, CCPO (β x)] (P : x, β x Prop) (hadm : x, admissible (P x)) :
admissible (fun (f : x, β x) => x, P x (f x)) := by
apply admissible_pi
intro
apply admissible_apply
apply hadm
end fun_order
section monotone_lemmas
theorem monotone_letFun
{α : Sort u} {β : Sort v} {γ : Sort w} [PartialOrder α] [PartialOrder β]
(v : γ) (k : α γ β)
(hmono : y, monotone (fun x => k x y)) :
monotone fun (x : α) => letFun v (k x) := hmono v
theorem monotone_ite
{α : Sort u} {β : Sort v} [PartialOrder α] [PartialOrder β]
(c : Prop) [Decidable c]
(k₁ : α β) (k₂ : α β)
(hmono₁ : monotone k₁) (hmono₂ : monotone k₂) :
monotone fun x => if c then k₁ x else k₂ x := by
split
· apply hmono₁
· apply hmono₂
theorem monotone_dite
{α : Sort u} {β : Sort v} [PartialOrder α] [PartialOrder β]
(c : Prop) [Decidable c]
(k₁ : α c β) (k₂ : α ¬ c β)
(hmono₁ : monotone k₁) (hmono₂ : monotone k₂) :
monotone fun x => dite c (k₁ x) (k₂ x) := by
split
· apply monotone_apply _ _ hmono₁
· apply monotone_apply _ _ hmono₂
end monotone_lemmas
section pprod_order
open PartialOrder
variable {α : Sort u}
variable {β : Sort v}
variable {γ : Sort w}
instance [PartialOrder α] [PartialOrder β] : PartialOrder (α ×' β) where
rel a b := a.1 b.1 a.2 b.2
rel_refl := rel_refl, rel_refl
rel_trans ha hb := rel_trans ha.1 hb.1, rel_trans ha.2 hb.2
rel_antisymm := fun {a} {b} ha hb => by
cases a; cases b;
dsimp at *
rw [rel_antisymm ha.1 hb.1, rel_antisymm ha.2 hb.2]
theorem monotone_pprod [PartialOrder α] [PartialOrder β] [PartialOrder γ]
{f : γ α} {g : γ β} (hf : monotone f) (hg : monotone g) :
monotone (fun x => PProd.mk (f x) (g x)) :=
fun _ _ h12 => hf _ _ h12, hg _ _ h12
theorem monotone_pprod_fst [PartialOrder α] [PartialOrder β] [PartialOrder γ]
{f : γ α ×' β} (hf : monotone f) : monotone (fun x => (f x).1) :=
fun _ _ h12 => (hf _ _ h12).1
theorem monotone_pprod_snd [PartialOrder α] [PartialOrder β] [PartialOrder γ]
{f : γ α ×' β} (hf : monotone f) : monotone (fun x => (f x).2) :=
fun _ _ h12 => (hf _ _ h12).2
def chain_pprod_fst [CCPO α] [CCPO β] (c : α ×' β Prop) : α Prop := fun a => b, c a, b
def chain_pprod_snd [CCPO α] [CCPO β] (c : α ×' β Prop) : β Prop := fun b => a, c a, b
theorem chain.pprod_fst [CCPO α] [CCPO β] (c : α ×' β Prop) (hchain : chain c) :
chain (chain_pprod_fst c) := by
intro a₁ a₂ b₁, h₁ b₂, h₂
cases hchain a₁, b₁ a₂, b₂ h₁ h₂
case inl h => left; exact h.1
case inr h => right; exact h.1
theorem chain.pprod_snd [CCPO α] [CCPO β] (c : α ×' β Prop) (hchain : chain c) :
chain (chain_pprod_snd c) := by
intro b₁ b₂ a₁, h₁ a₂, h₂
cases hchain a₁, b₁ a₂, b₂ h₁ h₂
case inl h => left; exact h.2
case inr h => right; exact h.2
instance [CCPO α] [CCPO β] : CCPO (α ×' β) where
csup c := CCPO.csup (chain_pprod_fst c), CCPO.csup (chain_pprod_snd c)
csup_spec := by
intro a, b c hchain
dsimp
constructor
next =>
intro h₁, h₂ a', b' cab
constructor <;> dsimp at *
· apply rel_trans ?_ h₁
apply le_csup hchain.pprod_fst
exact b', cab
· apply rel_trans ?_ h₂
apply le_csup hchain.pprod_snd
exact a', cab
next =>
intro h
constructor <;> dsimp
· apply csup_le hchain.pprod_fst
intro a' b', hcab
apply (h _ hcab).1
· apply csup_le hchain.pprod_snd
intro b' a', hcab
apply (h _ hcab).2
theorem admissible_pprod_fst {α : Sort u} {β : Sort v} [CCPO α] [CCPO β] (P : α Prop)
(hadm : admissible P) : admissible (fun (x : α ×' β) => P x.1) := by
intro c hchain h
apply hadm _ hchain.pprod_fst
intro x y, hxy
apply h x,y hxy
theorem admissible_pprod_snd {α : Sort u} {β : Sort v} [CCPO α] [CCPO β] (P : β Prop)
(hadm : admissible P) : admissible (fun (x : α ×' β) => P x.2) := by
intro c hchain h
apply hadm _ hchain.pprod_snd
intro y x, hxy
apply h x,y hxy
end pprod_order
section flat_order
variable {α : Sort u}
set_option linter.unusedVariables false in
/--
`FlatOrder b` wraps the type `α` with the flat partial order generated by `∀ x, b ⊑ x`.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
def FlatOrder {α : Sort u} (b : α) := α
variable {b : α}
/--
The flat partial order generated by `∀ x, b ⊑ x`.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
inductive FlatOrder.rel : (x y : FlatOrder b) Prop where
| bot : rel b x
| refl : rel x x
instance FlatOrder.instOrder : PartialOrder (FlatOrder b) where
rel := rel
rel_refl := .refl
rel_trans {x y z : α} (hxy : rel x y) (hyz : rel y z) := by
cases hxy <;> cases hyz <;> constructor
rel_antisymm {x y : α} (hxy : rel x y) (hyz : rel y x) : x = y := by
cases hxy <;> cases hyz <;> constructor
open Classical in
private theorem Classical.some_spec₂ {α : Sort _} {p : α Prop} {h : a, p a} (q : α Prop)
(hpq : a, p a q a) : q (choose h) := hpq _ <| choose_spec _
noncomputable def flat_csup (c : FlatOrder b Prop) : FlatOrder b := by
by_cases h : (x : FlatOrder b), c x x b
· exact Classical.choose h
· exact b
noncomputable instance FlatOrder.instCCPO : CCPO (FlatOrder b) where
csup := flat_csup
csup_spec := by
intro x c hc
unfold flat_csup
split
next hex =>
apply Classical.some_spec₂ (q := (· x ( y, c y y x)))
clear hex
intro z hz, hnb
constructor
· intro h y hy
apply PartialOrder.rel_trans _ h; clear h
cases hc y z hy hz
next => assumption
next h =>
cases h
· contradiction
· constructor
· intro h
cases h z hz
· contradiction
· constructor
next hnotex =>
constructor
· intro h y hy; clear h
suffices y = b by rw [this]; exact rel.bot
rw [not_exists] at hnotex
specialize hnotex y
rw [not_and] at hnotex
specialize hnotex hy
rw [@Classical.not_not] at hnotex
assumption
· intro; exact rel.bot
theorem admissible_flatOrder (P : FlatOrder b Prop) (hnot : P b) : admissible P := by
intro c hchain h
by_cases h' : (x : FlatOrder b), c x x b
· simp [CCPO.csup, flat_csup, h']
apply Classical.some_spec₂ (q := (P ·))
intro x hcx, hneb
apply h x hcx
· simp [CCPO.csup, flat_csup, h', hnot]
end flat_order
section mono_bind
/--
The class `MonoBind m` indicates that every `m α` has a `PartialOrder`, and that the bind operation
on `m` is monotone in both arguments with regard to that order.
This is intended to be used in the construction of `partial_fixpoint`, and not meant to be used otherwise.
-/
class MonoBind (m : Type u Type v) [Bind m] [ α, PartialOrder (m α)] where
bind_mono_left {a₁ a₂ : m α} {f : α m b} (h : a₁ a₂) : a₁ >>= f a₂ >>= f
bind_mono_right {a : m α} {f₁ f₂ : α m b} (h : x, f₁ x f₂ x) : a >>= f₁ a >>= f₂
theorem monotone_bind
(m : Type u Type v) [Bind m] [ α, PartialOrder (m α)] [MonoBind m]
{α β : Type u}
{γ : Type w} [PartialOrder γ]
(f : γ m α) (g : γ α m β)
(hmono₁ : monotone f)
(hmono₂ : monotone g) :
monotone (fun (x : γ) => f x >>= g x) := by
intro x₁ x₂ hx₁₂
apply PartialOrder.rel_trans
· apply MonoBind.bind_mono_left (hmono₁ _ _ hx₁₂)
· apply MonoBind.bind_mono_right (fun y => monotone_apply y _ hmono₂ _ _ hx₁₂)
instance : PartialOrder (Option α) := inferInstanceAs (PartialOrder (FlatOrder none))
noncomputable instance : CCPO (Option α) := inferInstanceAs (CCPO (FlatOrder none))
noncomputable instance : MonoBind Option where
bind_mono_left h := by
cases h
· exact FlatOrder.rel.bot
· exact FlatOrder.rel.refl
bind_mono_right h := by
cases Option _
· exact FlatOrder.rel.refl
· exact h _
theorem admissible_eq_some (P : Prop) (y : α) :
admissible (fun (x : Option α) => x = some y P) := by
apply admissible_flatOrder; simp
instance [Monad m] [inst : α, PartialOrder (m α)] : PartialOrder (ExceptT ε m α) := inst _
instance [Monad m] [ α, PartialOrder (m α)] [inst : α, CCPO (m α)] : CCPO (ExceptT ε m α) := inst _
instance [Monad m] [ α, PartialOrder (m α)] [ α, CCPO (m α)] [MonoBind m] : MonoBind (ExceptT ε m) where
bind_mono_left h₁₂ := by
apply MonoBind.bind_mono_left (m := m)
exact h₁₂
bind_mono_right h₁₂ := by
apply MonoBind.bind_mono_right (m := m)
intro x
cases x
· apply PartialOrder.rel_refl
· apply h₁₂
end mono_bind
namespace Example
def findF (P : Nat Bool) (rec : Nat Option Nat) (x : Nat) : Option Nat :=
if P x then
some x
else
rec (x + 1)
noncomputable def find (P : Nat Bool) : Nat Option Nat := fix (findF P) <| by
unfold findF
apply monotone_of_monotone_apply
intro n
split
· apply monotone_const
· apply monotone_apply
apply monotone_id
theorem find_eq : find P = findF P (find P) := fix_eq ..
theorem find_spec : n m, find P n = some m n m P m := by
unfold find
refine fix_induct (motive := fun (f : Nat Option Nat) => n m, f n = some m n m P m) _ ?hadm ?hstep
case hadm =>
-- apply admissible_pi_apply does not work well, hard to infer everything
exact admissible_pi_apply _ (fun n => admissible_pi _ (fun m => admissible_eq_some _ m))
case hstep =>
intro f ih n m heq
simp only [findF] at heq
split at heq
· simp_all
· obtain ih1, ih2 := ih _ _ heq
constructor
· exact Nat.le_trans (Nat.le_add_right _ _ ) ih1
· exact ih2
end Example
end Lean.Order

View File

@@ -0,0 +1,20 @@
/-
Copyright (c) 2024 Lean FRO, LLC. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Joachim Breitner
-/
prelude
import Init.Notation
namespace Lean.Order
/--
`monotonicity` performs one compositional step solving `monotone` goals,
using lemma tagged with `@[partial_fixpoint_monotone]`.
This tactic is mostly used internally by lean in `partial_fixpoint` definitions, but
can be useful on its own for debugging or when proving new `@[partial_fixpoint_monotone]` lemmas.
-/
scoped syntax (name := monotonicity) "monotonicity" : tactic
end Lean.Order

View File

@@ -18,6 +18,7 @@ inductive ExternEntry where
| inline (backend : Name) (pattern : String)
| standard (backend : Name) (fn : String)
| foreign (backend : Name) (fn : String)
deriving BEq, Hashable
/--
- `@[extern]`
@@ -36,7 +37,7 @@ inductive ExternEntry where
structure ExternAttrData where
arity? : Option Nat := none
entries : List ExternEntry
deriving Inhabited
deriving Inhabited, BEq, Hashable
-- def externEntry := leading_parser optional ident >> optional (nonReservedSymbol "inline ") >> strLit
-- @[builtin_attr_parser] def extern := leading_parser nonReservedSymbol "extern " >> optional numLit >> many externEntry

View File

@@ -7,6 +7,7 @@ prelude
import Init.Data.List.BasicAux
import Lean.Expr
import Lean.Meta.Instances
import Lean.Compiler.ExternAttr
import Lean.Compiler.InlineAttrs
import Lean.Compiler.Specialize
import Lean.Compiler.LCNF.Types
@@ -429,6 +430,80 @@ where
| .cases c => c.alts.forM fun alt => go alt.getCode
| .unreach .. | .return .. | .jmp .. => return ()
partial def Code.instantiateValueLevelParams (code : Code) (levelParams : List Name) (us : List Level) : Code :=
instCode code
where
instLevel (u : Level) :=
u.instantiateParams levelParams us
instExpr (e : Expr) :=
e.instantiateLevelParamsNoCache levelParams us
instParams (ps : Array Param) :=
ps.mapMono fun p => p.updateCore (instExpr p.type)
instAlt (alt : Alt) :=
match alt with
| .default k => alt.updateCode (instCode k)
| .alt _ ps k => alt.updateAlt! (instParams ps) (instCode k)
instArg (arg : Arg) : Arg :=
match arg with
| .type e => arg.updateType! (instExpr e)
| .fvar .. | .erased => arg
instLetValue (e : LetValue) : LetValue :=
match e with
| .const declName vs args => e.updateConst! declName (vs.mapMono instLevel) (args.mapMono instArg)
| .fvar fvarId args => e.updateFVar! fvarId (args.mapMono instArg)
| .proj .. | .value .. | .erased => e
instLetDecl (decl : LetDecl) :=
decl.updateCore (instExpr decl.type) (instLetValue decl.value)
instFunDecl (decl : FunDecl) :=
decl.updateCore (instExpr decl.type) (instParams decl.params) (instCode decl.value)
instCode (code : Code) :=
match code with
| .let decl k => code.updateLet! (instLetDecl decl) (instCode k)
| .jp decl k | .fun decl k => code.updateFun! (instFunDecl decl) (instCode k)
| .cases c => code.updateCases! (instExpr c.resultType) c.discr (c.alts.mapMono instAlt)
| .jmp fvarId args => code.updateJmp! fvarId (args.mapMono instArg)
| .return .. => code
| .unreach type => code.updateUnreach! (instExpr type)
inductive DeclValue where
| code (code : Code)
| extern (externAttrData : ExternAttrData)
deriving Inhabited, BEq
partial def DeclValue.size : DeclValue Nat
| .code c => c.size
| .extern .. => 0
def DeclValue.mapCode (f : Code Code) : DeclValue DeclValue :=
fun
| .code c => .code (f c)
| .extern e => .extern e
def DeclValue.mapCodeM [Monad m] (f : Code m Code) : DeclValue m DeclValue :=
fun v => do
match v with
| .code c => return .code ( f c)
| .extern .. => return v
def DeclValue.forCodeM [Monad m] (f : Code m Unit) : DeclValue m Unit :=
fun v => do
match v with
| .code c => f c
| .extern .. => return ()
def DeclValue.isCodeAndM [Monad m] (v : DeclValue) (f : Code m Bool) : m Bool :=
match v with
| .code c => f c
| .extern .. => pure false
/--
Declaration being processed by the Lean to Lean compiler passes.
-/
@@ -455,7 +530,7 @@ structure Decl where
The body of the declaration, usually changes as it progresses
through compiler passes.
-/
value : Code
value : DeclValue
/--
We set this flag to true during LCNF conversion. When we receive
a block of functions to be compiled, we set this flag to `true`
@@ -536,7 +611,9 @@ We use this function to decide whether we should inline a declaration tagged wit
`[inline_if_reduce]` or not.
-/
def Decl.isCasesOnParam? (decl : Decl) : Option Nat :=
go decl.value
match decl.value with
| .code c => go c
| .extern .. => none
where
go (code : Code) : Option Nat :=
match code with
@@ -550,49 +627,6 @@ def Decl.instantiateTypeLevelParams (decl : Decl) (us : List Level) : Expr :=
def Decl.instantiateParamsLevelParams (decl : Decl) (us : List Level) : Array Param :=
decl.params.mapMono fun param => param.updateCore (param.type.instantiateLevelParamsNoCache decl.levelParams us)
partial def Decl.instantiateValueLevelParams (decl : Decl) (us : List Level) : Code :=
instCode decl.value
where
instLevel (u : Level) :=
u.instantiateParams decl.levelParams us
instExpr (e : Expr) :=
e.instantiateLevelParamsNoCache decl.levelParams us
instParams (ps : Array Param) :=
ps.mapMono fun p => p.updateCore (instExpr p.type)
instAlt (alt : Alt) :=
match alt with
| .default k => alt.updateCode (instCode k)
| .alt _ ps k => alt.updateAlt! (instParams ps) (instCode k)
instArg (arg : Arg) : Arg :=
match arg with
| .type e => arg.updateType! (instExpr e)
| .fvar .. | .erased => arg
instLetValue (e : LetValue) : LetValue :=
match e with
| .const declName vs args => e.updateConst! declName (vs.mapMono instLevel) (args.mapMono instArg)
| .fvar fvarId args => e.updateFVar! fvarId (args.mapMono instArg)
| .proj .. | .value .. | .erased => e
instLetDecl (decl : LetDecl) :=
decl.updateCore (instExpr decl.type) (instLetValue decl.value)
instFunDecl (decl : FunDecl) :=
decl.updateCore (instExpr decl.type) (instParams decl.params) (instCode decl.value)
instCode (code : Code) :=
match code with
| .let decl k => code.updateLet! (instLetDecl decl) (instCode k)
| .jp decl k | .fun decl k => code.updateFun! (instFunDecl decl) (instCode k)
| .cases c => code.updateCases! (instExpr c.resultType) c.discr (c.alts.mapMono instAlt)
| .jmp fvarId args => code.updateJmp! fvarId (args.mapMono instArg)
| .return .. => code
| .unreach type => code.updateUnreach! (instExpr type)
/--
Return `true` if the arrow type contains an instance implicit argument.
-/
@@ -693,7 +727,7 @@ where
visit k
go : StateM NameSet Unit :=
decls.forM fun decl => visit decl.value
decls.forM (·.value.forCodeM visit)
def instantiateRangeArgs (e : Expr) (beginIdx endIdx : Nat) (args : Array Arg) : Expr :=
if !e.hasLooseBVars then

View File

@@ -123,7 +123,10 @@ def FunDeclCore.etaExpand (decl : FunDecl) : CompilerM FunDecl := do
decl.update decl.type params value
def Decl.etaExpand (decl : Decl) : CompilerM Decl := do
let some (params, value) etaExpandCore? decl.type decl.params decl.value | return decl
return { decl with params, value }
match decl.value with
| .code code =>
let some (params, newCode) etaExpandCore? decl.type decl.params code | return decl
return { decl with params, value := .code newCode}
| .extern .. => return decl
end Lean.Compiler.LCNF

View File

@@ -102,7 +102,7 @@ end CSE
Common sub-expression elimination
-/
def Decl.cse (decl : Decl) : CompilerM Decl := do
let value decl.value.cse
let value decl.value.mapCodeM (·.cse)
return { decl with value }
def cse (phase : Phase := .base) (occurrence := 0) : Pass :=

View File

@@ -261,7 +261,7 @@ def run (x : CheckM α) : CompilerM α :=
end Check
def Decl.check (decl : Decl) : CompilerM Unit := do
Check.run do Check.checkFunDeclCore decl.name decl.params decl.type decl.value
Check.run do decl.value.forCodeM (Check.checkFunDeclCore decl.name decl.params decl.type)
/--
Check whether every local declaration in the local context is used in one of given `decls`.
@@ -299,7 +299,7 @@ where
visitDecl (decl : Decl) : StateM FVarIdHashSet Unit := do
visitParams decl.params
visitCode decl.value
decl.value.forCodeM visitCode
visitDecls (decls : Array Decl) : StateM FVarIdHashSet Unit :=
decls.forM visitDecl

View File

@@ -148,7 +148,7 @@ def eraseCodeDecls (decls : Array CodeDecl) : CompilerM Unit := do
def eraseDecl (decl : Decl) : CompilerM Unit := do
eraseParams decl.params
eraseCode decl.value
decl.value.forCodeM eraseCode
abbrev Decl.erase (decl : Decl) : CompilerM Unit :=
eraseDecl decl

View File

@@ -38,6 +38,7 @@ end
instance : Hashable Code where
hash c := hashCode c
deriving instance Hashable for DeclValue
deriving instance Hashable for Decl
end Lean.Compiler.LCNF
end Lean.Compiler.LCNF

View File

@@ -95,6 +95,6 @@ def Code.elimDead (code : Code) : CompilerM Code :=
ElimDead.elimDead code |>.run' {}
def Decl.elimDead (decl : Decl) : CompilerM Decl := do
return { decl with value := ( decl.value.elimDead) }
return { decl with value := ( decl.value.mapCodeM Code.elimDead) }
end Lean.Compiler.LCNF

View File

@@ -513,7 +513,7 @@ def inferStep : InterpM Bool := do
let currentVal getFunVal idx
withReader (fun ctx => { ctx with currFnIdx := idx }) do
decl.params.forM fun p => updateVarAssignment p.fvarId .top
interpCode decl.value
decl.value.forCodeM interpCode
let newVal getFunVal idx
if currentVal != newVal then
return true
@@ -538,7 +538,7 @@ Use the information produced by the abstract interpreter to:
-/
partial def elimDead (assignment : Assignment) (decl : Decl) : CompilerM Decl := do
trace[Compiler.elimDeadBranches] s!"Eliminating {decl.name} with {repr (← assignment.toArray |>.mapM (fun (name, val) => do return (toString (← getBinderName name), val)))}"
return { decl with value := ( go decl.value) }
return { decl with value := ( decl.value.mapCodeM go) }
where
go (code : Code) : CompilerM Code := do
match code with

View File

@@ -141,8 +141,9 @@ partial def evalApp (declName : Name) (args : Array Arg) : FixParamM Unit := do
let key := (declName, values)
unless ( get).visited.contains key do
modify fun s => { s with visited := s.visited.insert key }
let assignment := mkAssignment decl values
withReader (fun ctx => { ctx with assignment }) <| evalCode decl.value
decl.value.forCodeM fun c =>
let assignment := mkAssignment decl values
withReader (fun ctx => { ctx with assignment }) <| evalCode c
end
@@ -169,8 +170,12 @@ def mkFixedParamsMap (decls : Array Decl) : NameMap (Array Bool) := Id.run do
let values := mkInitialValues decl.params.size
let assignment := mkAssignment decl values
let fixed := Array.mkArray decl.params.size true
match evalCode decl.value |>.run { main := decl, decls, assignment } |>.run { fixed } with
| .ok _ s | .error _ s => result := result.insert decl.name s.fixed
match decl.value with
| .code c =>
match evalCode c |>.run { main := decl, decls, assignment } |>.run { fixed } with
| .ok _ s | .error _ s => result := result.insert decl.name s.fixed
| .extern .. =>
result := result.insert decl.name fixed
return result
end Lean.Compiler.LCNF

View File

@@ -239,7 +239,7 @@ Iterate through `decl`, pushing local declarations that are only used in one
control flow arm into said arm in order to avoid useless computations.
-/
partial def floatLetIn (decl : Decl) : CompilerM Decl := do
let newValue go decl.value |>.run {}
let newValue decl.value.mapCodeM go |>.run {}
return { decl with value := newValue }
where
/--

View File

@@ -108,7 +108,7 @@ where
go (decl : Decl) : InternalizeM Decl := do
let type normExpr decl.type
let params decl.params.mapM internalizeParam
let value internalizeCode decl.value
let value decl.value.mapCodeM internalizeCode
return { decl with type, params, value }
/--

View File

@@ -133,7 +133,7 @@ this. This is because otherwise the calls to `myjp` in `f` and `g` would
produce out of scope join point jumps.
-/
partial def find (decl : Decl) : CompilerM FindState := do
let (_, candidates) go decl.value |>.run none |>.run {} |>.run' {}
let (_, candidates) decl.value.forCodeM go |>.run none |>.run {} |>.run' {}
return candidates
where
go : Code FindM Unit
@@ -178,7 +178,7 @@ and all calls to them with `jmp`s.
partial def replace (decl : Decl) (state : FindState) : CompilerM Decl := do
let mapper := fun acc cname _ => do return acc.insert cname ( mkFreshJpName)
let replaceCtx : ReplaceCtx state.candidates.foldM (init := .empty) mapper
let newValue go decl.value |>.run replaceCtx
let newValue decl.value.mapCodeM go |>.run replaceCtx
return { decl with value := newValue }
where
go (code : Code) : ReplaceM Code := do
@@ -389,7 +389,7 @@ position within the code so we can pull them out as far as possible, hopefully
enabling new inlining possibilities in the next simplifier run.
-/
partial def extend (decl : Decl) : CompilerM Decl := do
let newValue go decl.value |>.run {} |>.run' {} |>.run' {}
let newValue decl.value.mapCodeM go |>.run {} |>.run' {} |>.run' {}
let decl := { decl with value := newValue }
decl.pullFunDecls
where
@@ -510,8 +510,8 @@ After we have performed all of these optimizations we can take away the
code that has as little arguments as possible in the join points.
-/
partial def reduce (decl : Decl) : CompilerM Decl := do
let (_, analysis) goAnalyze decl.value |>.run {} |>.run {} |>.run' {}
let newValue goReduce decl.value |>.run analysis
let (_, analysis) decl.value.forCodeM goAnalyze |>.run {} |>.run {} |>.run' {}
let newValue decl.value.mapCodeM goReduce |>.run analysis
return { decl with value := newValue }
where
goAnalyzeFunDecl (fn : FunDecl) : ReduceAnalysisM Unit := do

View File

@@ -108,9 +108,10 @@ def mkAuxDecl (closure : Array Param) (decl : FunDecl) : LiftM LetDecl := do
where
go (nameNew : Name) (safe : Bool) (inlineAttr? : Option InlineAttributeKind) : InternalizeM Decl := do
let params := ( closure.mapM internalizeParam) ++ ( decl.params.mapM internalizeParam)
let value internalizeCode decl.value
let type value.inferType
let code internalizeCode decl.value
let type code.inferType
let type mkForallParams params type
let value := .code code
let decl := { name := nameNew, levelParams := [], params, type, value, safe, inlineAttr?, recursive := false : Decl }
return decl.setLevelParams
@@ -149,7 +150,7 @@ mutual
end
def main (decl : Decl) : LiftM Decl := do
let value withParams decl.params <| visitCode decl.value
let value withParams decl.params <| decl.value.mapCodeM visitCode
return { decl with value }
end LambdaLifting

View File

@@ -139,6 +139,10 @@ mutual
| .jmp _ args => visitArgs args
end
def visitDeclValue : DeclValue Visitor
| .code c => visitCode c
| .extern .. => id
end CollectLevelParams
open Lean.CollectLevelParams
@@ -149,7 +153,7 @@ Collect universe level parameters collecting in the type, parameters, and value,
set `decl.levelParams` with the resulting value.
-/
def Decl.setLevelParams (decl : Decl) : Decl :=
let levelParams := (visitCode decl.value visitParams decl.params visitType decl.type) {} |>.params.toList
let levelParams := (visitDeclValue decl.value visitParams decl.params visitType decl.type) {} |>.params.toList
{ decl with levelParams }
end Lean.Compiler.LCNF

View File

@@ -28,10 +28,10 @@ and `[specialize]` since they can be partially applied.
-/
def shouldGenerateCode (declName : Name) : CoreM Bool := do
if ( isCompIrrelevant |>.run') then return false
let env getEnv
if isExtern env declName then return true
let some info getDeclInfo? declName | return false
unless info.hasValue (allowOpaque := true) do return false
let env getEnv
if isExtern env declName then return false
if hasMacroInlineAttribute env declName then return false
if ( Meta.isMatcher declName) then return false
if isCasesOnRecursor env declName then return false

View File

@@ -105,6 +105,11 @@ mutual
return f!"⊥ : {← ppExpr type}"
else
return ""
partial def ppDeclValue (b : DeclValue) : M Format := do
match b with
| .code c => ppCode c
| .extern .. => return "extern"
end
def run (x : M α) : CompilerM α :=
@@ -121,7 +126,7 @@ def ppLetValue (e : LetValue) : CompilerM Format :=
def ppDecl (decl : Decl) : CompilerM Format :=
PP.run do
return f!"def {decl.name}{← PP.ppParams decl.params} : {← PP.ppExpr (← PP.getFunType decl.params decl.type)} :={indentD (← PP.ppCode decl.value)}"
return f!"def {decl.name}{← PP.ppParams decl.params} : {← PP.ppExpr (← PP.getFunType decl.params decl.type)} :={indentD (← PP.ppDeclValue decl.value)}"
def ppFunDecl (decl : FunDecl) : CompilerM Format :=
PP.run do

View File

@@ -57,7 +57,7 @@ where
| .cases (cases : CasesCore Code) => cases.alts.forM (go ·.getCode)
| .jmp .. | .return .. | .unreach .. => return ()
start (decls : Array Decl) : StateRefT (Array LetValue) CompilerM Unit :=
decls.forM (go ·.value)
decls.forM (·.value.forCodeM go)
partial def getJps : Probe Decl FunDecl := fun decls => do
let (_, res) start decls |>.run #[]
@@ -72,10 +72,10 @@ where
| .jmp .. | .return .. | .unreach .. => return ()
start (decls : Array Decl) : StateRefT (Array FunDecl) CompilerM Unit :=
decls.forM fun decl => go decl.value
decls.forM (·.value.forCodeM go)
partial def filterByLet (f : LetDecl CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let decl k => do if ( f decl) then return true else go k
@@ -84,7 +84,7 @@ where
| .jmp .. | .return .. | .unreach .. => return false
partial def filterByFun (f : FunDecl CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let _ k | .jp _ k => go k
@@ -93,7 +93,7 @@ where
| .jmp .. | .return .. | .unreach .. => return false
partial def filterByJp (f : FunDecl CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let _ k => go k
@@ -103,7 +103,7 @@ where
| .jmp .. | .return .. | .unreach .. => return false
partial def filterByFunDecl (f : FunDecl CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let _ k => go k
@@ -112,7 +112,7 @@ where
| .jmp .. | .return .. | .unreach .. => return false
partial def filterByCases (f : Cases CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let _ k => go k
@@ -121,7 +121,7 @@ where
| .jmp .. | .return .. | .unreach .. => return false
partial def filterByJmp (f : FVarId Array Arg CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let _ k => go k
@@ -131,7 +131,7 @@ where
| .return .. | .unreach .. => return false
partial def filterByReturn (f : FVarId CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let _ k => go k
@@ -141,7 +141,7 @@ where
| .return var => f var
partial def filterByUnreach (f : Expr CompilerM Bool) : Probe Decl Decl :=
filter (fun decl => go decl.value)
filter (·.value.isCodeAndM go)
where
go : Code CompilerM Bool
| .let _ k => go k

View File

@@ -172,8 +172,8 @@ open PullFunDecls
Pull local function declarations and join points in the given declaration.
-/
def Decl.pullFunDecls (decl : Decl) : CompilerM Decl := do
let (value, ps) pull decl.value |>.run []
let value := attach ps.toArray value
let (value, ps) decl.value.mapCodeM pull |>.run []
let value := value.mapCode (attach ps.toArray)
return { decl with value }
def pullFunDecls : Pass :=

View File

@@ -96,8 +96,8 @@ open PullLetDecls
def Decl.pullLetDecls (decl : Decl) (isCandidateFn : LetDecl FVarIdSet CompilerM Bool) : CompilerM Decl := do
PullM.run (isCandidateFn := isCandidateFn) do
withParams decl.params do
let value pullDecls decl.value
let value attachToPull value
let value decl.value.mapCodeM pullDecls
let value value.mapCodeM attachToPull
return { decl with value }
def Decl.pullInstances (decl : Decl) : CompilerM Decl :=

View File

@@ -108,7 +108,7 @@ partial def visit (code : Code) : FindUsedM Unit := do
def collectUsedParams (decl : Decl) : CompilerM FVarIdSet := do
let params := decl.params.foldl (init := {}) fun s p => s.insert p.fvarId
let (_, { used, .. }) visit decl.value |>.run { decl, params } |>.run {}
let (_, { used, .. }) decl.value.forCodeM visit |>.run { decl, params } |>.run {}
return used
end FindUsed
@@ -146,37 +146,40 @@ end ReduceArity
open FindUsed ReduceArity Internalize
def Decl.reduceArity (decl : Decl) : CompilerM (Array Decl) := do
let used collectUsedParams decl
if used.size == decl.params.size then
return #[decl] -- Declarations uses all parameters
else
trace[Compiler.reduceArity] "{decl.name}, used params: {used.toList.map mkFVar}"
let mask := decl.params.map fun param => used.contains param.fvarId
let auxName := decl.name ++ `_redArg
let mkAuxDecl : CompilerM Decl := do
let params := decl.params.filter fun param => used.contains param.fvarId
let value reduce decl.value |>.run { declName := decl.name, auxDeclName := auxName, paramMask := mask }
let type value.inferType
let type mkForallParams params type
let auxDecl := { decl with name := auxName, levelParams := [], type, params, value }
auxDecl.saveMono
return auxDecl
let updateDecl : InternalizeM Decl := do
let params decl.params.mapM internalizeParam
let mut args := #[]
for used in mask, param in params do
if used then
args := args.push param.toArg
let letDecl mkAuxLetDecl (.const auxName [] args)
let value := .let letDecl (.return letDecl.fvarId)
let decl := { decl with params, value, inlineAttr? := some .inline, recursive := false }
decl.saveMono
return decl
let unusedParams := decl.params.filter fun param => !used.contains param.fvarId
let auxDecl mkAuxDecl
let decl updateDecl |>.run' {}
eraseParams unusedParams
return #[auxDecl, decl]
match decl.value with
| .code code =>
let used collectUsedParams decl
if used.size == decl.params.size then
return #[decl] -- Declarations uses all parameters
else
trace[Compiler.reduceArity] "{decl.name}, used params: {used.toList.map mkFVar}"
let mask := decl.params.map fun param => used.contains param.fvarId
let auxName := decl.name ++ `_redArg
let mkAuxDecl : CompilerM Decl := do
let params := decl.params.filter fun param => used.contains param.fvarId
let value decl.value.mapCodeM reduce |>.run { declName := decl.name, auxDeclName := auxName, paramMask := mask }
let type code.inferType
let type mkForallParams params type
let auxDecl := { decl with name := auxName, levelParams := [], type, params, value }
auxDecl.saveMono
return auxDecl
let updateDecl : InternalizeM Decl := do
let params decl.params.mapM internalizeParam
let mut args := #[]
for used in mask, param in params do
if used then
args := args.push param.toArg
let letDecl mkAuxLetDecl (.const auxName [] args)
let value := .code (.let letDecl (.return letDecl.fvarId))
let decl := { decl with params, value, inlineAttr? := some .inline, recursive := false }
decl.saveMono
return decl
let unusedParams := decl.params.filter fun param => !used.contains param.fvarId
let auxDecl mkAuxDecl
let decl updateDecl |>.run' {}
eraseParams unusedParams
return #[auxDecl, decl]
| .extern .. => return #[decl]
def reduceArity : Pass where
phase := .mono
@@ -187,4 +190,4 @@ def reduceArity : Pass where
builtin_initialize
registerTraceClass `Compiler.reduceArity (inherited := true)
end Lean.Compiler.LCNF
end Lean.Compiler.LCNF

View File

@@ -68,7 +68,7 @@ open ReduceJpArity
Try to reduce arity of join points
-/
def Decl.reduceJpArity (decl : Decl) : CompilerM Decl := do
let value reduce decl.value |>.run {}
let value decl.value.mapCodeM reduce |>.run {}
return { decl with value }
def reduceJpArity (phase := Phase.base) : Pass :=

View File

@@ -55,7 +55,7 @@ def Decl.applyRenaming (decl : Decl) (r : Renaming) : CompilerM Decl := do
return decl
else
let params decl.params.mapMonoM (·.applyRenaming r)
let value decl.value.applyRenaming r
let value decl.value.mapCodeM (·.applyRenaming r)
return { decl with params, value }
end Lean.Compiler.LCNF
end Lean.Compiler.LCNF

View File

@@ -22,19 +22,20 @@ namespace Lean.Compiler.LCNF
open Simp
def Decl.simp? (decl : Decl) : SimpM (Option Decl) := do
updateFunDeclInfo decl.value
let .code code := decl.value | return none
updateFunDeclInfo code
traceM `Compiler.simp.inline.info do return m!"{decl.name}:{Format.nest 2 (← (← get).funDeclInfoMap.format)}"
traceM `Compiler.simp.step do ppDecl decl
let value simp decl.value
let code simp code
let s get
let value value.applyRenaming s.binderRenaming
traceM `Compiler.simp.step.new do return m!"{decl.name} :=\n{← ppCode value}"
trace[Compiler.simp.stat] "{decl.name}, size: {value.size}, # visited: {s.visited}, # inline: {s.inline}, # inline local: {s.inlineLocal}"
if let some value simpJpCases? value then
let decl := { decl with value }
let code code.applyRenaming s.binderRenaming
traceM `Compiler.simp.step.new do return m!"{decl.name} :=\n{← ppCode code}"
trace[Compiler.simp.stat] "{decl.name}, size: {code.size}, # visited: {s.visited}, # inline: {s.inline}, # inline local: {s.inlineLocal}"
if let some code simpJpCases? code then
let decl := { decl with value := .code code }
decl.reduceJpArity
else if ( get).simplified then
return some { decl with value }
return some { decl with value := .code code }
else
return none

View File

@@ -43,6 +43,7 @@ def inlineCandidate? (e : LetValue) : SimpM (Option InlineCandidateInfo) := do
unless ( read).config.inlineDefs do
return none
let some decl getDecl? declName | return none
let .code code := decl.value | return none
let shouldInline : SimpM Bool := do
if !decl.inlineIfReduceAttr && decl.recursive then return false
if mustInline then return true
@@ -63,9 +64,8 @@ def inlineCandidate? (e : LetValue) : SimpM (Option InlineCandidateInfo) := do
if decl.alwaysInlineAttr then return true
-- TODO: check inlining quota
if decl.inlineAttr || decl.inlineIfReduceAttr then return true
unless decl.noinlineAttr do
if ( isSmall decl.value) then return true
return false
if decl.noinlineAttr then return false
isSmall code
unless ( shouldInline) do return none
/- check arity -/
let arity := decl.getArity
@@ -77,7 +77,7 @@ def inlineCandidate? (e : LetValue) : SimpM (Option InlineCandidateInfo) := do
let arg := args[paramIdx]!
unless ( arg.isConstructorApp) do return none
let params := decl.instantiateParamsLevelParams us
let value := decl.instantiateValueLevelParams us
let value := code.instantiateValueLevelParams decl.levelParams us
let type := decl.instantiateTypeLevelParams us
incInline
return some {

View File

@@ -69,11 +69,14 @@ where
visit fvarId projs
else
let some decl getDecl? declName | failure
guard (decl.getArity == args.size)
let params := decl.instantiateParamsLevelParams us
let code := decl.instantiateValueLevelParams us
let code betaReduce params code args (mustInline := true)
visitCode code projs
match decl.value with
| .code code =>
guard (decl.getArity == args.size)
let params := decl.instantiateParamsLevelParams us
let code := code.instantiateValueLevelParams decl.levelParams us
let code betaReduce params code args (mustInline := true)
visitCode code projs
| .extern .. => failure
visitCode (code : Code) (projs : List Nat) : OptionT (StateRefT (Array CodeDecl) SimpM) FVarId := do
match code with

View File

@@ -222,6 +222,7 @@ def mkSpecDecl (decl : Decl) (us : List Level) (argMask : Array (Option Arg)) (p
eraseDecl decl
where
go (decl : Decl) (nameNew : Name) : InternalizeM Decl := do
let .code code := decl.value | panic! "can only specialize decls with code"
let mut params params.mapM internalizeParam
let decls decls.mapM internalizeCodeDecl
for param in decl.params, arg in argMask do
@@ -235,11 +236,12 @@ where
for param in decl.params[argMask.size:] do
let param := { param with type := param.type.instantiateLevelParamsNoCache decl.levelParams us }
params := params.push ( internalizeParam param)
let value := decl.instantiateValueLevelParams us
let value internalizeCode value
let value := attachCodeDecls decls value
let type value.inferType
let code := code.instantiateValueLevelParams decl.levelParams us
let code internalizeCode code
let code := attachCodeDecls decls code
let type code.inferType
let type mkForallParams params type
let value := .code code
let safe := decl.safe
let recursive := decl.recursive
let decl := { name := nameNew, levelParams := levelParamsNew, params, type, value, safe, recursive, inlineAttr? := none : Decl }
@@ -268,6 +270,7 @@ mutual
let some paramsInfo getSpecParamInfo? declName | return none
unless ( shouldSpecialize paramsInfo args) do return none
let some decl getDecl? declName | return none
let .code _ := decl.value | return none
trace[Compiler.specialize.candidate] "{e.toExpr}, {paramsInfo}"
let (argMask, params, decls) Collector.collect paramsInfo args
let keyBody := .const declName us (argMask.filterMap id)
@@ -290,7 +293,7 @@ mutual
let specDecl specDecl.simp {}
let specDecl specDecl.simp { etaPoly := true, inlinePartial := true, implementedBy := true }
let value withReader (fun _ => { declName := specDecl.name }) do
withParams specDecl.params <| visitCode specDecl.value
withParams specDecl.params <| specDecl.value.mapCodeM visitCode
let specDecl := { specDecl with value }
modify fun s => { s with decls := s.decls.push specDecl }
return some (.const specDecl.name usNew argsNew)
@@ -325,7 +328,7 @@ def main (decl : Decl) : SpecializeM Decl := do
if ( decl.isTemplateLike) then
return decl
else
let value withParams decl.params <| visitCode decl.value
let value withParams decl.params <| decl.value.mapCodeM visitCode
return { decl with value }
end Specialize

View File

@@ -235,12 +235,17 @@ Assert that the pass under test produces `Decl`s that do not contain
`Expr.const constName` in their `Code.let` values anymore.
-/
def assertDoesNotContainConstAfter (constName : Name) (msg : String) : TestInstaller :=
assertForEachDeclAfterEachOccurrence (fun _ decl => !decl.value.containsConst constName) msg
assertForEachDeclAfterEachOccurrence
fun _ decl =>
match decl.value with
| .code c => !c.containsConst constName
| .extern .. => true
msg
def assertNoFun : TestInstaller :=
assertAfter do
for decl in ( getDecls) do
decl.value.forM fun
decl.value.forCodeM fun
| .fun .. => throwError "declaration `{decl.name}` contains a local function declaration"
| _ => return ()

View File

@@ -6,6 +6,7 @@ Authors: Leonardo de Moura
prelude
import Lean.Meta.Transform
import Lean.Meta.Match.MatcherInfo
import Lean.Compiler.ExternAttr
import Lean.Compiler.ImplementedByAttr
import Lean.Compiler.LCNF.ToLCNF
@@ -96,31 +97,48 @@ The steps for this are roughly:
def toDecl (declName : Name) : CompilerM Decl := do
let declName := if let some name := isUnsafeRecName? declName then name else declName
let some info getDeclInfo? declName | throwError "declaration `{declName}` not found"
let some value := info.value? (allowOpaque := true) | throwError "declaration `{declName}` does not have a value"
let (type, value) Meta.MetaM.run' do
let type toLCNFType info.type
let value Meta.lambdaTelescope value fun xs body => do Meta.mkLambdaFVars xs ( Meta.etaExpand body)
let value replaceUnsafeRecNames value
let value macroInline value
/- Recall that some declarations tagged with `macro_inline` contain matchers. -/
let value inlineMatchers value
/- Recall that `inlineMatchers` may have exposed `ite`s and `dite`s which are tagged as `[macro_inline]`. -/
let value macroInline value
/-
Remark: we have disabled the following transformatbion, we will perform it at phase 2, after code specialization.
It prevents many optimizations (e.g., "cases-of-ctor").
-/
-- let value ← applyCasesOnImplementedBy value
return (type, value)
let value toLCNF value
let safe := !info.isPartial && !info.isUnsafe
let inlineAttr? := getInlineAttribute? ( getEnv) declName
let decl if let .fun decl (.return _) := value then
eraseFunDecl decl (recursive := false)
pure { name := declName, params := decl.params, type, value := decl.value, levelParams := info.levelParams, safe, inlineAttr? : Decl }
if let some externAttrData := getExternAttrData? ( getEnv) declName then
let paramsFromTypeBinders (expr : Expr) : CompilerM (Array Param) := do
let mut params := #[]
let mut currentExpr := expr
repeat
match currentExpr with
| .forallE binderName type body _ =>
let borrow := isMarkedBorrowed type
params := params.push ( mkParam binderName type borrow)
currentExpr := body
| _ => break
return params
let type Meta.MetaM.run' (toLCNFType info.type)
let params paramsFromTypeBinders type
return { name := declName, params, type, value := .extern externAttrData, levelParams := info.levelParams, safe, inlineAttr? }
else
pure { name := declName, params := #[], type, value, levelParams := info.levelParams, safe, inlineAttr? }
/- `toLCNF` may eta-reduce simple declarations. -/
decl.etaExpand
let some value := info.value? (allowOpaque := true) | throwError "declaration `{declName}` does not have a value"
let (type, value) Meta.MetaM.run' do
let type toLCNFType info.type
let value Meta.lambdaTelescope value fun xs body => do Meta.mkLambdaFVars xs ( Meta.etaExpand body)
let value replaceUnsafeRecNames value
let value macroInline value
/- Recall that some declarations tagged with `macro_inline` contain matchers. -/
let value inlineMatchers value
/- Recall that `inlineMatchers` may have exposed `ite`s and `dite`s which are tagged as `[macro_inline]`. -/
let value macroInline value
/-
Remark: we have disabled the following transformatbion, we will perform it at phase 2, after code specialization.
It prevents many optimizations (e.g., "cases-of-ctor").
-/
-- let value ← applyCasesOnImplementedBy value
return (type, value)
let code toLCNF value
let decl if let .fun decl (.return _) := code then
eraseFunDecl decl (recursive := false)
pure { name := declName, params := decl.params, type, value := .code decl.value, levelParams := info.levelParams, safe, inlineAttr? : Decl }
else
pure { name := declName, params := #[], type, value := .code code, levelParams := info.levelParams, safe, inlineAttr? }
/- `toLCNF` may eta-reduce simple declarations. -/
decl.etaExpand
end Lean.Compiler.LCNF

View File

@@ -110,7 +110,4 @@ def Code.toExpr (code : Code) (xs : Array FVarId := #[]) : Expr :=
def FunDeclCore.toExpr (decl : FunDecl) (xs : Array FVarId := #[]) : Expr :=
run' decl.toExprM xs
def Decl.toExpr (decl : Decl) : Expr :=
run do withParams decl.params do mkLambdaM decl.params ( decl.value.toExprM)
end Lean.Compiler.LCNF

View File

@@ -143,7 +143,7 @@ where
go : ToMonoM Decl := do
let type toMonoType decl.type
let params decl.params.mapM (·.toMono)
let value decl.value.toMono
let value decl.value.mapCodeM (·.toMono)
let decl := { decl with type, params, value, levelParams := [] }
decl.saveMono
return decl

View File

@@ -6,6 +6,7 @@ Authors: Marc Huisinga, Wojciech Nawrocki
-/
prelude
import Lean.Data.Lsp.Basic
import Lean.Data.Lsp.CancelParams
import Lean.Data.Lsp.Capabilities
import Lean.Data.Lsp.Client
import Lean.Data.Lsp.Communication

View File

@@ -6,7 +6,6 @@ Authors: Marc Huisinga, Wojciech Nawrocki
-/
prelude
import Lean.Data.Json
import Lean.Data.JsonRpc
/-! Defines most of the 'Basic Structures' in the LSP specification
(https://microsoft.github.io/language-server-protocol/specifications/specification-current/),
@@ -19,10 +18,6 @@ namespace Lsp
open Json
structure CancelParams where
id : JsonRpc.RequestID
deriving Inhabited, BEq, ToJson, FromJson
abbrev DocumentUri := String
/-- We adopt the convention that zero-based UTF-16 positions as sent by LSP clients

View File

@@ -0,0 +1,25 @@
/-
Copyright (c) 2020 Marc Huisinga. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Marc Huisinga, Wojciech Nawrocki
-/
prelude
import Lean.Data.JsonRpc
/-! # Defines `Lean.Lsp.CancelParams`.
This is separate from `Lean.Data.Lsp.Basic` to reduce transitive dependencies.
-/
namespace Lean
namespace Lsp
open Json
structure CancelParams where
id : JsonRpc.RequestID
deriving Inhabited, BEq, ToJson, FromJson
end Lsp
end Lean

View File

@@ -6,7 +6,6 @@ Authors: Marc Huisinga, Wojciech Nawrocki
-/
prelude
import Init.Data.String
import Init.Data.Array
import Lean.Data.Lsp.Basic
import Lean.Data.Position
import Lean.DeclarationRange

View File

@@ -49,3 +49,8 @@ variable {_ : BEq α} {_ : Hashable α}
@[inline] def fold {β : Type v} (f : β α β) (init : β) (s : PersistentHashSet α) : β :=
Id.run $ s.foldM f init
def toList (s : PersistentHashSet α) : List α :=
s.set.toList.map (·.1)
end PersistentHashSet

View File

@@ -131,14 +131,18 @@ def throwCalcFailure (steps : Array CalcStepView) (expectedType result : Expr) :
if isDefEqGuarded r er then
let mut failed := false
unless isDefEqGuarded lhs elhs do
let (lhs, elhs) addPPExplicitToExposeDiff lhs elhs
let (lhsTy, elhsTy) addPPExplicitToExposeDiff ( inferType lhs) ( inferType elhs)
logErrorAt steps[0]!.term m!"\
invalid 'calc' step, left-hand side is{indentD m!"{lhs} : { inferType lhs}"}\n\
but is expected to be{indentD m!"{elhs} : { inferType elhs}"}"
invalid 'calc' step, left-hand side is{indentD m!"{lhs} : {lhsTy}"}\n\
but is expected to be{indentD m!"{elhs} : {elhsTy}"}"
failed := true
unless isDefEqGuarded rhs erhs do
let (rhs, erhs) addPPExplicitToExposeDiff rhs erhs
let (rhsTy, erhsTy) addPPExplicitToExposeDiff ( inferType rhs) ( inferType erhs)
logErrorAt steps.back!.term m!"\
invalid 'calc' step, right-hand side is{indentD m!"{rhs} : { inferType rhs}"}\n\
but is expected to be{indentD m!"{erhs} : { inferType erhs}"}"
invalid 'calc' step, right-hand side is{indentD m!"{rhs} : {rhsTy}"}\n\
but is expected to be{indentD m!"{erhs} : {erhsTy}"}"
failed := true
if failed then
throwAbortTerm

View File

@@ -38,6 +38,7 @@ def elabCheckTactic : CommandElab := fun stx => do
| [next] => do
let (val, _, _) matchCheckGoalType stx (next.getType)
if !( Meta.withReducible <| isDefEq val expTerm) then
let (val, expTerm) addPPExplicitToExposeDiff val expTerm
throwErrorAt stx
m!"Term reduces to{indentExpr val}\nbut is expected to reduce to {indentExpr expTerm}"
| _ => do

View File

@@ -16,3 +16,4 @@ import Lean.Elab.Deriving.FromToJson
import Lean.Elab.Deriving.SizeOf
import Lean.Elab.Deriving.Hashable
import Lean.Elab.Deriving.Ord
import Lean.Elab.Deriving.ToExpr

View File

@@ -0,0 +1,237 @@
/-
Copyright (c) 2024 Lean FRO, LLC. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Kyle Miller
-/
prelude
import Lean.Meta.Transform
import Lean.Elab.Deriving.Basic
import Lean.Elab.Deriving.Util
import Lean.ToLevel
import Lean.ToExpr
/-!
# `ToExpr` deriving handler
This module defines a `ToExpr` deriving handler for inductive types.
It supports mutually inductive types as well.
The `ToExpr` deriving handlers support universe level polymorphism, via the `Lean.ToLevel` class.
To use `ToExpr` in places where there is universe polymorphism, make sure a `[ToLevel.{u}]` instance is available,
though be aware that the `ToLevel` mechanism does not support `max` or `imax` expressions.
Implementation note: this deriving handler was initially modeled after the `Repr` deriving handler, but
1. we need to account for universe levels,
2. the `ToExpr` class has two fields rather than one, and
3. we don't handle structures specially.
-/
namespace Lean.Elab.Deriving.ToExpr
open Lean Elab Parser.Term
open Meta Command Deriving
/--
Given `args := #[e₁, e₂, …, eₙ]`, constructs the syntax `Expr.app (… (Expr.app (Expr.app f e₁) e₂) …) eₙ`.
-/
def mkAppNTerm (f : Term) (args : Array Term) : MetaM Term :=
args.foldlM (fun a b => ``(Expr.app $a $b)) f
/-- Fixes the output of `mkInductiveApp` to explicitly reference universe levels. -/
def updateIndType (indVal : InductiveVal) (t : Term) : TermElabM Term :=
let levels := indVal.levelParams.toArray.map mkIdent
match t with
| `(@$f $args*) => `(@$f.{$levels,*} $args*)
| _ => throwError "(internal error) expecting output of `mkInductiveApp`"
/--
Creates a term that evaluates to an expression representing the inductive type.
Uses `toExpr` and `toTypeExpr` for the arguments to the type constructor.
-/
def mkToTypeExpr (indVal : InductiveVal) (argNames : Array Name) : TermElabM Term := do
let levels indVal.levelParams.toArray.mapM (fun u => `(Lean.toLevel.{$(mkIdent u)}))
forallTelescopeReducing indVal.type fun xs _ => do
let mut args : Array Term := #[]
for argName in argNames, x in xs do
let a := mkIdent argName
if Meta.isType x then
args := args.push <| ``(toTypeExpr $a)
else
args := args.push <| ``(toExpr $a)
mkAppNTerm ( ``(Expr.const $(quote indVal.name) [$levels,*])) args
/--
Creates the body of the `toExpr` function for the `ToExpr` instance, which is a `match` expression
that calls `toExpr` and `toTypeExpr` to assemble an expression for a given term.
For recursive inductive types, `auxFunName` refers to the `ToExpr` instance for the current type.
For mutually recursive types, we rely on the local instances set up by `mkLocalInstanceLetDecls`.
-/
def mkToExprBody (header : Header) (indVal : InductiveVal) (auxFunName : Name) (levelInsts : Array Term) :
TermElabM Term := do
let discrs mkDiscrs header indVal
let alts mkAlts
`(match $[$discrs],* with $alts:matchAlt*)
where
/-- Create the `match` cases, one per constructor. -/
mkAlts : TermElabM (Array (TSyntax ``matchAlt)) := do
let levels levelInsts.mapM fun inst => `($(inst).toLevel)
let mut alts := #[]
for ctorName in indVal.ctors do
let ctorInfo getConstInfoCtor ctorName
let alt forallTelescopeReducing ctorInfo.type fun xs _ => do
let mut patterns := #[]
-- add `_` pattern for indices, before the constructor's pattern
for _ in [:indVal.numIndices] do
patterns := patterns.push ( `(_))
let mut ctorArgs := #[]
let mut rhsArgs : Array Term := #[]
let mkArg (x : Expr) (a : Term) : TermElabM Term := do
if ( inferType x).isAppOf indVal.name then
`($(mkIdent auxFunName) $levelInsts* $a)
else if Meta.isType x then
``(toTypeExpr $a)
else
``(toExpr $a)
-- add `_` pattern for inductive parameters, which are inaccessible
for i in [:ctorInfo.numParams] do
let a := mkIdent header.argNames[i]!
ctorArgs := ctorArgs.push ( `(_))
rhsArgs := rhsArgs.push <| mkArg xs[i]! a
for i in [:ctorInfo.numFields] do
let a := mkIdent ( mkFreshUserName `a)
ctorArgs := ctorArgs.push a
rhsArgs := rhsArgs.push <| mkArg xs[ctorInfo.numParams + i]! a
patterns := patterns.push ( `(@$(mkIdent ctorName):ident $ctorArgs:term*))
let rhs : Term mkAppNTerm ( ``(Expr.const $(quote ctorInfo.name) [$levels,*])) rhsArgs
`(matchAltExpr| | $[$patterns:term],* => $rhs)
alts := alts.push alt
return alts
/--
For nested and mutually recursive inductive types, we define `partial` instances,
and the strategy is to have local `ToExpr` instances in scope for the body of each instance.
This way, each instance can freely use `toExpr` and `toTypeExpr` for each of the types in `ctx`.
This is a modified copy of `Lean.Elab.Deriving.mkLocalInstanceLetDecls`,
since we need to include the `toTypeExpr` field in the `letDecl`
Note that, for simplicity, each instance gets its own definition of each others' `toTypeExpr` fields.
These are very simple fields, so avoiding the duplication is not worth it.
-/
def mkLocalInstanceLetDecls (ctx : Deriving.Context) (argNames : Array Name) (levelInsts : Array Term) :
TermElabM (Array (TSyntax ``Parser.Term.letDecl)) := do
let mut letDecls := #[]
for indVal in ctx.typeInfos, auxFunName in ctx.auxFunNames do
let currArgNames mkInductArgNames indVal
let numParams := indVal.numParams
let currIndices := currArgNames[numParams:]
let binders mkImplicitBinders currIndices
let argNamesNew := argNames[:numParams] ++ currIndices
let indType mkInductiveApp indVal argNamesNew
let instName mkFreshUserName `localinst
let toTypeExpr mkToTypeExpr indVal argNames
-- Recall that mutually inductive types all use the same universe levels, hence we pass the same ToLevel instances to each aux function.
let letDecl `(Parser.Term.letDecl| $(mkIdent instName):ident $binders:implicitBinder* : ToExpr $indType :=
{ toExpr := $(mkIdent auxFunName) $levelInsts*,
toTypeExpr := $toTypeExpr })
letDecls := letDecls.push letDecl
return letDecls
open TSyntax.Compat in
/--
Makes a `toExpr` function for the given inductive type.
The implementation of each `toExpr` function for a (mutual) inductive type is given as top-level private definitions.
These are assembled into `ToExpr` instances in `mkInstanceCmds`.
For mutual/nested inductive types, then each of the types' `ToExpr` instances are provided as local instances,
to wire together the recursion (necessitating these auxiliary definitions being `partial`).
-/
def mkAuxFunction (ctx : Deriving.Context) (i : Nat) : TermElabM Command := do
let auxFunName := ctx.auxFunNames[i]!
let indVal := ctx.typeInfos[i]!
let header mkHeader ``ToExpr 1 indVal
/- We make the `ToLevel` instances be explicit here so that we can pass the instances from the instances to the
aux functions. This lets us ensure universe level variables are being lined up,
without needing to use `ident.{u₁,…,uₙ}` syntax, which could conditionally be incorrect
depending on the ambient CommandElabM scope state.
TODO(kmill): deriving handlers should run in a scope with no `universes` or `variables`. -/
let (toLevelInsts, levelBinders) := Array.unzip <| indVal.levelParams.toArray.mapM fun u => do
let inst := mkIdent ( mkFreshUserName `inst)
return (inst, `(explicitBinderF| ($inst : ToLevel.{$(mkIdent u)})))
let mut body mkToExprBody header indVal auxFunName toLevelInsts
if ctx.usePartial then
let letDecls mkLocalInstanceLetDecls ctx header.argNames toLevelInsts
body mkLet letDecls body
/- We need to alter the last binder (the one for the "target") to have explicit universe levels
so that the `ToLevel` instance arguments can use them. -/
let addLevels binder :=
match binder with
| `(bracketedBinderF| ($a : $ty)) => do `(bracketedBinderF| ($a : $( updateIndType indVal ty)))
| _ => throwError "(internal error) expecting inst binder"
let binders := header.binders.pop ++ levelBinders ++ #[ addLevels header.binders.back!]
if ctx.usePartial then
`(private partial def $(mkIdent auxFunName):ident $binders:bracketedBinder* : Expr := $body:term)
else
`(private def $(mkIdent auxFunName):ident $binders:bracketedBinder* : Expr := $body:term)
/--
Creates all the auxiliary functions (using `mkAuxFunction`) for the (mutual) inductive type(s).
Wraps the resulting definition commands in `mutual ... end`.
-/
def mkAuxFunctions (ctx : Deriving.Context) : TermElabM Syntax := do
let mut auxDefs := #[]
for i in [:ctx.typeInfos.size] do
auxDefs := auxDefs.push ( mkAuxFunction ctx i)
`(mutual $auxDefs:command* end)
open TSyntax.Compat in
/--
Assuming all of the auxiliary definitions exist,
creates all the `instance` commands for the `ToExpr` instances for the (mutual) inductive type(s).
This is a modified copy of `Lean.Elab.Deriving.mkInstanceCmds` to account for `ToLevel` instances.
-/
def mkInstanceCmds (ctx : Deriving.Context) (typeNames : Array Name) :
TermElabM (Array Command) := do
let mut instances := #[]
for indVal in ctx.typeInfos, auxFunName in ctx.auxFunNames do
if typeNames.contains indVal.name then
let argNames mkInductArgNames indVal
let binders mkImplicitBinders argNames
let binders := binders ++ ( mkInstImplicitBinders ``ToExpr indVal argNames)
let (toLevelInsts, levelBinders) := Array.unzip <| indVal.levelParams.toArray.mapM fun u => do
let inst := mkIdent ( mkFreshUserName `inst)
return (inst, `(instBinderF| [$inst : ToLevel.{$(mkIdent u)}]))
let binders := binders ++ levelBinders
let indType updateIndType indVal ( mkInductiveApp indVal argNames)
let toTypeExpr mkToTypeExpr indVal argNames
let instCmd `(instance $binders:implicitBinder* : ToExpr $indType where
toExpr := $(mkIdent auxFunName) $toLevelInsts*
toTypeExpr := $toTypeExpr)
instances := instances.push instCmd
return instances
/--
Returns all the commands necessary to construct the `ToExpr` instances.
-/
def mkToExprInstanceCmds (declNames : Array Name) : TermElabM (Array Syntax) := do
let ctx mkContext "toExpr" declNames[0]!
let cmds := #[ mkAuxFunctions ctx] ++ ( mkInstanceCmds ctx declNames)
trace[Elab.Deriving.toExpr] "\n{cmds}"
return cmds
/--
The main entry point to the `ToExpr` deriving handler.
-/
def mkToExprInstanceHandler (declNames : Array Name) : CommandElabM Bool := do
if ( declNames.allM isInductive) && declNames.size > 0 then
let cmds withFreshMacroScope <| liftTermElabM <| mkToExprInstanceCmds declNames
-- Enable autoimplicits, used for universe levels.
withScope (fun scope => { scope with opts := autoImplicit.set scope.opts true }) do
elabCommand (mkNullNode cmds)
return true
else
return false
builtin_initialize
registerDerivingHandler ``Lean.ToExpr mkToExprInstanceHandler
registerTraceClass `Elab.Deriving.toExpr
end Lean.Elab.Deriving.ToExpr

View File

@@ -691,6 +691,9 @@ private def addProjections (r : ElabHeaderResult) (fieldInfos : Array StructFiel
let env getEnv
let env ofExceptKernelException (mkProjections env r.view.declName projNames.toList r.view.isClass)
setEnv env
for fieldInfo in fieldInfos do
if fieldInfo.isSubobject then
addDeclarationRangesFromSyntax fieldInfo.declName r.view.ref fieldInfo.ref
private def registerStructure (structName : Name) (infos : Array StructFieldInfo) : TermElabM Unit := do
let fields infos.filterMapM fun info => do
@@ -775,14 +778,14 @@ private def setSourceInstImplicit (type : Expr) : Expr :=
/--
Creates a projection function to a non-subobject parent.
-/
private partial def mkCoercionToCopiedParent (levelParams : List Name) (params : Array Expr) (view : StructView) (source : Expr) (parentStructName : Name) (parentType : Expr) : MetaM StructureParentInfo := do
private partial def mkCoercionToCopiedParent (levelParams : List Name) (params : Array Expr) (view : StructView) (source : Expr) (parent : StructParentInfo) (parentType : Expr) : MetaM StructureParentInfo := do
let isProp Meta.isProp parentType
let env getEnv
let structName := view.declName
let sourceFieldNames := getStructureFieldsFlattened env structName
let binfo := if view.isClass && isClass env parentStructName then BinderInfo.instImplicit else BinderInfo.default
let binfo := if view.isClass && isClass env parent.structName then BinderInfo.instImplicit else BinderInfo.default
let mut declType instantiateMVars ( mkForallFVars params ( mkForallFVars #[source] parentType))
if view.isClass && isClass env parentStructName then
if view.isClass && isClass env parent.structName then
declType := setSourceInstImplicit declType
declType := declType.inferImplicit params.size true
let rec copyFields (parentType : Expr) : MetaM Expr := do
@@ -823,7 +826,8 @@ private partial def mkCoercionToCopiedParent (levelParams : List Name) (params :
-- (Instances will get instance reducibility in `Lean.Elab.Command.addParentInstances`.)
if !binfo.isInstImplicit && !( Meta.isProp parentType) then
setReducibleAttribute declName
return { structName := parentStructName, subobject := false, projFn := declName }
addDeclarationRangesFromSyntax declName view.ref parent.ref
return { structName := parent.structName, subobject := false, projFn := declName }
private def mkRemainingProjections (levelParams : List Name) (params : Array Expr) (view : StructView)
(parents : Array StructParentInfo) (fieldInfos : Array StructFieldInfo) : TermElabM (Array StructureParentInfo) := do
@@ -844,7 +848,7 @@ private def mkRemainingProjections (levelParams : List Name) (params : Array Exp
pure { structName := parent.structName, subobject := true, projFn := info.declName }
else
let parent_type := ( instantiateMVars parent.type).replace fun e => parentFVarToConst[e]?
mkCoercionToCopiedParent levelParams params view source parent.structName parent_type)
mkCoercionToCopiedParent levelParams params view source parent parent_type)
parentInfos := parentInfos.push parentInfo
if let some fvar := parent.fvar? then
parentFVarToConst := parentFVarToConst.insert fvar <|

View File

@@ -44,3 +44,5 @@ import Lean.Elab.Tactic.DiscrTreeKey
import Lean.Elab.Tactic.BVDecide
import Lean.Elab.Tactic.BoolToPropSimps
import Lean.Elab.Tactic.Classical
import Lean.Elab.Tactic.Grind
import Lean.Elab.Tactic.Monotonicity

View File

@@ -82,7 +82,7 @@ instance : ToExpr Gate where
| .and => mkConst ``Gate.and
| .xor => mkConst ``Gate.xor
| .beq => mkConst ``Gate.beq
| .imp => mkConst ``Gate.imp
| .or => mkConst ``Gate.or
toTypeExpr := mkConst ``Gate
instance : ToExpr BVPred where

View File

@@ -91,7 +91,7 @@ where
| .and => ``Std.Tactic.BVDecide.Reflect.Bool.and_congr
| .xor => ``Std.Tactic.BVDecide.Reflect.Bool.xor_congr
| .beq => ``Std.Tactic.BVDecide.Reflect.Bool.beq_congr
| .imp => ``Std.Tactic.BVDecide.Reflect.Bool.imp_congr
| .or => ``Std.Tactic.BVDecide.Reflect.Bool.or_congr
/--
Construct the reified version of `Bool.not subExpr`.
@@ -136,7 +136,7 @@ def mkIte (discr lhs rhs : ReifiedBVLogical) (discrExpr lhsExpr rhsExpr : Expr)
lhsEvalExpr lhsProof?
rhsEvalExpr rhsProof? | return none
return mkApp9
(mkConst ``Std.Tactic.BVDecide.Reflect.Bool.ite_congr)
(mkConst ``Std.Tactic.BVDecide.Reflect.Bool.cond_congr)
discrExpr lhsExpr rhsExpr
discrEvalExpr lhsEvalExpr rhsEvalExpr
discrProof lhsProof rhsProof

View File

@@ -22,67 +22,70 @@ This function adds the two lemmas:
- `discrExpr = false → atomExpr = rhsExpr`
It assumes that `discrExpr`, `lhsExpr` and `rhsExpr` are the expressions corresponding to `discr`,
`lhs` and `rhs`. Furthermore it assumes that `atomExpr` is of the form
`if discrExpr = true then lhsExpr else rhsExpr`.
`bif discrExpr then lhsExpr else rhsExpr`.
-/
def addIfLemmas (discr : ReifiedBVLogical) (atom lhs rhs : ReifiedBVExpr)
def addCondLemmas (discr : ReifiedBVLogical) (atom lhs rhs : ReifiedBVExpr)
(discrExpr atomExpr lhsExpr rhsExpr : Expr) : LemmaM Unit := do
let some trueLemma mkIfTrueLemma discr atom lhs rhs discrExpr atomExpr lhsExpr rhsExpr | return ()
let some trueLemma mkCondTrueLemma discr atom lhs discrExpr atomExpr lhsExpr rhsExpr | return ()
LemmaM.addLemma trueLemma
let some falseLemma mkIfFalseLemma discr atom lhs rhs discrExpr atomExpr lhsExpr rhsExpr | return ()
let some falseLemma mkCondFalseLemma discr atom rhs discrExpr atomExpr lhsExpr rhsExpr | return ()
LemmaM.addLemma falseLemma
where
mkIfTrueLemma (discr : ReifiedBVLogical) (atom lhs rhs : ReifiedBVExpr)
(discrExpr atomExpr lhsExpr rhsExpr : Expr) : M (Option SatAtBVLogical) :=
mkIfLemma true discr atom lhs rhs discrExpr atomExpr lhsExpr rhsExpr
mkIfFalseLemma (discr : ReifiedBVLogical) (atom lhs rhs : ReifiedBVExpr)
(discrExpr atomExpr lhsExpr rhsExpr : Expr) : M (Option SatAtBVLogical) :=
mkIfLemma false discr atom lhs rhs discrExpr atomExpr lhsExpr rhsExpr
mkIfLemma (discrVal : Bool) (discr : ReifiedBVLogical) (atom lhs rhs : ReifiedBVExpr)
mkCondTrueLemma (discr : ReifiedBVLogical) (atom lhs : ReifiedBVExpr)
(discrExpr atomExpr lhsExpr rhsExpr : Expr) : M (Option SatAtBVLogical) := do
let resExpr := if discrVal then lhsExpr else rhsExpr
let resValExpr := if discrVal then lhs else rhs
let lemmaName :=
if discrVal then
``Std.Tactic.BVDecide.Reflect.BitVec.if_true
else
``Std.Tactic.BVDecide.Reflect.BitVec.if_false
let discrValExpr := toExpr discrVal
let discrVal ReifiedBVLogical.mkBoolConst discrVal
let resExpr := lhsExpr
let resValExpr := lhs
let lemmaName := ``Std.Tactic.BVDecide.Reflect.BitVec.cond_true
let eqDiscrExpr mkAppM ``BEq.beq #[discrExpr, discrValExpr]
let eqDiscr ReifiedBVLogical.mkGate discr discrVal discrExpr discrValExpr .beq
let notDiscrExpr := mkApp (mkConst ``Bool.not) discrExpr
let notDiscr ReifiedBVLogical.mkNot discr discrExpr
let eqBVExpr mkAppM ``BEq.beq #[atomExpr, resExpr]
let some eqBVPred ReifiedBVPred.mkBinPred atom resValExpr atomExpr resExpr .eq | return none
let eqBV ReifiedBVLogical.ofPred eqBVPred
let imp ReifiedBVLogical.mkGate eqDiscr eqBV eqDiscrExpr eqBVExpr .imp
let imp ReifiedBVLogical.mkGate notDiscr eqBV notDiscrExpr eqBVExpr .or
let proof := do
let evalExpr ReifiedBVLogical.mkEvalExpr imp.expr
let congrProof := ( imp.evalsAtAtoms).getD (ReifiedBVLogical.mkRefl evalExpr)
let lemmaProof := mkApp4 (mkConst lemmaName) (toExpr lhs.width) discrExpr lhsExpr rhsExpr
let trueExpr := mkConst ``Bool.true
let eqDiscrTrueExpr mkEq eqDiscrExpr trueExpr
let eqBVExprTrueExpr mkEq eqBVExpr trueExpr
let impExpr mkArrow eqDiscrTrueExpr eqBVExprTrueExpr
-- construct a `Decidable` instance for the implication using forall_prop_decidable
let decEqDiscrTrue := mkApp2 (mkConst ``instDecidableEqBool) eqDiscrExpr trueExpr
let decEqBVExprTrue := mkApp2 (mkConst ``instDecidableEqBool) eqBVExpr trueExpr
let impDecidable := mkApp4 (mkConst ``forall_prop_decidable)
eqDiscrTrueExpr
(.lam .anonymous eqDiscrTrueExpr eqBVExprTrueExpr .default)
decEqDiscrTrue
(.lam .anonymous eqDiscrTrueExpr decEqBVExprTrue .default)
let decideImpExpr := mkApp2 (mkConst ``Decidable.decide) impExpr impDecidable
-- !discr || (atom == rhs)
let impExpr := mkApp2 (mkConst ``Bool.or) notDiscrExpr eqBVExpr
return mkApp4
(mkConst ``Std.Tactic.BVDecide.Reflect.Bool.lemma_congr)
decideImpExpr
impExpr
evalExpr
congrProof
lemmaProof
return some imp.bvExpr, proof, imp.expr
mkCondFalseLemma (discr : ReifiedBVLogical) (atom rhs : ReifiedBVExpr)
(discrExpr atomExpr lhsExpr rhsExpr : Expr) : M (Option SatAtBVLogical) := do
let resExpr := rhsExpr
let resValExpr := rhs
let lemmaName := ``Std.Tactic.BVDecide.Reflect.BitVec.cond_false
let eqBVExpr mkAppM ``BEq.beq #[atomExpr, resExpr]
let some eqBVPred ReifiedBVPred.mkBinPred atom resValExpr atomExpr resExpr .eq | return none
let eqBV ReifiedBVLogical.ofPred eqBVPred
let imp ReifiedBVLogical.mkGate discr eqBV discrExpr eqBVExpr .or
let proof := do
let evalExpr ReifiedBVLogical.mkEvalExpr imp.expr
let congrProof := ( imp.evalsAtAtoms).getD (ReifiedBVLogical.mkRefl evalExpr)
let lemmaProof := mkApp4 (mkConst lemmaName) (toExpr rhs.width) discrExpr lhsExpr rhsExpr
-- discr || (atom == rhs)
let impExpr := mkApp2 (mkConst ``Bool.or) discrExpr eqBVExpr
return mkApp4
(mkConst ``Std.Tactic.BVDecide.Reflect.Bool.lemma_congr)
impExpr
evalExpr
congrProof
lemmaProof

View File

@@ -220,15 +220,12 @@ where
.rotateRight
``BVUnOp.rotateRight
``Std.Tactic.BVDecide.Reflect.BitVec.rotateRight_congr
| ite _ discrExpr _ lhsExpr rhsExpr =>
let_expr Eq α discrExpr val := discrExpr | return none
let_expr Bool := α | return none
let_expr Bool.true := val | return none
| cond _ discrExpr lhsExpr rhsExpr =>
let some atom ReifiedBVExpr.bitVecAtom x true | return none
let some discr ReifiedBVLogical.of discrExpr | return none
let some lhs goOrAtom lhsExpr | return none
let some rhs goOrAtom rhsExpr | return none
addIfLemmas discr atom lhs rhs discrExpr x lhsExpr rhsExpr
addCondLemmas discr atom lhs rhs discrExpr x lhsExpr rhsExpr
return some atom
| _ => return none
@@ -392,10 +389,7 @@ where
| Bool => gateReflection lhsExpr rhsExpr .beq
| BitVec _ => goPred t
| _ => return none
| ite _ discrExpr _ lhsExpr rhsExpr =>
let_expr Eq α discrExpr val := discrExpr | return none
let_expr Bool := α | return none
let_expr Bool.true := val | return none
| cond _ discrExpr lhsExpr rhsExpr =>
let some discr goOrAtom discrExpr | return none
let some lhs goOrAtom lhsExpr | return none
let some rhs goOrAtom rhsExpr | return none

View File

@@ -28,6 +28,18 @@ namespace Frontend.Normalize
open Lean.Meta
open Std.Tactic.BVDecide.Normalize
builtin_simproc [bv_normalize] reduceCond (cond _ _ _) := fun e => do
let_expr f@cond α c tb eb := e | return .continue
let r Simp.simp c
if r.expr.cleanupAnnotations.isConstOf ``Bool.true then
let pr := mkApp (mkApp4 (mkConst ``Bool.cond_pos f.constLevels!) α c tb eb) ( r.getProof)
return .visit { expr := tb, proof? := pr }
else if r.expr.cleanupAnnotations.isConstOf ``Bool.false then
let pr := mkApp (mkApp4 (mkConst ``Bool.cond_neg f.constLevels!) α c tb eb) ( r.getProof)
return .visit { expr := eb, proof? := pr }
else
return .continue
builtin_simproc [bv_normalize] eqToBEq (((_ : Bool) = (_ : Bool))) := fun e => do
let_expr Eq _ lhs rhs := e | return .continue
match_expr rhs with
@@ -127,8 +139,6 @@ builtin_simproc [bv_normalize] bv_add_const' (((_ : BitVec _) + (_ : BitVec _))
else
return .continue
attribute [builtin_bv_normalize_proc] reduceIte
/-- Return a number `k` such that `2^k = n`. -/
private def Nat.log2Exact (n : Nat) : Option Nat := do
guard <| n 0

View File

@@ -0,0 +1,69 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Tactics
import Lean.Meta.Tactic.Grind
import Lean.Elab.Command
import Lean.Elab.Tactic.Basic
import Lean.Elab.Tactic.Config
namespace Lean.Elab.Tactic
open Meta
declare_config_elab elabGrindConfig Grind.Config
open Command Term in
@[builtin_command_elab Lean.Parser.Command.grindPattern]
def elabGrindPattern : CommandElab := fun stx => do
match stx with
| `(grind_pattern $thmName:ident => $terms,*) => do
liftTermElabM do
let declName resolveGlobalConstNoOverload thmName
discard <| addTermInfo thmName ( mkConstWithLevelParams declName)
let info getConstInfo declName
forallTelescope info.type fun xs _ => do
let patterns terms.getElems.mapM fun term => do
let pattern elabTerm term none
synthesizeSyntheticMVarsUsingDefault
let pattern instantiateMVars pattern
let pattern Grind.preprocessPattern pattern
return pattern.abstract xs
Grind.addEMatchTheorem declName xs.size patterns.toList
| _ => throwUnsupportedSyntax
def grind (mvarId : MVarId) (config : Grind.Config) (mainDeclName : Name) (fallback : Grind.Fallback) : MetaM Unit := do
let mvarIds Grind.main mvarId config mainDeclName fallback
unless mvarIds.isEmpty do
throwError "`grind` failed\n{goalsToMessageData mvarIds}"
private def elabFallback (fallback? : Option Term) : TermElabM (Grind.GoalM Unit) := do
let some fallback := fallback? | return (pure ())
let type := mkApp (mkConst ``Grind.GoalM) (mkConst ``Unit)
let value withLCtx {} {} do Term.elabTermAndSynthesize fallback type
let auxDeclName if let .const declName _ := value then
pure declName
else
let auxDeclName Term.mkAuxName `_grind_fallback
let decl := Declaration.defnDecl {
name := auxDeclName
levelParams := []
type, value, hints := .opaque, safety := .safe
}
addAndCompile decl
pure auxDeclName
unsafe evalConst (Grind.GoalM Unit) auxDeclName
@[builtin_tactic Lean.Parser.Tactic.grind] def evalApplyRfl : Tactic := fun stx => do
match stx with
| `(tactic| grind $config:optConfig $[on_failure $fallback?]?) =>
let fallback elabFallback fallback?
logWarningAt stx "The `grind` tactic is experimental and still under development. Avoid using it in production projects"
let declName := ( Term.getDeclName?).getD `_grind
let config elabGrindConfig config
withMainContext do liftMetaFinishingTactic (grind · config declName fallback)
| _ => throwUnsupportedSyntax
end Lean.Elab.Tactic

View File

@@ -579,12 +579,25 @@ private def checkAltsOfOptInductionAlts (optInductionAlts : Syntax) : TacticM Un
throwErrorAt alt "more than one wildcard alternative '| _ => ...' used"
found := true
def getInductiveValFromMajor (major : Expr) : TacticM InductiveVal :=
def getInductiveValFromMajor (induction : Bool) (major : Expr) : TacticM InductiveVal :=
liftMetaMAtMain fun mvarId => do
let majorType inferType major
let majorType whnf majorType
matchConstInduct majorType.getAppFn
(fun _ => Meta.throwTacticEx `induction mvarId m!"major premise type is not an inductive type {indentExpr majorType}")
(fun _ => do
let tacticName := if induction then `induction else `cases
let mut hint := m!"\n\nExplanation: the '{tacticName}' tactic is for constructor-based reasoning \
as well as for applying custom {tacticName} principles with a 'using' clause or a registered '@[{tacticName}_eliminator]' theorem. \
The above type neither is an inductive type nor has a registered theorem."
if majorType.isProp then
hint := m!"{hint}\n\n\
Consider using the 'by_cases' tactic, which does true/false reasoning for propositions."
else if majorType.isType then
hint := m!"{hint}\n\n\
Type universes are not inductive types, and type-constructor-based reasoning is not possible. \
This is a strong limitation. According to Lean's underlying theory, the only provable distinguishing \
feature of types is their cardinalities."
Meta.throwTacticEx tacticName mvarId m!"major premise type is not an inductive type{indentExpr majorType}{hint}")
(fun val _ => pure val)
/--
@@ -627,7 +640,7 @@ private def getElimNameInfo (optElimId : Syntax) (targets : Array Expr) (inducti
return getElimInfo elimName
unless targets.size == 1 do
throwError "eliminator must be provided when multiple targets are used (use 'using <eliminator-name>'), and no default eliminator has been registered using attribute `[eliminator]`"
let indVal getInductiveValFromMajor targets[0]!
let indVal getInductiveValFromMajor induction targets[0]!
if induction && indVal.all.length != 1 then
throwError "'induction' tactic does not support mutually inductive types, the eliminator '{mkRecName indVal.name}' has multiple motives"
if induction && indVal.isNested then

View File

@@ -0,0 +1,223 @@
/-
Copyright (c) 2024 Lean FRO, LLC. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Joachim Breitner
-/
prelude
import Lean.Meta.Tactic.Split
import Lean.Elab.RecAppSyntax
import Lean.Elab.Tactic.Basic
import Init.Internal.Order
namespace Lean.Meta.Monotonicity
open Lean Meta
open Lean.Order
partial def headBetaUnderLambda (f : Expr) : Expr := Id.run do
let mut f := f.headBeta
if f.isLambda then
while f.bindingBody!.isHeadBetaTarget do
f := f.updateLambda! f.bindingInfo! f.bindingDomain! f.bindingBody!.headBeta
return f
/-- Environment extensions for monotonicity lemmas -/
builtin_initialize monotoneExt :
SimpleScopedEnvExtension (Name × Array DiscrTree.Key) (DiscrTree Name)
registerSimpleScopedEnvExtension {
addEntry := fun dt (n, ks) => dt.insertCore ks n
initial := {}
}
builtin_initialize registerBuiltinAttribute {
name := `partial_fixpoint_monotone
descr := "monotonicity theorem"
add := fun decl _ kind => MetaM.run' do
let declTy := ( getConstInfo decl).type
let (xs, _, targetTy) withReducible <| forallMetaTelescopeReducing declTy
let_expr monotone α inst_α β inst_β f := targetTy |
throwError "@[partial_fixpoint_monotone] attribute only applies to lemmas proving {.ofConstName ``monotone}"
let f := f.headBeta
let f if f.isLambda then pure f else etaExpand f
let f := headBetaUnderLambda f
lambdaBoundedTelescope f 1 fun _ e => do
let key withReducible <| DiscrTree.mkPath e
monotoneExt.add (decl, key) kind
}
/--
Finds tagged monotonicity theorems of the form `monotone (fun x => e)`.
-/
def findMonoThms (e : Expr) : MetaM (Array Name) := do
(monotoneExt.getState ( getEnv)).getMatch e
private def defaultFailK (f : Expr) (monoThms : Array Name) : MetaM α :=
let extraMsg := if monoThms.isEmpty then m!"" else
m!"Tried to apply {.andList (monoThms.toList.map (m!"'{·}'"))}, but failed."
throwError "Failed to prove monotonicity of:{indentExpr f}\n{extraMsg}"
private def applyConst (goal : MVarId) (name : Name) : MetaM (List MVarId) := do
mapError (f := (m!"Could not apply {.ofConstName name}:{indentD ·}")) do
goal.applyConst name (cfg := { synthAssignedInstances := false})
/--
Base case for solveMonoStep: Handles goals of the form
```
monotone (fun f => f.1.2 x y)
```
It's tricky to solve them compositionally from the outside in, so here we construct the proof
from the inside out.
-/
partial def solveMonoCall (α inst_α : Expr) (e : Expr) : MetaM (Option Expr) := do
if e.isApp && !e.appArg!.hasLooseBVars then
let some hmono solveMonoCall α inst_α e.appFn! | return none
let hmonoType inferType hmono
let_expr monotone _ _ _ inst _ := hmonoType | throwError "solveMonoCall {e}: unexpected type {hmonoType}"
let some inst whnfUntil inst ``instOrderPi | throwError "solveMonoCall {e}: unexpected instance {inst}"
let_expr instOrderPi γ δ inst inst | throwError "solveMonoCall {e}: whnfUntil failed?{indentExpr inst}"
return mkAppOptM ``monotone_apply #[γ, δ, α, inst_α, inst, e.appArg!, none, hmono]
if e.isProj then
let some hmono solveMonoCall α inst_α e.projExpr! | return none
let hmonoType inferType hmono
let_expr monotone _ _ _ inst _ := hmonoType | throwError "solveMonoCall {e}: unexpected type {hmonoType}"
let some inst whnfUntil inst ``instPartialOrderPProd | throwError "solveMonoCall {e}: unexpected instance {inst}"
let_expr instPartialOrderPProd β γ inst_β inst_γ inst | throwError "solveMonoCall {e}: whnfUntil failed?{indentExpr inst}"
let n := if e.projIdx! == 0 then ``monotone_pprod_fst else ``monotone_pprod_snd
return mkAppOptM n #[β, γ, α, inst_β, inst_γ, inst_α, none, hmono]
if e == .bvar 0 then
let hmono mkAppOptM ``monotone_id #[α, inst_α]
return some hmono
return none
def solveMonoStep (failK : {α}, Expr Array Name MetaM α := @defaultFailK) (goal : MVarId) : MetaM (List MVarId) :=
goal.withContext do
trace[Elab.Tactic.monotonicity] "monotonicity at\n{goal}"
let type goal.getType
if type.isForall then
let (_, goal) goal.intro1P
return [goal]
match_expr type with
| monotone α inst_α β inst_β f =>
-- Ensure f is not headed by a redex and headed by at least one lambda, and clean some
-- redexes left by some of the lemmas we tend to apply
let f instantiateMVars f
let f := f.headBeta
let f if f.isLambda then pure f else etaExpand f
let f := headBetaUnderLambda f
let e := f.bindingBody!
-- No recursive calls left
if !e.hasLooseBVars then
return applyConst goal ``monotone_const
-- NB: `e` is now an open term.
-- Look through mdata
if e.isMData then
let f' := f.updateLambdaE! f.bindingDomain! e.mdataExpr!
let goal' mkFreshExprSyntheticOpaqueMVar (mkApp type.appFn! f')
goal.assign goal'
return [goal'.mvarId!]
-- Float letE to the environment
if let .letE n t v b _nonDep := e then
if t.hasLooseBVars || v.hasLooseBVars then
failK f #[]
let goal' withLetDecl n t v fun x => do
let b' := f.updateLambdaE! f.bindingDomain! (b.instantiate1 x)
let goal' mkFreshExprSyntheticOpaqueMVar (mkApp type.appFn! b')
goal.assign ( mkLetFVars #[x] goal')
pure goal'
return [goal'.mvarId!]
-- Float `letFun` to the environment.
-- `applyConst` tends to reduce the redex
match_expr e with
| letFun γ _ v b =>
if γ.hasLooseBVars || v.hasLooseBVars then
failK f #[]
let b' := f.updateLambdaE! f.bindingDomain! b
let p mkAppOptM ``monotone_letFun #[α, β, γ, inst_α, inst_β, v, b']
let new_goals mapError (f := (m!"Could not apply {p}:{indentD ·}")) do
goal.apply p
let [new_goal] := new_goals
| throwError "Unexpected number of goals after {.ofConstName ``monotone_letFun}."
let (_, new_goal)
if b.isLambda then
new_goal.intro b.bindingName!
else
new_goal.intro1
return [new_goal]
| _ => pure ()
-- Handle lambdas, preserving the name of the binder
if e.isLambda then
let [new_goal] applyConst goal ``monotone_of_monotone_apply
| throwError "Unexpected number of goals after {.ofConstName ``monotone_of_monotone_apply}."
let (_, new_goal) new_goal.intro e.bindingName!
return [new_goal]
-- A recursive call directly here
if e.isBVar then
return applyConst goal ``monotone_id
-- A recursive call
if let some hmono solveMonoCall α inst_α e then
trace[Elab.Tactic.monotonicity] "Found recursive call {e}:{indentExpr hmono}"
unless goal.checkedAssign hmono do
trace[Elab.Tactic.monotonicity] "Failed to assign {hmono} : {← inferType hmono} to goal"
failK f #[]
return []
let monoThms withLocalDeclD `f f.bindingDomain! fun f =>
-- The discrimination tree does not like open terms
findMonoThms (e.instantiate1 f)
trace[Elab.Tactic.monotonicity] "Found monoThms: {monoThms.map MessageData.ofConstName}"
for monoThm in monoThms do
let new_goals? try
let new_goals applyConst goal monoThm
trace[Elab.Tactic.monotonicity] "Succeeded with {.ofConstName monoThm}"
pure (some new_goals)
catch e =>
trace[Elab.Tactic.monotonicity] "{e.toMessageData}"
pure none
if let some new_goals := new_goals? then
return new_goals
-- Split match-expressions
if let some info := isMatcherAppCore? ( getEnv) e then
let candidate id do
let args := e.getAppArgs
for i in [info.getFirstDiscrPos : info.getFirstDiscrPos + info.numDiscrs] do
if args[i]!.hasLooseBVars then
return false
return true
if candidate then
-- We could be even more deliberate here and use the `lifter` lemmas
-- for the match statements instead of the `split` tactic.
-- For now using `splitMatch` works fine.
return Split.splitMatch goal e
failK f monoThms
| _ =>
throwError "Unexpected goal:{goal}"
partial def solveMono (failK : {α}, Expr Array Name MetaM α := defaultFailK) (goal : MVarId) : MetaM Unit := do
let new_goals solveMonoStep failK goal
new_goals.forM (solveMono failK)
open Elab Tactic in
@[builtin_tactic Lean.Order.monotonicity]
def evalMonotonicity : Tactic := fun _stx =>
liftMetaTactic Lean.Meta.Monotonicity.solveMonoStep
end Lean.Meta.Monotonicity
builtin_initialize Lean.registerTraceClass `Elab.Tactic.monotonicity

View File

@@ -63,7 +63,7 @@ def isNumeral? (e : Expr) : Option (Expr × Nat) :=
if e.isConstOf ``Nat.zero then
(mkConst ``Nat, 0)
else if let Expr.app (Expr.app (Expr.app (Expr.const ``OfNat.ofNat ..) α ..)
(Expr.lit (Literal.natVal n) ..) ..) .. := e then
(Expr.lit (Literal.natVal n) ..) ..) .. := e.consumeMData then
some (α, n)
else
none

View File

@@ -769,6 +769,11 @@ opaque quickLt (a : @& Expr) (b : @& Expr) : Bool
@[extern "lean_expr_lt"]
opaque lt (a : @& Expr) (b : @& Expr) : Bool
def quickComp (a b : Expr) : Ordering :=
if quickLt a b then .lt
else if quickLt b a then .gt
else .eq
/--
Return true iff `a` and `b` are alpha equivalent.
Binder annotations are ignored.
@@ -1643,6 +1648,23 @@ def isFalse (e : Expr) : Bool :=
def isTrue (e : Expr) : Bool :=
e.cleanupAnnotations.isConstOf ``True
/--
`getForallArity type` returns the arity of a `forall`-type. This function consumes nested annotations,
and performs pending beta reductions. It does **not** use whnf.
Examples:
- If `a` is `Nat`, `getForallArity a` returns `0`
- If `a` is `Nat → Bool`, `getForallArity a` returns `1`
-/
partial def getForallArity : Expr Nat
| .mdata _ b => getForallArity b
| .forallE _ _ b _ => getForallArity b + 1
| e =>
if e.isHeadBetaTarget then
getForallArity e.headBeta
else
let e' := e.cleanupAnnotations
if e != e' then getForallArity e' else 0
/--
Checks if an expression is a "natural number numeral in normal form",
i.e. of type `Nat`, and explicitly of the form `OfNat.ofNat n`

View File

@@ -4,6 +4,7 @@ Released under Apache 2.0 license as described in the file LICENSE.
Authors: Mario Carneiro
-/
prelude
import Lean.Parser.Syntax
import Lean.Meta.Tactic.Simp.RegisterCommand
import Lean.Elab.Command
import Lean.Elab.SetOption

View File

@@ -176,6 +176,14 @@ def mkEqOfHEq (h : Expr) : MetaM Expr := do
| _ =>
throwAppBuilderException ``eq_of_heq m!"heterogeneous equality proof expected{indentExpr h}"
/-- Given `h : Eq a b`, returns a proof of `HEq a b`. -/
def mkHEqOfEq (h : Expr) : MetaM Expr := do
let hType infer h
let some (α, a, b) := hType.eq?
| throwAppBuilderException ``heq_of_eq m!"equality proof expected{indentExpr h}"
let u getLevel α
return mkApp4 (mkConst ``heq_of_eq [u]) α a b h
/--
If `e` is `@Eq.refl α a`, return `a`.
-/

View File

@@ -605,6 +605,9 @@ variable [MonadControlT MetaM n] [Monad n]
@[inline] def modifyCache (f : Cache Cache) : MetaM Unit :=
modify fun { mctx, cache, zetaDeltaFVarIds, postponed, diag } => { mctx, cache := f cache, zetaDeltaFVarIds, postponed, diag }
def resetCache : MetaM Unit :=
modifyCache fun _ => {}
@[inline] def modifyInferTypeCache (f : InferTypeCache InferTypeCache) : MetaM Unit :=
modifyCache fun ic, c1, c2, c3, c4, c5 => f ic, c1, c2, c3, c4, c5
@@ -1777,9 +1780,9 @@ private def withMVarContextImp (mvarId : MVarId) (x : MetaM α) : MetaM α := do
withLocalContextImp mvarDecl.lctx mvarDecl.localInstances x
/--
Execute `x` using the given metavariable `LocalContext` and `LocalInstances`.
The type class resolution cache is flushed when executing `x` if its `LocalInstances` are
different from the current ones. -/
Executes `x` using the given metavariable `LocalContext` and `LocalInstances`.
The type class resolution cache is flushed when executing `x` if its `LocalInstances` are
different from the current ones. -/
def _root_.Lean.MVarId.withContext (mvarId : MVarId) : n α n α :=
mapMetaM <| withMVarContextImp mvarId
@@ -1789,13 +1792,25 @@ private def withMCtxImp (mctx : MetavarContext) (x : MetaM α) : MetaM α := do
try x finally setMCtx mctx'
/--
`withMCtx mctx k` replaces the metavariable context and then executes `k`.
The metavariable context is restored after executing `k`.
`withMCtx mctx k` replaces the metavariable context and then executes `k`.
The metavariable context is restored after executing `k`.
This method is used to implement the type class resolution procedure. -/
This method is used to implement the type class resolution procedure. -/
def withMCtx (mctx : MetavarContext) : n α n α :=
mapMetaM <| withMCtxImp mctx
/--
`withoutModifyingMCtx k` executes `k` and then restores the metavariable context.
-/
def withoutModifyingMCtx : n α n α :=
mapMetaM fun x => do
let mctx getMCtx
try
x
finally
resetCache
setMCtx mctx
@[inline] private def approxDefEqImp (x : MetaM α) : MetaM α :=
withConfig (fun config => { config with foApprox := true, ctxApprox := true, quasiPatternApprox := true}) x

View File

@@ -17,14 +17,6 @@ namespace Lean.Meta
private def ensureType (e : Expr) : MetaM Unit := do
discard <| getLevel e
def throwLetTypeMismatchMessage {α} (fvarId : FVarId) : MetaM α := do
let lctx getLCtx
match lctx.find? fvarId with
| some (LocalDecl.ldecl _ _ _ t v _ _) => do
let vType inferType v
throwError "invalid let declaration, term{indentExpr v}\nhas type{indentExpr vType}\nbut is expected to have type{indentExpr t}"
| _ => unreachable!
private def checkConstant (constName : Name) (us : List Level) : MetaM Unit := do
let cinfo getConstInfo constName
unless us.length == cinfo.levelParams.length do
@@ -177,6 +169,15 @@ where
catch _ =>
return (a, b)
def throwLetTypeMismatchMessage {α} (fvarId : FVarId) : MetaM α := do
let lctx getLCtx
match lctx.find? fvarId with
| some (LocalDecl.ldecl _ _ _ t v _ _) => do
let vType inferType v
let (vType, t) addPPExplicitToExposeDiff vType t
throwError "invalid let declaration, term{indentExpr v}\nhas type{indentExpr vType}\nbut is expected to have type{indentExpr t}"
| _ => unreachable!
/--
Return error message "has type{givenType}\nbut is expected to have type{expectedType}"
-/

View File

@@ -167,7 +167,7 @@ def isDefEqStringLit (s t : Expr) : MetaM LBool := do
Remark: `n` may be 0. -/
def isEtaUnassignedMVar (e : Expr) : MetaM Bool := do
match e.etaExpanded? with
| some (Expr.mvar mvarId) =>
| some (.mvar mvarId) =>
if ( mvarId.isReadOnlyOrSyntheticOpaque) then
pure false
else if ( mvarId.isAssigned) then
@@ -361,9 +361,9 @@ private partial def isDefEqBindingAux (lctx : LocalContext) (fvars : Array Expr)
let fvars := fvars.push (mkFVar fvarId)
isDefEqBindingAux lctx fvars b₁ b₂ (ds₂.push d₂)
match e₁, e₂ with
| Expr.forallE n d₁ b₁ _, Expr.forallE _ d₂ b₂ _ => process n d₁ d₂ b₁ b₂
| Expr.lam n d₁ b₁ _, Expr.lam _ d₂ b₂ _ => process n d₁ d₂ b₁ b₂
| _, _ =>
| .forallE n d₁ b₁ _, .forallE _ d₂ b₂ _ => process n d₁ d₂ b₁ b₂
| .lam n d₁ b₁ _, .lam _ d₂ b₂ _ => process n d₁ d₂ b₁ b₂
| _, _ =>
withLCtx' lctx do
isDefEqBindingDomain fvars ds₂ do
Meta.isExprDefEqAux (e₁.instantiateRev fvars) (e₂.instantiateRev fvars)
@@ -1037,13 +1037,13 @@ def checkedAssignImpl (mvarId : MVarId) (val : Expr) : MetaM Bool := do
private def processAssignmentFOApproxAux (mvar : Expr) (args : Array Expr) (v : Expr) : MetaM Bool :=
match v with
| .mdata _ e => processAssignmentFOApproxAux mvar args e
| Expr.app f a =>
| .mdata _ e => processAssignmentFOApproxAux mvar args e
| .app f a =>
if args.isEmpty then
pure false
else
Meta.isExprDefEqAux args.back! a <&&> Meta.isExprDefEqAux (mkAppRange mvar 0 (args.size - 1) args) f
| _ => pure false
| _ => pure false
/--
Auxiliary method for applying first-order unification. It is an approximation.
@@ -1177,7 +1177,7 @@ private partial def processAssignment (mvarApp : Expr) (v : Expr) : MetaM Bool :
let arg simpAssignmentArg arg
let args := args.set i arg
match arg with
| Expr.fvar fvarId =>
| .fvar fvarId =>
if args[0:i].any fun prevArg => prevArg == arg then
useFOApprox args
else if mvarDecl.lctx.contains fvarId && !cfg.quasiPatternApprox then
@@ -1233,7 +1233,7 @@ private def processAssignment' (mvarApp : Expr) (v : Expr) : MetaM Bool := do
private def isDeltaCandidate? (t : Expr) : MetaM (Option ConstantInfo) := do
match t.getAppFn with
| Expr.const c _ =>
| .const c _ =>
match ( getUnfoldableConst? c) with
| r@(some info) => if info.hasValue then return r else return none
| _ => return none
@@ -1375,8 +1375,8 @@ private def unfoldBothDefEq (fn : Name) (t s : Expr) : MetaM LBool := do
private def sameHeadSymbol (t s : Expr) : Bool :=
match t.getAppFn, s.getAppFn with
| Expr.const c₁ _, Expr.const c₂ _ => c₁ == c₂
| _, _ => false
| .const c₁ _, .const c₂ _ => c₁ == c₂
| _, _ => false
/--
- If headSymbol (unfold t) == headSymbol s, then unfold t
@@ -1521,8 +1521,8 @@ private def isDefEqDelta (t s : Expr) : MetaM LBool := do
unfoldNonProjFnDefEq tInfo sInfo t s
private def isAssigned : Expr MetaM Bool
| Expr.mvar mvarId => mvarId.isAssigned
| _ => pure false
| .mvar mvarId => mvarId.isAssigned
| _ => pure false
private def expandDelayedAssigned? (t : Expr) : MetaM (Option Expr) := do
let tFn := t.getAppFn
@@ -1647,8 +1647,8 @@ private partial def isDefEqQuick (t s : Expr) : MetaM LBool :=
| .sort u, .sort v => toLBoolM <| isLevelDefEqAux u v
| .lam .., .lam .. => if t == s then pure LBool.true else toLBoolM <| isDefEqBinding t s
| .forallE .., .forallE .. => if t == s then pure LBool.true else toLBoolM <| isDefEqBinding t s
-- | Expr.mdata _ t _, s => isDefEqQuick t s
-- | t, Expr.mdata _ s _ => isDefEqQuick t s
-- | .mdata _ t _, s => isDefEqQuick t s
-- | t, .mdata _ s _ => isDefEqQuick t s
| .fvar fvarId₁, .fvar fvarId₂ => do
if fvarId₁ == fvarId₂ then
return .true

View File

@@ -66,30 +66,31 @@ structure GeneralizeIndicesSubgoal where
numEqs : Nat
/--
Similar to `generalizeTargets` but customized for the `casesOn` motive.
Given a metavariable `mvarId` representing the
```
Ctx, h : I A j, D |- T
```
where `fvarId` is `h`s id, and the type `I A j` is an inductive datatype where `A` are parameters,
and `j` the indices. Generate the goal
```
Ctx, h : I A j, D, j' : J, h' : I A j' |- j == j' -> h == h' -> T
```
Remark: `(j == j' -> h == h')` is a "telescopic" equality.
Remark: `j` is sequence of terms, and `j'` a sequence of free variables.
The result contains the fields
- `mvarId`: the new goal
- `indicesFVarIds`: `j'` ids
- `fvarId`: `h'` id
- `numEqs`: number of equations in the target -/
def generalizeIndices (mvarId : MVarId) (fvarId : FVarId) : MetaM GeneralizeIndicesSubgoal :=
Given a metavariable `mvarId` representing the goal
```
Ctx |- T
```
and an expression `e : I A j`, where `I A j` is an inductive datatype where `A` are parameters,
and `j` the indices. Generate the goal
```
Ctx, j' : J, h' : I A j' |- j == j' -> e == h' -> T
```
Remark: `(j == j' -> e == h')` is a "telescopic" equality.
Remark: `j` is sequence of terms, and `j'` a sequence of free variables.
The result contains the fields
- `mvarId`: the new goal
- `indicesFVarIds`: `j'` ids
- `fvarId`: `h'` id
- `numEqs`: number of equations in the target
If `varName?` is not none, it is used to name `h'`.
-/
def generalizeIndices' (mvarId : MVarId) (e : Expr) (varName? : Option Name := none) : MetaM GeneralizeIndicesSubgoal :=
mvarId.withContext do
let lctx getLCtx
let localInsts getLocalInstances
mvarId.checkNotAssigned `generalizeIndices
let fvarDecl fvarId.getDecl
let type whnf fvarDecl.type
let type whnfD ( inferType e)
type.withApp fun f args => matchConstInduct f (fun _ => throwTacticEx `generalizeIndices mvarId "inductive type expected") fun val _ => do
unless val.numIndices > 0 do throwTacticEx `generalizeIndices mvarId "indexed inductive type expected"
unless args.size == val.numIndices + val.numParams do throwTacticEx `generalizeIndices mvarId "ill-formed inductive datatype"
@@ -98,9 +99,10 @@ def generalizeIndices (mvarId : MVarId) (fvarId : FVarId) : MetaM GeneralizeIndi
let IAType inferType IA
forallTelescopeReducing IAType fun newIndices _ => do
let newType := mkAppN IA newIndices
withLocalDeclD fvarDecl.userName newType fun h' =>
let varName if let some varName := varName? then pure varName else mkFreshUserName `x
withLocalDeclD varName newType fun h' =>
withNewEqs indices newIndices fun newEqs newRefls => do
let (newEqType, newRefl) mkEqAndProof fvarDecl.toExpr h'
let (newEqType, newRefl) mkEqAndProof e h'
let newRefls := newRefls.push newRefl
withLocalDeclD `h newEqType fun newEq => do
let newEqs := newEqs.push newEq
@@ -112,7 +114,7 @@ def generalizeIndices (mvarId : MVarId) (fvarId : FVarId) : MetaM GeneralizeIndi
let auxType mkForallFVars newIndices auxType
let newMVar mkFreshExprMVarAt lctx localInsts auxType MetavarKind.syntheticOpaque tag
/- assign mvarId := newMVar indices h refls -/
mvarId.assign (mkAppN (mkApp (mkAppN newMVar indices) fvarDecl.toExpr) newRefls)
mvarId.assign (mkAppN (mkApp (mkAppN newMVar indices) e) newRefls)
let (indicesFVarIds, newMVarId) newMVar.mvarId!.introNP newIndices.size
let (fvarId, newMVarId) newMVarId.intro1P
return {
@@ -122,6 +124,29 @@ def generalizeIndices (mvarId : MVarId) (fvarId : FVarId) : MetaM GeneralizeIndi
numEqs := newEqs.size
}
/--
Similar to `generalizeTargets` but customized for the `casesOn` motive.
Given a metavariable `mvarId` representing the
```
Ctx, h : I A j, D |- T
```
where `fvarId` is `h`s id, and the type `I A j` is an inductive datatype where `A` are parameters,
and `j` the indices. Generate the goal
```
Ctx, h : I A j, D, j' : J, h' : I A j' |- j == j' -> h == h' -> T
```
Remark: `(j == j' -> h == h')` is a "telescopic" equality.
Remark: `j` is sequence of terms, and `j'` a sequence of free variables.
The result contains the fields
- `mvarId`: the new goal
- `indicesFVarIds`: `j'` ids
- `fvarId`: `h'` id
- `numEqs`: number of equations in the target -/
def generalizeIndices (mvarId : MVarId) (fvarId : FVarId) : MetaM GeneralizeIndicesSubgoal :=
mvarId.withContext do
let fvarDecl fvarId.getDecl
generalizeIndices' mvarId fvarDecl.toExpr fvarDecl.userName
structure CasesSubgoal extends InductionSubgoal where
ctorName : Name

View File

@@ -7,19 +7,47 @@ prelude
import Lean.Meta.Tactic.Grind.Attr
import Lean.Meta.Tactic.Grind.RevertAll
import Lean.Meta.Tactic.Grind.Types
import Lean.Meta.Tactic.Grind.Preprocessor
import Lean.Meta.Tactic.Grind.Util
import Lean.Meta.Tactic.Grind.Cases
import Lean.Meta.Tactic.Grind.Injection
import Lean.Meta.Tactic.Grind.Core
import Lean.Meta.Tactic.Grind.Canon
import Lean.Meta.Tactic.Grind.MarkNestedProofs
import Lean.Meta.Tactic.Grind.Inv
import Lean.Meta.Tactic.Grind.Proof
import Lean.Meta.Tactic.Grind.Propagate
import Lean.Meta.Tactic.Grind.PP
import Lean.Meta.Tactic.Grind.Simp
import Lean.Meta.Tactic.Grind.Ctor
import Lean.Meta.Tactic.Grind.Parser
import Lean.Meta.Tactic.Grind.EMatchTheorem
import Lean.Meta.Tactic.Grind.EMatch
import Lean.Meta.Tactic.Grind.Main
namespace Lean
/-! Trace options for `grind` users -/
builtin_initialize registerTraceClass `grind
builtin_initialize registerTraceClass `grind.eq
builtin_initialize registerTraceClass `grind.assert
builtin_initialize registerTraceClass `grind.eqc
builtin_initialize registerTraceClass `grind.internalize
builtin_initialize registerTraceClass `grind.ematch
builtin_initialize registerTraceClass `grind.ematch.pattern
builtin_initialize registerTraceClass `grind.ematch.pattern.search
builtin_initialize registerTraceClass `grind.ematch.instance
builtin_initialize registerTraceClass `grind.ematch.instance.assignment
builtin_initialize registerTraceClass `grind.issues
builtin_initialize registerTraceClass `grind.add
builtin_initialize registerTraceClass `grind.pre
builtin_initialize registerTraceClass `grind.simp
/-! Trace options for `grind` developers -/
builtin_initialize registerTraceClass `grind.debug
builtin_initialize registerTraceClass `grind.debug.proofs
builtin_initialize registerTraceClass `grind.debug.congr
builtin_initialize registerTraceClass `grind.debug.proof
builtin_initialize registerTraceClass `grind.debug.proj
builtin_initialize registerTraceClass `grind.debug.parent
builtin_initialize registerTraceClass `grind.debug.final
builtin_initialize registerTraceClass `grind.debug.forallPropagator
end Lean

View File

@@ -0,0 +1,159 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Util
import Lean.Meta.Basic
import Lean.Meta.FunInfo
import Lean.Util.FVarSubset
import Lean.Util.PtrSet
import Lean.Util.FVarSubset
namespace Lean.Meta.Grind
namespace Canon
/-!
A canonicalizer module for the `grind` tactic. The canonicalizer defined in `Meta/Canonicalizer.lean` is
not suitable for the `grind` tactic. It was designed for tactics such as `omega`, where the goal is
to detect when two structurally different atoms are definitionally equal.
The `grind` tactic, on the other hand, uses congruence closure. Moreover, types, type formers, proofs, and instances
are considered supporting elements and are not factored into congruence detection.
This module minimizes the number of `isDefEq` checks by comparing two terms `a` and `b` only if they instances,
types, or type formers and are the `i`-th arguments of two different `f`-applications. This approach is
sufficient for the congruence closure procedure used by the `grind` tactic.
To further optimize `isDefEq` checks, instances are compared using `TransparencyMode.instances`, which reduces
the number of constants that need to be unfolded. If diagnostics are enabled, instances are compared using
the default transparency mode too for sanity checking, and discrepancies are reported.
Types and type formers are always checked using default transparency.
Remark:
The canonicalizer minimizes issues with non-canonical instances and structurally different but definitionally equal types,
but it does not solve all problems. For example, consider a situation where we have `(a : BitVec n)`
and `(b : BitVec m)`, along with instances `inst1 n : Add (BitVec n)` and `inst2 m : Add (BitVec m)` where `inst1`
and `inst2` are structurally different. Now consider the terms `a + a` and `b + b`. After canonicalization, the two
additions will still use structurally different (and definitionally different) instances: `inst1 n` and `inst2 m`.
Furthermore, `grind` will not be able to infer that `HEq (a + a) (b + b)` even if we add the assumptions `n = m` and `HEq a b`.
-/
structure State where
argMap : PHashMap (Expr × Nat) (List Expr) := {}
canon : PHashMap Expr Expr := {}
proofCanon : PHashMap Expr Expr := {}
deriving Inhabited
/--
Helper function for canonicalizing `e` occurring as the `i`th argument of an `f`-application.
`isInst` is true if `e` is an type class instance.
Recall that we use `TransparencyMode.instances` for checking whether two instances are definitionally equal or not.
Thus, if diagnostics are enabled, we also check them using `TransparencyMode.default`. If the result is different
we report to the user.
-/
def canonElemCore (f : Expr) (i : Nat) (e : Expr) (isInst : Bool) : StateT State MetaM Expr := do
let s get
if let some c := s.canon.find? e then
return c
let key := (f, i)
let cs := s.argMap.find? key |>.getD []
for c in cs do
if ( isDefEq e c) then
-- We used to check `c.fvarsSubset e` because it is not
-- in general safe to replace `e` with `c` if `c` has more free variables than `e`.
-- However, we don't revert previously canonicalized elements in the `grind` tactic.
modify fun s => { s with canon := s.canon.insert e c }
return c
if isInst then
if ( isDiagnosticsEnabled <&&> pure (c.fvarsSubset e) <&&> (withDefault <| isDefEq e c)) then
-- TODO: consider storing this information in some structure that can be browsed later.
trace[grind.issues] "the following `grind` static elements are definitionally equal with `default` transparency, but not with `instances` transparency{indentExpr e}\nand{indentExpr c}"
modify fun s => { s with canon := s.canon.insert e e, argMap := s.argMap.insert key (e::cs) }
return e
abbrev canonType (f : Expr) (i : Nat) (e : Expr) := withDefault <| canonElemCore f i e false
abbrev canonInst (f : Expr) (i : Nat) (e : Expr) := withReducibleAndInstances <| canonElemCore f i e true
/--
Return type for the `shouldCanon` function.
-/
private inductive ShouldCanonResult where
| /- Nested types (and type formers) are canonicalized. -/
canonType
| /- Nested instances are canonicalized. -/
canonInst
| /-
Term is not a proof, type (former), nor an instance.
Thus, it must be recursively visited by the canonizer.
-/
visit
deriving Inhabited
/--
See comments at `ShouldCanonResult`.
-/
def shouldCanon (pinfos : Array ParamInfo) (i : Nat) (arg : Expr) : MetaM ShouldCanonResult := do
if h : i < pinfos.size then
let pinfo := pinfos[i]
if pinfo.isInstImplicit then
return .canonInst
else if pinfo.isProp then
return .visit
if ( isTypeFormer arg) then
return .canonType
else
return .visit
unsafe def canonImpl (e : Expr) : StateT State MetaM Expr := do
visit e |>.run' mkPtrMap
where
visit (e : Expr) : StateRefT (PtrMap Expr Expr) (StateT State MetaM) Expr := do
unless e.isApp || e.isForall do return e
-- Check whether it is cached
if let some r := ( get).find? e then
return r
let e' match e with
| .app .. => e.withApp fun f args => do
if f.isConstOf ``Lean.Grind.nestedProof && args.size == 2 then
let prop := args[0]!
let prop' visit prop
if let some r := ( getThe State).proofCanon.find? prop' then
pure r
else
let e' := if ptrEq prop prop' then e else mkAppN f (args.set! 0 prop')
modifyThe State fun s => { s with proofCanon := s.proofCanon.insert prop' e' }
pure e'
else
let pinfos := ( getFunInfo f).paramInfo
let mut modified := false
let mut args := args
for i in [:args.size] do
let arg := args[i]!
let arg' match ( shouldCanon pinfos i arg) with
| .canonType => canonType f i arg
| .canonInst => canonInst f i arg
| .visit => visit arg
unless ptrEq arg arg' do
args := args.set! i arg'
modified := true
pure <| if modified then mkAppN f args else e
| .forallE _ d b _ =>
-- Recall that we have `ForallProp.lean`.
let d' visit d
-- Remark: users may not want to convert `p → q` into `¬p q`
let b' if b.hasLooseBVars then pure b else visit b
pure <| e.updateForallE! d' b'
| _ => unreachable!
modify fun s => s.insert e e'
return e'
/-- Canonicalizes nested types, type formers, and instances in `e`. -/
def canon (e : Expr) : StateT State MetaM Expr :=
unsafe canonImpl e
end Canon
end Lean.Meta.Grind

View File

@@ -11,28 +11,29 @@ namespace Lean.Meta.Grind
The `grind` tactic includes an auxiliary `cases` tactic that is not intended for direct use by users.
This method implements it.
This tactic is automatically applied when introducing local declarations with a type tagged with `[grind_cases]`.
It is also used for "case-splitting" on terms during the search.
It differs from the user-facing Lean `cases` tactic in the following ways:
- It avoids unnecessary `revert` and `intro` operations.
- It does not introduce new local declarations for each minor premise. Instead, the `grind` tactic preprocessor is responsible for introducing them.
- It assumes that the major premise (i.e., the parameter `fvarId`) is the latest local declaration in the current goal.
- If the major premise type is an indexed family, auxiliary declarations and (heterogeneous) equalities are introduced.
However, these equalities are not resolved using `unifyEqs`. Instead, the `grind` tactic employs union-find and
congruence closure to process these auxiliary equalities. This approach avoids applying substitution to propositions
that have already been internalized by `grind`.
-/
def cases (mvarId : MVarId) (fvarId : FVarId) : MetaM (List MVarId) := mvarId.withContext do
def cases (mvarId : MVarId) (e : Expr) : MetaM (List MVarId) := mvarId.withContext do
let tag mvarId.getTag
let type whnf ( fvarId.getType)
let type whnf ( inferType e)
let .const declName _ := type.getAppFn | throwInductiveExpected type
let .inductInfo _ getConstInfo declName | throwInductiveExpected type
let recursorInfo mkRecursorInfo (mkCasesOnName declName)
let k (mvarId : MVarId) (fvarId : FVarId) (indices : Array Expr) (clearMajor : Bool) : MetaM (List MVarId) := do
let recursor mkRecursorAppPrefix mvarId `grind.cases fvarId recursorInfo indices
let mut recursor := mkApp (mkAppN recursor indices) (mkFVar fvarId)
let k (mvarId : MVarId) (fvarId : FVarId) (indices : Array FVarId) : MetaM (List MVarId) := do
let indicesExpr := indices.map mkFVar
let recursor mkRecursorAppPrefix mvarId `grind.cases fvarId recursorInfo indicesExpr
let mut recursor := mkApp (mkAppN recursor indicesExpr) (mkFVar fvarId)
let mut recursorType inferType recursor
let mut mvarIdsNew := #[]
for _ in [:recursorInfo.numMinors] do
@@ -41,22 +42,22 @@ def cases (mvarId : MVarId) (fvarId : FVarId) : MetaM (List MVarId) := mvarId.wi
recursorType := recursorTypeNew
let mvar mkFreshExprSyntheticOpaqueMVar targetNew tag
recursor := mkApp recursor mvar
let mvarIdNew if clearMajor then
mvar.mvarId!.clear fvarId
else
pure mvar.mvarId!
let mvarIdNew mvar.mvarId!.tryClearMany (indices.push fvarId)
mvarIdsNew := mvarIdsNew.push mvarIdNew
mvarId.assign recursor
return mvarIdsNew.toList
if recursorInfo.numIndices > 0 then
let s generalizeIndices mvarId fvarId
let s generalizeIndices' mvarId e
s.mvarId.withContext do
k s.mvarId s.fvarId (s.indicesFVarIds.map mkFVar) (clearMajor := false)
k s.mvarId s.fvarId s.indicesFVarIds
else if let .fvar fvarId := e then
k mvarId fvarId #[]
else
let indices getMajorTypeIndices mvarId `grind.cases recursorInfo type
k mvarId fvarId indices (clearMajor := true)
let mvarId mvarId.assert ( mkFreshUserName `x) type e
let (fvarId, mvarId) mvarId.intro1
mvarId.withContext do k mvarId fvarId #[]
where
throwInductiveExpected {α} (type : Expr) : MetaM α := do
throwTacticEx `grind.cases mvarId m!"(non-recursive) inductive type expected at {mkFVar fvarId}{indentExpr type}"
throwTacticEx `grind.cases mvarId m!"(non-recursive) inductive type expected at {e}{indentExpr type}"
end Lean.Meta.Grind

View File

@@ -4,144 +4,15 @@ Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Lean.Meta.Tactic.Grind.Types
import Init.Grind.Util
import Lean.Meta.LitValues
import Lean.Meta.Tactic.Grind.Types
import Lean.Meta.Tactic.Grind.Inv
import Lean.Meta.Tactic.Grind.PP
import Lean.Meta.Tactic.Grind.Ctor
import Lean.Meta.Tactic.Grind.Internalize
namespace Lean.Meta.Grind
/-- Helper function for pretty printing the state for debugging purposes. -/
def ppENodeRef (e : Expr) : GoalM Format := do
let some n getENode? e | return "_"
return f!"#{n.idx}"
/-- Returns expressions in the given expression equivalence class. -/
partial def getEqc (e : Expr) : GoalM (List Expr) :=
go e e []
where
go (first : Expr) (e : Expr) (acc : List Expr) : GoalM (List Expr) := do
let next getNext e
let acc := e :: acc
if isSameExpr first next then
return acc
else
go first next acc
/-- Returns all equivalence classes in the current goal. -/
partial def getEqcs : GoalM (List (List Expr)) := do
let mut r := []
for (_, node) in ( get).enodes do
if isSameExpr node.root node.self then
r := ( getEqc node.self) :: r
return r
/-- Helper function for pretty printing the state for debugging purposes. -/
def ppENodeDeclValue (e : Expr) : GoalM Format := do
if e.isApp && !( isLitValue e) then
e.withApp fun f args => do
let r if f.isConst then
ppExpr f
else
ppENodeRef f
let mut r := r
for arg in args do
r := r ++ " " ++ ( ppENodeRef arg)
return r
else
ppExpr e
/-- Helper function for pretty printing the state for debugging purposes. -/
def ppENodeDecl (e : Expr) : GoalM Format := do
let mut r := f!"{← ppENodeRef e} := {← ppENodeDeclValue e}"
let n getENode e
unless isSameExpr e n.root do
r := r ++ f!" ↦ {← ppENodeRef n.root}"
if n.interpreted then
r := r ++ ", [val]"
if n.ctor then
r := r ++ ", [ctor]"
return r
/-- Pretty print goal state for debugging purposes. -/
def ppState : GoalM Format := do
let mut r := f!"Goal:"
let nodes := ( get).enodes.toArray.map (·.2)
let nodes := nodes.qsort fun a b => a.idx < b.idx
for node in nodes do
r := r ++ "\n" ++ ( ppENodeDecl node.self)
let eqcs getEqcs
for eqc in eqcs do
if eqc.length > 1 then
r := r ++ "\n" ++ "{" ++ (Format.joinSep ( eqc.mapM ppENodeRef) ", ") ++ "}"
return r
/--
Returns `true` if `e` is `True`, `False`, or a literal value.
See `LitValues` for supported literals.
-/
def isInterpreted (e : Expr) : MetaM Bool := do
if e.isTrue || e.isFalse then return true
isLitValue e
/--
Creates an `ENode` for `e` if one does not already exist.
This method assumes `e` has been hashconsed.
-/
def mkENode (e : Expr) (generation : Nat) : GoalM Unit := do
if ( alreadyInternalized e) then return ()
let ctor := ( isConstructorAppCore? e).isSome
let interpreted isInterpreted e
mkENodeCore e interpreted ctor generation
private def pushNewEqCore (lhs rhs proof : Expr) (isHEq : Bool) : GoalM Unit :=
modify fun s => { s with newEqs := s.newEqs.push { lhs, rhs, proof, isHEq } }
@[inline] private def pushNewEq (lhs rhs proof : Expr) : GoalM Unit :=
pushNewEqCore lhs rhs proof (isHEq := false)
@[inline] private def pushNewHEq (lhs rhs proof : Expr) : GoalM Unit :=
pushNewEqCore lhs rhs proof (isHEq := true)
/--
Adds `e` to congruence table.
-/
def addCongrTable (_e : Expr) : GoalM Unit := do
-- TODO
return ()
partial def internalize (e : Expr) (generation : Nat) : GoalM Unit := do
if ( alreadyInternalized e) then return ()
match e with
| .bvar .. => unreachable!
| .sort .. => return ()
| .fvar .. | .letE .. | .lam .. | .forallE .. =>
mkENodeCore e (ctor := false) (interpreted := false) (generation := generation)
| .lit .. | .const .. =>
mkENode e generation
| .mvar ..
| .mdata ..
| .proj .. =>
trace[grind.issues] "unexpected term during internalization{indentExpr e}"
mkENodeCore e (ctor := false) (interpreted := false) (generation := generation)
| .app .. =>
if ( isLitValue e) then
-- We do not want to internalize the components of a literal value.
mkENode e generation
else e.withApp fun f args => do
unless f.isConst do
internalize f generation
let info getFunInfo f
let shouldInternalize (i : Nat) : GoalM Bool := do
if h : i < info.paramInfo.size then
let pinfo := info.paramInfo[i]
if pinfo.binderInfo.isInstImplicit || pinfo.isProp then
return false
return true
for h : i in [: args.size] do
let arg := args[i]
if ( shouldInternalize i) then
unless ( isTypeFormer arg) do
internalize arg generation
mkENode e generation
addCongrTable e
/--
The fields `target?` and `proof?` in `e`'s `ENode` are encoding a transitivity proof
@@ -163,39 +34,95 @@ where
proof? := proofNew?
}
private def markAsInconsistent : GoalM Unit :=
modify fun s => { s with inconsistent := true }
/--
Remove `root` parents from the congruence table.
This is an auxiliary function performed while merging equivalence classes.
-/
private def removeParents (root : Expr) : GoalM ParentSet := do
let parents getParentsAndReset root
for parent in parents do
-- Recall that we may have `Expr.forallE` in `parents` because of `ForallProp.lean`
if ( pure parent.isApp <&&> isCongrRoot parent) then
trace[grind.debug.parent] "remove: {parent}"
modify fun s => { s with congrTable := s.congrTable.erase { e := parent } }
return parents
def isInconsistent : GoalM Bool :=
return ( get).inconsistent
/--
Reinsert parents into the congruence table and detect new equalities.
This is an auxiliary function performed while merging equivalence classes.
-/
private def reinsertParents (parents : ParentSet) : GoalM Unit := do
for parent in parents do
if ( pure parent.isApp <&&> isCongrRoot parent) then
trace[grind.debug.parent] "reinsert: {parent}"
addCongrTable parent
/-- Closes the goal when `True` and `False` are in the same equivalence class. -/
private def closeGoalWithTrueEqFalse : GoalM Unit := do
let mvarId := ( get).mvarId
unless ( mvarId.isAssigned) do
let trueEqFalse mkEqFalseProof ( getTrueExpr)
let falseProof mkEqMP trueEqFalse (mkConst ``True.intro)
closeGoal falseProof
/-- Closes the goal when `lhs` and `rhs` are both literal values and belong to the same equivalence class. -/
private def closeGoalWithValuesEq (lhs rhs : Expr) : GoalM Unit := do
let p mkEq lhs rhs
let hp mkEqProof lhs rhs
let d mkDecide p
let pEqFalse := mkApp3 (mkConst ``eq_false_of_decide) p d.appArg! (mkApp2 (mkConst ``Eq.refl [1]) (mkConst ``Bool) (mkConst ``false))
let falseProof mkEqMP pEqFalse hp
closeGoal falseProof
/--
Updates the modification time to `gmt` for the parents of `root`.
The modification time is used to decide which terms are considered during e-matching.
-/
private partial def updateMT (root : Expr) : GoalM Unit := do
let gmt := ( get).gmt
for parent in ( getParents root) do
let node getENode parent
if node.mt < gmt then
setENode parent { node with mt := gmt }
updateMT parent
private partial def addEqStep (lhs rhs proof : Expr) (isHEq : Bool) : GoalM Unit := do
trace[grind.eq] "{lhs} {if isHEq then "" else "="} {rhs}"
let lhsNode getENode lhs
let rhsNode getENode rhs
if isSameExpr lhsNode.root rhsNode.root then
-- `lhs` and `rhs` are already in the same equivalence class.
trace[grind.debug] "{← ppENodeRef lhs} and {← ppENodeRef rhs} are already in the same equivalence class"
return ()
trace[grind.eqc] "{← if isHEq then mkHEq lhs rhs else mkEq lhs rhs}"
let lhsRoot getENode lhsNode.root
let rhsRoot getENode rhsNode.root
let mut valueInconsistency := false
let mut trueEqFalse := false
if lhsRoot.interpreted && rhsRoot.interpreted then
if lhsNode.root.isTrue || rhsNode.root.isTrue then
markAsInconsistent
trueEqFalse := true
else
valueInconsistency := true
if (lhsRoot.interpreted && !rhsRoot.interpreted)
|| (lhsRoot.ctor && !rhsRoot.ctor)
|| (lhsRoot.size > rhsRoot.size && !rhsRoot.interpreted && !rhsRoot.ctor) then
go rhs lhs rhsNode lhsNode rhsRoot lhsRoot true
else
go lhs rhs lhsNode rhsNode lhsRoot rhsRoot false
if trueEqFalse then
closeGoalWithTrueEqFalse
unless ( isInconsistent) do
if lhsRoot.ctor && rhsRoot.ctor then
propagateCtor lhsRoot.self rhsRoot.self
unless ( isInconsistent) do
if valueInconsistency then
closeGoalWithValuesEq lhsRoot.self rhsRoot.self
trace[grind.debug] "after addEqStep, {← ppState}"
checkInvariants
where
go (lhs rhs : Expr) (lhsNode rhsNode lhsRoot rhsRoot : ENode) (flipped : Bool) : GoalM Unit := do
trace[grind.debug] "adding {← ppENodeRef lhs} ↦ {← ppENodeRef rhs}"
let mut valueInconsistency := false
if lhsRoot.interpreted && rhsRoot.interpreted then
if lhsNode.root.isTrue || rhsNode.root.isTrue then
markAsInconsistent
else
valueInconsistency := true
-- TODO: process valueInconsistency := true
/-
We have the following `target?/proof?`
`lhs -> ... -> lhsNode.root`
@@ -210,14 +137,12 @@ where
proof? := proof
flipped
}
-- TODO: Remove parents from congruence table
-- TODO: set propagateBool
updateRoots lhs rhsNode.root true -- TODO
let parents removeParents lhsRoot.self
updateRoots lhs rhsNode.root
trace[grind.debug] "{← ppENodeRef lhs} new root {← ppENodeRef rhsNode.root}, {← ppENodeRef (← getRoot lhs)}"
-- TODO: Reinsert parents into congruence table
setENode lhsNode.root { lhsRoot with
reinsertParents parents
setENode lhsNode.root { ( getENode lhsRoot.self) with -- We must retrieve `lhsRoot` since it was updated.
next := rhsRoot.next
root := rhsNode.root
}
setENode rhsNode.root { rhsRoot with
next := lhsRoot.next
@@ -225,22 +150,35 @@ where
hasLambdas := rhsRoot.hasLambdas || lhsRoot.hasLambdas
heqProofs := isHEq || rhsRoot.heqProofs || lhsRoot.heqProofs
}
-- TODO: copy parentst from lhsRoot parents to rhsRoot parents
copyParentsTo parents rhsNode.root
unless ( isInconsistent) do
for parent in parents do
propagateUp parent
unless ( isInconsistent) do
updateMT rhsRoot.self
updateRoots (lhs : Expr) (rootNew : Expr) (_propagateBool : Bool) : GoalM Unit := do
updateRoots (lhs : Expr) (rootNew : Expr) : GoalM Unit := do
let rec loop (e : Expr) : GoalM Unit := do
-- TODO: propagateBool
let n getENode e
setENode e { n with root := rootNew }
unless ( isInconsistent) do
propagateDown e
if isSameExpr lhs n.next then return ()
loop n.next
loop lhs
/-- Ensures collection of equations to be processed is empty. -/
def resetNewEqs : GoalM Unit :=
private def resetNewEqs : GoalM Unit :=
modify fun s => { s with newEqs := #[] }
partial def addEqCore (lhs rhs proof : Expr) (isHEq : Bool) : GoalM Unit := do
/-- Pops and returns the next equality to be processed. -/
private def popNextEq? : GoalM (Option NewEq) := do
let r := ( get).newEqs.back?
if r.isSome then
modify fun s => { s with newEqs := s.newEqs.pop }
return r
private partial def addEqCore (lhs rhs proof : Expr) (isHEq : Bool) : GoalM Unit := do
addEqStep lhs rhs proof isHEq
processTodo
where
@@ -248,21 +186,29 @@ where
if ( isInconsistent) then
resetNewEqs
return ()
let some { lhs, rhs, proof, isHEq } := ( get).newEqs.back? | return ()
checkSystem "grind"
let some { lhs, rhs, proof, isHEq } := ( popNextEq?) | return ()
addEqStep lhs rhs proof isHEq
processTodo
/-- Adds a new equality `lhs = rhs`. It assumes `lhs` and `rhs` have already been internalized. -/
def addEq (lhs rhs proof : Expr) : GoalM Unit := do
addEqCore lhs rhs proof false
/-- Adds a new heterogeneous equality `HEq lhs rhs`. It assumes `lhs` and `rhs` have already been internalized. -/
def addHEq (lhs rhs proof : Expr) : GoalM Unit := do
addEqCore lhs rhs proof true
/--
Adds a new `fact` justified by the given proof and using the given generation.
-/
/-- Internalizes `lhs` and `rhs`, and then adds equality `lhs = rhs`. -/
def addNewEq (lhs rhs proof : Expr) (generation : Nat) : GoalM Unit := do
internalize lhs generation
internalize rhs generation
addEq lhs rhs proof
/-- Adds a new `fact` justified by the given proof and using the given generation. -/
def add (fact : Expr) (proof : Expr) (generation := 0) : GoalM Unit := do
trace[grind.add] "{proof} : {fact}"
trace[grind.assert] "{fact}"
if ( isInconsistent) then return ()
resetNewEqs
let_expr Not p := fact
@@ -270,10 +216,9 @@ def add (fact : Expr) (proof : Expr) (generation := 0) : GoalM Unit := do
go p true
where
go (p : Expr) (isNeg : Bool) : GoalM Unit := do
trace[grind.add] "isNeg: {isNeg}, {p}"
match_expr p with
| Eq _ lhs rhs => goEq p lhs rhs isNeg false
| HEq _ _ lhs rhs => goEq p lhs rhs isNeg true
| HEq _ lhs _ rhs => goEq p lhs rhs isNeg true
| _ =>
internalize p generation
if isNeg then
@@ -290,10 +235,8 @@ where
internalize rhs generation
addEqCore lhs rhs proof isHEq
/--
Adds a new hypothesis.
-/
def addHyp (fvarId : FVarId) (generation := 0) : GoalM Unit := do
/-- Adds a new hypothesis. -/
def addHypothesis (fvarId : FVarId) (generation := 0) : GoalM Unit := do
add ( fvarId.getType) (mkFVar fvarId) generation
end Lean.Meta.Grind

View File

@@ -0,0 +1,53 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Lean.Meta.Tactic.Grind.Types
namespace Lean.Meta.Grind
private partial def propagateInjEqs (eqs : Expr) (proof : Expr) : GoalM Unit := do
-- Remark: we must use `shareCommon` before using `pushEq` and `pushHEq`.
-- This is needed because the result type of the injection theorem may allocate
match_expr eqs with
| And left right =>
propagateInjEqs left (.proj ``And 0 proof)
propagateInjEqs right (.proj ``And 1 proof)
| Eq _ lhs rhs =>
pushEq ( shareCommon lhs) ( shareCommon rhs) proof
| HEq _ lhs _ rhs =>
pushHEq ( shareCommon lhs) ( shareCommon rhs) proof
| _ =>
trace[grind.issues] "unexpected injectivity theorem result type{indentExpr eqs}"
return ()
/--
Given constructors `a` and `b`, propagate equalities if they are the same,
and close goal if they are different.
-/
def propagateCtor (a b : Expr) : GoalM Unit := do
let aType whnfD ( inferType a)
let bType whnfD ( inferType b)
unless ( withDefault <| isDefEq aType bType) do
return ()
let ctor₁ := a.getAppFn
let ctor₂ := b.getAppFn
if ctor₁ == ctor₂ then
let .const ctorName _ := a.getAppFn | return ()
let injDeclName := Name.mkStr ctorName "inj"
unless ( getEnv).contains injDeclName do return ()
let info getConstInfo injDeclName
let n := info.type.getForallArity
let mask : Array (Option Expr) := mkArray n none
let mask := mask.set! (n-1) (some ( mkEqProof a b))
let injLemma mkAppOptM injDeclName mask
propagateInjEqs ( inferType injLemma) injLemma
else
let .const declName _ := aType.getAppFn | return ()
let noConfusionDeclName := Name.mkStr declName "noConfusion"
unless ( getEnv).contains noConfusionDeclName do return ()
closeGoal ( mkNoConfusion ( getFalseExpr) ( mkEqProof a b))
end Lean.Meta.Grind

View File

@@ -0,0 +1,35 @@
/-
Copyright (c) 2025 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Util
import Init.Simproc
import Lean.Meta.Tactic.Simp.Simproc
namespace Lean.Meta.Grind
/--
Returns `Grind.doNotSimp e`.
Recall that `Grind.doNotSimp` is an identity function, but the following simproc is used to prevent the term `e` from being simplified.
-/
def markAsDoNotSimp (e : Expr) : MetaM Expr :=
mkAppM ``Grind.doNotSimp #[e]
builtin_dsimproc_decl reduceDoNotSimp (Grind.doNotSimp _) := fun e => do
let_expr Grind.doNotSimp _ _ e | return .continue
return .done e
/-- Adds `reduceDoNotSimp` to `s` -/
def addDoNotSimp (s : Simprocs) : CoreM Simprocs := do
s.add ``reduceDoNotSimp (post := false)
/-- Erases `Grind.doNotSimp` annotations. -/
def eraseDoNotSimp (e : Expr) : CoreM Expr := do
let pre (e : Expr) := do
let_expr Grind.doNotSimp _ a := e | return .continue e
return .continue a
Core.transform e (pre := pre)
end Lean.Meta.Grind

View File

@@ -0,0 +1,354 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Lean.Meta.Tactic.Grind.Types
import Lean.Meta.Tactic.Grind.Intro
import Lean.Meta.Tactic.Grind.DoNotSimp
namespace Lean.Meta.Grind
namespace EMatch
/-! This module implements a simple E-matching procedure as a backtracking search. -/
/-- We represent an `E-matching` problem as a list of constraints. -/
inductive Cnstr where
| /-- Matches pattern `pat` with term `e` -/
«match» (pat : Expr) (e : Expr)
| /-- Matches offset pattern `pat+k` with term `e` -/
offset (pat : Expr) (k : Nat) (e : Expr)
| /-- This constraint is used to encode multi-patterns. -/
«continue» (pat : Expr)
deriving Inhabited
/--
Internal "marker" for representing unassigned elemens in the `assignment` field.
This is a small hack to avoid one extra level of indirection by using `Option Expr` at `assignment`.
-/
private def unassigned : Expr := mkConst (Name.mkSimple "[grind_unassigned]")
private def assignmentToMessageData (assignment : Array Expr) : Array MessageData :=
assignment.reverse.map fun e =>
if isSameExpr e unassigned then m!"_" else m!"{e}"
/--
Choice point for the backtracking search.
The state of the procedure contains a stack of choices.
-/
structure Choice where
/-- Contraints to be processed. -/
cnstrs : List Cnstr
/-- Maximum term generation found so far. -/
gen : Nat
/-- Partial assignment so far. Recall that pattern variables are encoded as de-Bruijn variables. -/
assignment : Array Expr
deriving Inhabited
/-- Context for the E-matching monad. -/
structure Context where
/-- `useMT` is `true` if we are using the mod-time optimization. It is always set to false for new `EMatchTheorem`s. -/
useMT : Bool := true
/-- `EMatchTheorem` being processed. -/
thm : EMatchTheorem := default
deriving Inhabited
/-- State for the E-matching monad -/
structure State where
/-- Choices that still have to be processed. -/
choiceStack : List Choice := []
deriving Inhabited
abbrev M := ReaderT Context $ StateRefT State GoalM
def M.run' (x : M α) : GoalM α :=
x {} |>.run' {}
/--
Assigns `bidx := e` in `c`. If `bidx` is already assigned in `c`, we check whether
`e` and `c.assignment[bidx]` are in the same equivalence class.
This function assumes `bidx < c.assignment.size`.
Recall that we initialize the assignment array with the number of theorem parameters.
-/
private def assign? (c : Choice) (bidx : Nat) (e : Expr) : OptionT GoalM Choice := do
if h : bidx < c.assignment.size then
let v := c.assignment[bidx]
if isSameExpr v unassigned then
return { c with assignment := c.assignment.set bidx e }
else
guard ( isEqv v e)
return c
else
-- `Choice` was not properly initialized
unreachable!
/--
Returns `true` if the function `pFn` of a pattern is equivalent to the function `eFn`.
Recall that we ignore universe levels in patterns.
-/
private def eqvFunctions (pFn eFn : Expr) : Bool :=
(pFn.isFVar && pFn == eFn)
|| (pFn.isConst && eFn.isConstOf pFn.constName!)
/-- Matches a pattern argument. See `matchArgs?`. -/
private def matchArg? (c : Choice) (pArg : Expr) (eArg : Expr) : OptionT GoalM Choice := do
if isPatternDontCare pArg then
return c
else if pArg.isBVar then
assign? c pArg.bvarIdx! eArg
else if let some pArg := groundPattern? pArg then
guard ( isEqv pArg eArg)
return c
else if let some (pArg, k) := isOffsetPattern? pArg then
assert! Option.isNone <| isOffsetPattern? pArg
assert! !isPatternDontCare pArg
return { c with cnstrs := .offset pArg k eArg :: c.cnstrs }
else
return { c with cnstrs := .match pArg eArg :: c.cnstrs }
private def Choice.updateGen (c : Choice) (gen : Nat) : Choice :=
{ c with gen := Nat.max gen c.gen }
private def pushChoice (c : Choice) : M Unit :=
modify fun s => { s with choiceStack := c :: s.choiceStack }
/--
Matches arguments of pattern `p` with term `e`. Returns `some` if successful,
and `none` otherwise. It may update `c`s assignment and list of contraints to be
processed.
-/
private partial def matchArgs? (c : Choice) (p : Expr) (e : Expr) : OptionT GoalM Choice := do
if !p.isApp then return c -- Done
let pArg := p.appArg!
let eArg := e.appArg!
let c matchArg? c pArg eArg
matchArgs? c p.appFn! e.appFn!
/--
Matches pattern `p` with term `e` with respect to choice `c`.
We traverse the equivalence class of `e` looking for applications compatible with `p`.
For each candidate application, we match the arguments and may update `c`s assignments and contraints.
We add the updated choices to the choice stack.
-/
private partial def processMatch (c : Choice) (p : Expr) (e : Expr) : M Unit := do
let maxGeneration getMaxGeneration
let pFn := p.getAppFn
let numArgs := p.getAppNumArgs
let mut curr := e
repeat
let n getENode curr
-- Remark: we use `<` because the instance generation is the maximum term generation + 1
if n.generation < maxGeneration
-- uses heterogeneous equality or is the root of its congruence class
&& (n.heqProofs || n.isCongrRoot)
&& eqvFunctions pFn curr.getAppFn
&& curr.getAppNumArgs == numArgs then
if let some c matchArgs? c p curr |>.run then
pushChoice (c.updateGen n.generation)
curr getNext curr
if isSameExpr curr e then break
/--
Matches offset pattern `pArg+k` with term `e` with respect to choice `c`.
-/
private partial def processOffset (c : Choice) (pArg : Expr) (k : Nat) (e : Expr) : M Unit := do
let maxGeneration getMaxGeneration
let mut curr := e
repeat
let n getENode curr
if n.generation < maxGeneration then
if let some (eArg, k') isOffset? curr |>.run then
if k' < k then
let c := c.updateGen n.generation
pushChoice { c with cnstrs := .offset pArg (k - k') eArg :: c.cnstrs }
else if k' == k then
if let some c matchArg? c pArg eArg |>.run then
pushChoice (c.updateGen n.generation)
else if k' > k then
let eArg' := mkNatAdd eArg (mkNatLit (k' - k))
let eArg' shareCommon ( canon eArg')
internalize eArg' (n.generation+1)
if let some c matchArg? c pArg eArg' |>.run then
pushChoice (c.updateGen n.generation)
else if let some k' evalNat curr |>.run then
if k' >= k then
let eArg' := mkNatLit (k' - k)
let eArg' shareCommon ( canon eArg')
internalize eArg' (n.generation+1)
if let some c matchArg? c pArg eArg' |>.run then
pushChoice (c.updateGen n.generation)
curr getNext curr
if isSameExpr curr e then break
/-- Processes `continue` contraint used to implement multi-patterns. -/
private def processContinue (c : Choice) (p : Expr) : M Unit := do
let some apps := ( getThe Goal).appMap.find? p.toHeadIndex
| return ()
let maxGeneration getMaxGeneration
for app in apps do
let n getENode app
if n.generation < maxGeneration
&& (n.heqProofs || n.isCongrRoot) then
if let some c matchArgs? c p app |>.run then
let gen := n.generation
let c := { c with gen := Nat.max gen c.gen }
modify fun s => { s with choiceStack := c :: s.choiceStack }
/-- Helper function for marking parts of `match`-equation theorem as "do-not-simplify" -/
private partial def annotateMatchEqnType (prop : Expr) : M Expr := do
if let .forallE n d b bi := prop then
withLocalDecl n bi ( markAsDoNotSimp d) fun x => do
mkForallFVars #[x] ( annotateMatchEqnType (b.instantiate1 x))
else
let_expr f@Eq α lhs rhs := prop | return prop
return mkApp3 f α ( markAsDoNotSimp lhs) rhs
/--
Stores new theorem instance in the state.
Recall that new instances are internalized later, after a full round of ematching.
-/
private def addNewInstance (origin : Origin) (proof : Expr) (generation : Nat) : M Unit := do
let proof instantiateMVars proof
if grind.debug.proofs.get ( getOptions) then
check proof
let mut prop inferType proof
if Match.isMatchEqnTheorem ( getEnv) origin.key then
prop annotateMatchEqnType prop
trace[grind.ematch.instance] "{← origin.pp}: {prop}"
addTheoremInstance proof prop (generation+1)
/--
After processing a (multi-)pattern, use the choice assignment to instantiate the proof.
Missing parameters are synthesized using type inference and type class synthesis."
-/
private partial def instantiateTheorem (c : Choice) : M Unit := withDefault do withNewMCtxDepth do
let thm := ( read).thm
unless ( markTheoremInstance thm.proof c.assignment) do
return ()
trace[grind.ematch.instance.assignment] "{← thm.origin.pp}: {assignmentToMessageData c.assignment}"
let proof thm.getProofWithFreshMVarLevels
let numParams := thm.numParams
assert! c.assignment.size == numParams
let (mvars, bis, _) forallMetaBoundedTelescope ( inferType proof) numParams
if mvars.size != thm.numParams then
trace[grind.issues] "unexpected number of parameters at {← thm.origin.pp}"
return ()
-- Apply assignment
for h : i in [:mvars.size] do
let v := c.assignment[numParams - i - 1]!
unless isSameExpr v unassigned do
let mvarId := mvars[i].mvarId!
unless ( isDefEq ( mvarId.getType) ( inferType v) <&&> mvarId.checkedAssign v) do
trace[grind.issues] "type error constructing proof for {← thm.origin.pp}\nwhen assigning metavariable {mvars[i]} with {indentExpr v}"
return ()
-- Synthesize instances
for mvar in mvars, bi in bis do
if bi.isInstImplicit && !( mvar.mvarId!.isAssigned) then
let type inferType mvar
unless ( synthesizeInstance mvar type) do
trace[grind.issues] "failed to synthesize instance when instantiating {← thm.origin.pp}{indentExpr type}"
return ()
let proof := mkAppN proof mvars
if ( mvars.allM (·.mvarId!.isAssigned)) then
addNewInstance thm.origin proof c.gen
else
let mvars mvars.filterM fun mvar => return !( mvar.mvarId!.isAssigned)
if let some mvarBad mvars.findM? fun mvar => return !( isProof mvar) then
trace[grind.issues] "failed to instantiate {← thm.origin.pp}, failed to instantiate non propositional argument with type{indentExpr (← inferType mvarBad)}"
let proof mkLambdaFVars (binderInfoForMVars := .default) mvars ( instantiateMVars proof)
addNewInstance thm.origin proof c.gen
where
synthesizeInstance (x type : Expr) : MetaM Bool := do
let .some val trySynthInstance type | return false
isDefEq x val
/-- Process choice stack until we don't have more choices to be processed. -/
private def processChoices : M Unit := do
let maxGeneration getMaxGeneration
while !( get).choiceStack.isEmpty do
checkSystem "ematch"
if ( checkMaxInstancesExceeded) then return ()
let c modifyGet fun s : State => (s.choiceStack.head!, { s with choiceStack := s.choiceStack.tail! })
if c.gen < maxGeneration then
match c.cnstrs with
| [] => instantiateTheorem c
| .match p e :: cnstrs => processMatch { c with cnstrs } p e
| .offset p k e :: cnstrs => processOffset { c with cnstrs } p k e
| .continue p :: cnstrs => processContinue { c with cnstrs } p
private def main (p : Expr) (cnstrs : List Cnstr) : M Unit := do
let some apps := ( getThe Goal).appMap.find? p.toHeadIndex
| return ()
let numParams := ( read).thm.numParams
let assignment := mkArray numParams unassigned
let useMT := ( read).useMT
let gmt := ( getThe Goal).gmt
for app in apps do
if ( checkMaxInstancesExceeded) then return ()
let n getENode app
if (n.heqProofs || n.isCongrRoot) &&
(!useMT || n.mt == gmt) then
if let some c matchArgs? { cnstrs, assignment, gen := n.generation } p app |>.run then
modify fun s => { s with choiceStack := [c] }
processChoices
def ematchTheorem (thm : EMatchTheorem) : M Unit := do
if ( checkMaxInstancesExceeded) then return ()
withReader (fun ctx => { ctx with thm }) do
let ps := thm.patterns
match ps, ( read).useMT with
| [p], _ => main p []
| p::ps, false => main p (ps.map (.continue ·))
| _::_, true => tryAll ps []
| _, _ => unreachable!
where
/--
When using the mod-time optimization with multi-patterns,
we must start ematching at each different pattern. That is,
if we have `[p₁, p₂, p₃]`, we must execute
- `main p₁ [.continue p₂, .continue p₃]`
- `main p₂ [.continue p₁, .continue p₃]`
- `main p₃ [.continue p₁, .continue p₂]`
-/
tryAll (ps : List Expr) (cs : List Cnstr) : M Unit := do
match ps with
| [] => return ()
| p::ps =>
main p (cs.reverse ++ (ps.map (.continue ·)))
tryAll ps (.continue p :: cs)
def ematchTheorems (thms : PArray EMatchTheorem) : M Unit := do
thms.forM ematchTheorem
end EMatch
open EMatch
/-- Performs one round of E-matching, and returns new instances. -/
def ematch : GoalM Unit := do
let go (thms newThms : PArray EMatchTheorem) : EMatch.M Unit := do
withReader (fun ctx => { ctx with useMT := true }) <| ematchTheorems thms
withReader (fun ctx => { ctx with useMT := false }) <| ematchTheorems newThms
if ( checkMaxInstancesExceeded <||> checkMaxEmatchExceeded) then
return ()
else
go ( get).thms ( get).newThms |>.run'
modify fun s => { s with
thms := s.thms ++ s.newThms
newThms := {}
gmt := s.gmt + 1
numEmatch := s.numEmatch + 1
}
/-- Performs one round of E-matching, and assert new instances. -/
def ematchAndAssert? (goal : Goal) : GrindM (Option (List Goal)) := do
let numInstances := goal.numInstances
let goal GoalM.run' goal ematch
if goal.numInstances == numInstances then
return none
assertAll goal
def ematchStar (goal : Goal) : GrindM (List Goal) := do
iterate goal ematchAndAssert?
end Lean.Meta.Grind

View File

@@ -0,0 +1,667 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Util
import Init.Grind.Tactics
import Lean.HeadIndex
import Lean.PrettyPrinter
import Lean.Util.FoldConsts
import Lean.Util.CollectFVars
import Lean.Meta.Basic
import Lean.Meta.InferType
import Lean.Meta.Eqns
import Lean.Meta.Tactic.Grind.Util
namespace Lean.Meta.Grind
def mkOffsetPattern (pat : Expr) (k : Nat) : Expr :=
mkApp2 (mkConst ``Grind.offset) pat (mkRawNatLit k)
private def detectOffsets (pat : Expr) : MetaM Expr := do
let pre (e : Expr) := do
if e == pat then
-- We only consider nested offset patterns
return .continue e
else match e with
| .letE .. | .lam .. | .forallE .. => return .done e
| _ =>
let some (e, k) isOffset? e
| return .continue e
if k == 0 then return .continue e
return .continue <| mkOffsetPattern e k
Core.transform pat (pre := pre)
def isOffsetPattern? (pat : Expr) : Option (Expr × Nat) := Id.run do
let_expr Grind.offset pat k := pat | none
let .lit (.natVal k) := k | none
return some (pat, k)
def preprocessPattern (pat : Expr) (normalizePattern := true) : MetaM Expr := do
let pat instantiateMVars pat
let pat unfoldReducible pat
let pat if normalizePattern then normalize pat else pure pat
let pat detectOffsets pat
let pat foldProjs pat
return pat
inductive Origin where
/-- A global declaration in the environment. -/
| decl (declName : Name)
/-- A local hypothesis. -/
| fvar (fvarId : FVarId)
/--
A proof term provided directly to a call to `grind` where `ref`
is the provided grind argument. The `id` is a unique identifier for the call.
-/
| stx (id : Name) (ref : Syntax)
| other
deriving Inhabited, Repr
/-- A unique identifier corresponding to the origin. -/
def Origin.key : Origin Name
| .decl declName => declName
| .fvar fvarId => fvarId.name
| .stx id _ => id
| .other => `other
def Origin.pp [Monad m] [MonadEnv m] [MonadError m] (o : Origin) : m MessageData := do
match o with
| .decl declName => return MessageData.ofConst ( mkConstWithLevelParams declName)
| .fvar fvarId => return mkFVar fvarId
| .stx _ ref => return ref
| .other => return "[unknown]"
/-- A theorem for heuristic instantiation based on E-matching. -/
structure EMatchTheorem where
/--
It stores universe parameter names for universe polymorphic proofs.
Recall that it is non-empty only when we elaborate an expression provided by the user.
When `proof` is just a constant, we can use the universe parameter names stored in the declaration.
-/
levelParams : Array Name
proof : Expr
numParams : Nat
patterns : List Expr
/-- Contains all symbols used in `pattterns`. -/
symbols : List HeadIndex
origin : Origin
deriving Inhabited
/-- Set of E-matching theorems. -/
structure EMatchTheorems where
/-- The key is a symbol from `EMatchTheorem.symbols`. -/
private map : PHashMap Name (List EMatchTheorem) := {}
/-- Set of theorem names that have been inserted using `insert`. -/
private thmNames : PHashSet Name := {}
deriving Inhabited
/--
Inserts a `thm` with symbols `[s_1, ..., s_n]` to `s`.
We add `s_1 -> { thm with symbols := [s_2, ..., s_n] }`.
When `grind` internalizes a term containing symbol `s`, we
process all theorems `thm` associated with key `s`.
If their `thm.symbols` is empty, we say they are activated.
Otherwise, we reinsert into `map`.
-/
def EMatchTheorems.insert (s : EMatchTheorems) (thm : EMatchTheorem) : EMatchTheorems := Id.run do
let .const declName :: syms := thm.symbols
| unreachable!
let thm := { thm with symbols := syms }
let { map, thmNames } := s
let thmNames := thmNames.insert thm.origin.key
if let some thms := map.find? declName then
return { map := map.insert declName (thm::thms), thmNames }
else
return { map := map.insert declName [thm], thmNames }
/--
Retrieves theorems from `s` associated with the given symbol. See `EMatchTheorem.insert`.
The theorems are removed from `s`.
-/
@[inline]
def EMatchTheorems.retrieve? (s : EMatchTheorems) (sym : Name) : Option (List EMatchTheorem × EMatchTheorems) :=
if let some thms := s.map.find? sym then
some (thms, { s with map := s.map.erase sym })
else
none
/-- Returns `true` if `declName` is the name of a theorem that was inserted using `insert`. -/
def EMatchTheorems.containsTheoremName (s : EMatchTheorems) (declName : Name) : Bool :=
s.thmNames.contains declName
def EMatchTheorem.getProofWithFreshMVarLevels (thm : EMatchTheorem) : MetaM Expr := do
if thm.proof.isConst && thm.levelParams.isEmpty then
let declName := thm.proof.constName!
let info getConstInfo declName
if info.levelParams.isEmpty then
return thm.proof
else
mkConstWithFreshMVarLevels declName
else if thm.levelParams.isEmpty then
return thm.proof
else
let us thm.levelParams.mapM fun _ => mkFreshLevelMVar
return thm.proof.instantiateLevelParamsArray thm.levelParams us
private builtin_initialize ematchTheoremsExt : SimpleScopedEnvExtension EMatchTheorem EMatchTheorems
registerSimpleScopedEnvExtension {
addEntry := EMatchTheorems.insert
initial := {}
}
-- TODO: create attribute?
private def forbiddenDeclNames := #[``Eq, ``HEq, ``Iff, ``And, ``Or, ``Not]
private def isForbidden (declName : Name) := forbiddenDeclNames.contains declName
private def dontCare := mkConst (Name.mkSimple "[grind_dontcare]")
def mkGroundPattern (e : Expr) : Expr :=
mkAnnotation `grind.ground_pat e
def groundPattern? (e : Expr) : Option Expr :=
annotation? `grind.ground_pat e
private def isGroundPattern (e : Expr) : Bool :=
groundPattern? e |>.isSome
def isPatternDontCare (e : Expr) : Bool :=
e == dontCare
private def isAtomicPattern (e : Expr) : Bool :=
e.isBVar || isPatternDontCare e || isGroundPattern e
partial def ppPattern (pattern : Expr) : MessageData := Id.run do
if let some e := groundPattern? pattern then
return m!"`[{e}]"
else if isPatternDontCare pattern then
return m!"?"
else match pattern with
| .bvar idx => return m!"#{idx}"
| _ =>
let mut r := m!"{pattern.getAppFn}"
for arg in pattern.getAppArgs do
let mut argFmt ppPattern arg
if !isAtomicPattern arg then
argFmt := MessageData.paren argFmt
r := r ++ " " ++ argFmt
return r
namespace NormalizePattern
structure State where
symbols : Array HeadIndex := #[]
symbolSet : Std.HashSet HeadIndex := {}
bvarsFound : Std.HashSet Nat := {}
abbrev M := StateRefT State MetaM
private def saveSymbol (h : HeadIndex) : M Unit := do
unless ( get).symbolSet.contains h do
modify fun s => { s with symbols := s.symbols.push h, symbolSet := s.symbolSet.insert h }
private def foundBVar (idx : Nat) : M Bool :=
return ( get).bvarsFound.contains idx
private def saveBVar (idx : Nat) : M Unit := do
modify fun s => { s with bvarsFound := s.bvarsFound.insert idx }
private def getPatternFn? (pattern : Expr) : Option Expr :=
if !pattern.isApp then
none
else match pattern.getAppFn with
| f@(.const declName _) => if isForbidden declName then none else some f
| f@(.fvar _) => some f
| _ => none
/--
Returns a bit-mask `mask` s.t. `mask[i]` is true if the the corresponding argument is
- a type (that is not a proposition) or type former, or
- a proof, or
- an instance implicit argument
When `mask[i]`, we say the corresponding argument is a "support" argument.
-/
def getPatternSupportMask (f : Expr) (numArgs : Nat) : MetaM (Array Bool) := do
forallBoundedTelescope ( inferType f) numArgs fun xs _ => do
xs.mapM fun x => do
if ( isProp x) then
return false
else if ( isTypeFormer x <||> isProof x) then
return true
else
return ( x.fvarId!.getDecl).binderInfo matches .instImplicit
private partial def go (pattern : Expr) (root := false) : M Expr := do
if root && !pattern.hasLooseBVars then
throwError "invalid pattern, it does not have pattern variables"
if let some (e, k) := isOffsetPattern? pattern then
let e goArg e (isSupport := false)
if e == dontCare then
return dontCare
else
return mkOffsetPattern e k
let some f := getPatternFn? pattern
| throwError "invalid pattern, (non-forbidden) application expected{indentExpr pattern}"
assert! f.isConst || f.isFVar
saveSymbol f.toHeadIndex
let mut args := pattern.getAppArgs
let supportMask getPatternSupportMask f args.size
for i in [:args.size] do
let arg := args[i]!
let isSupport := supportMask[i]?.getD false
args := args.set! i ( goArg arg isSupport)
return mkAppN f args
where
goArg (arg : Expr) (isSupport : Bool) : M Expr := do
if !arg.hasLooseBVars then
if arg.hasMVar then
pure dontCare
else
pure <| mkGroundPattern arg
else match arg with
| .bvar idx =>
if isSupport && ( foundBVar idx) then
pure dontCare
else
saveBVar idx
pure arg
| _ =>
if isSupport then
pure dontCare
else if let some _ := getPatternFn? arg then
go arg
else
pure dontCare
def main (patterns : List Expr) : MetaM (List Expr × List HeadIndex × Std.HashSet Nat) := do
let (patterns, s) patterns.mapM go |>.run {}
return (patterns, s.symbols.toList, s.bvarsFound)
def normalizePattern (e : Expr) : M Expr := do
go e
end NormalizePattern
/--
Returns `true` if free variables in `type` are not in `thmVars` or are in `fvarsFound`.
We use this function to check whether `type` is fully instantiated.
-/
private def checkTypeFVars (thmVars : FVarIdSet) (fvarsFound : FVarIdSet) (type : Expr) : Bool :=
let typeFVars := (collectFVars {} type).fvarIds
typeFVars.all fun fvarId => !thmVars.contains fvarId || fvarsFound.contains fvarId
/--
Given an type class instance type `instType`, returns true if free variables in input parameters
1- are not in `thmVars`, or
2- are in `fvarsFound`.
Remark: `fvarsFound` is a subset of `thmVars`
-/
private def canBeSynthesized (thmVars : FVarIdSet) (fvarsFound : FVarIdSet) (instType : Expr) : MetaM Bool := do
forallTelescopeReducing instType fun xs type => type.withApp fun classFn classArgs => do
for x in xs do
unless checkTypeFVars thmVars fvarsFound ( inferType x) do return false
forallBoundedTelescope ( inferType classFn) type.getAppNumArgs fun params _ => do
for param in params, classArg in classArgs do
let paramType inferType param
if !paramType.isAppOf ``semiOutParam && !paramType.isAppOf ``outParam then
unless checkTypeFVars thmVars fvarsFound classArg do
return false
return true
/--
Auxiliary type for the `checkCoverage` function.
-/
inductive CheckCoverageResult where
| /-- `checkCoverage` succeeded -/
ok
| /--
`checkCoverage` failed because some of the theorem parameters are missing,
`pos` contains their positions
-/
missing (pos : List Nat)
/--
After we process a set of patterns, we obtain the set of de Bruijn indices in these patterns.
We say they are pattern variables. This function checks whether the set of pattern variables is sufficient for
instantiating the theorem with proof `thmProof`. The theorem has `numParams` parameters.
The missing parameters:
1- we may be able to infer them using type inference or type class synthesis, or
2- they are propositions, and may become hypotheses of the instantiated theorem.
For type class instance parameters, we must check whether the free variables in class input parameters are available.
-/
private def checkCoverage (thmProof : Expr) (numParams : Nat) (bvarsFound : Std.HashSet Nat) : MetaM CheckCoverageResult := do
if bvarsFound.size == numParams then return .ok
forallBoundedTelescope ( inferType thmProof) numParams fun xs _ => do
assert! numParams == xs.size
let patternVars := bvarsFound.toList.map fun bidx => xs[numParams - bidx - 1]!.fvarId!
-- `xs` as a `FVarIdSet`.
let thmVars : FVarIdSet := RBTree.ofList <| xs.toList.map (·.fvarId!)
-- Collect free variables occurring in `e`, and insert the ones that are in `thmVars` into `fvarsFound`
let update (fvarsFound : FVarIdSet) (e : Expr) : FVarIdSet :=
(collectFVars {} e).fvarIds.foldl (init := fvarsFound) fun s fvarId =>
if thmVars.contains fvarId then s.insert fvarId else s
-- Theorem variables found so far. We initialize with the variables occurring in patterns
-- Remark: fvarsFound is a subset of thmVars
let mut fvarsFound : FVarIdSet := RBTree.ofList patternVars
for patternVar in patternVars do
let type patternVar.getType
fvarsFound := update fvarsFound type
if fvarsFound.size == numParams then return .ok
-- Now, we keep traversing remaining variables and collecting
-- `processed` contains the variables we have already processed.
let mut processed : FVarIdSet := RBTree.ofList patternVars
let mut modified := false
repeat
modified := false
for x in xs do
let fvarId := x.fvarId!
unless processed.contains fvarId do
let xType inferType x
if fvarsFound.contains fvarId then
-- Collect free vars in `x`s type and mark as processed
fvarsFound := update fvarsFound xType
processed := processed.insert fvarId
modified := true
else if ( isProp xType) then
-- If `x` is a proposition, and all theorem variables in `x`s type have already been found
-- add it to `fvarsFound` and mark it as processed.
if checkTypeFVars thmVars fvarsFound xType then
fvarsFound := fvarsFound.insert fvarId
processed := processed.insert fvarId
modified := true
else if ( fvarId.getDecl).binderInfo matches .instImplicit then
-- If `x` is instance implicit, check whether
-- we have found all free variables needed to synthesize instance
if ( canBeSynthesized thmVars fvarsFound xType) then
fvarsFound := fvarsFound.insert fvarId
fvarsFound := update fvarsFound xType
processed := processed.insert fvarId
modified := true
if fvarsFound.size == numParams then
return .ok
if !modified then
break
let mut pos := #[]
for h : i in [:xs.size] do
let fvarId := xs[i].fvarId!
unless fvarsFound.contains fvarId do
pos := pos.push i
return .missing pos.toList
/--
Given a theorem with proof `proof` and `numParams` parameters, returns a message
containing the parameters at positions `paramPos`.
-/
private def ppParamsAt (proof : Expr) (numParams : Nat) (paramPos : List Nat) : MetaM MessageData := do
forallBoundedTelescope ( inferType proof) numParams fun xs _ => do
let mut msg := m!""
let mut first := true
for h : i in [:xs.size] do
if paramPos.contains i then
let x := xs[i]
if first then first := false else msg := msg ++ "\n"
msg := msg ++ m!"{x} : {← inferType x}"
addMessageContextFull msg
/--
Creates an E-matching theorem for a theorem with proof `proof`, `numParams` parameters, and the given set of patterns.
Pattern variables are represented using de Bruijn indices.
-/
def mkEMatchTheoremCore (origin : Origin) (levelParams : Array Name) (numParams : Nat) (proof : Expr) (patterns : List Expr) : MetaM EMatchTheorem := do
let (patterns, symbols, bvarFound) NormalizePattern.main patterns
trace[grind.ematch.pattern] "{MessageData.ofConst proof}: {patterns.map ppPattern}"
if let .missing pos checkCoverage proof numParams bvarFound then
let pats : MessageData := m!"{patterns.map ppPattern}"
throwError "invalid pattern(s) for `{← origin.pp}`{indentD pats}\nthe following theorem parameters cannot be instantiated:{indentD (← ppParamsAt proof numParams pos)}"
return {
proof, patterns, numParams, symbols
levelParams, origin
}
private def getProofFor (declName : Name) : CoreM Expr := do
let .thmInfo info getConstInfo declName
| throwError "`{declName}` is not a theorem"
let us := info.levelParams.map mkLevelParam
return mkConst declName us
/--
Creates an E-matching theorem for `declName` with `numParams` parameters, and the given set of patterns.
Pattern variables are represented using de Bruijn indices.
-/
def mkEMatchTheorem (declName : Name) (numParams : Nat) (patterns : List Expr) : MetaM EMatchTheorem := do
mkEMatchTheoremCore (.decl declName) #[] numParams ( getProofFor declName) patterns
/--
Given a theorem with proof `proof` and type of the form `∀ (a_1 ... a_n), lhs = rhs`,
creates an E-matching pattern for it using `addEMatchTheorem n [lhs]`
If `normalizePattern` is true, it applies the `grind` simplification theorems and simprocs to the pattern.
-/
def mkEMatchEqTheoremCore (origin : Origin) (levelParams : Array Name) (proof : Expr) (normalizePattern : Bool) (useLhs : Bool) : MetaM EMatchTheorem := do
let (numParams, patterns) forallTelescopeReducing ( inferType proof) fun xs type => do
let (lhs, rhs) match_expr type with
| Eq _ lhs rhs => pure (lhs, rhs)
| Iff lhs rhs => pure (lhs, rhs)
| HEq _ lhs _ rhs => pure (lhs, rhs)
| _ => throwError "invalid E-matching equality theorem, conclusion must be an equality{indentExpr type}"
let pat := if useLhs then lhs else rhs
let pat preprocessPattern pat normalizePattern
return (xs.size, [pat.abstract xs])
mkEMatchTheoremCore origin levelParams numParams proof patterns
/--
Given theorem with name `declName` and type of the form `∀ (a_1 ... a_n), lhs = rhs`,
creates an E-matching pattern for it using `addEMatchTheorem n [lhs]`
If `normalizePattern` is true, it applies the `grind` simplification theorems and simprocs to the
pattern.
-/
def mkEMatchEqTheorem (declName : Name) (normalizePattern := true) (useLhs : Bool := true) : MetaM EMatchTheorem := do
mkEMatchEqTheoremCore (.decl declName) #[] ( getProofFor declName) normalizePattern useLhs
/--
Adds an E-matching theorem to the environment.
See `mkEMatchTheorem`.
-/
def addEMatchTheorem (declName : Name) (numParams : Nat) (patterns : List Expr) : MetaM Unit := do
ematchTheoremsExt.add ( mkEMatchTheorem declName numParams patterns)
/--
Adds an E-matching equality theorem to the environment.
See `mkEMatchEqTheorem`.
-/
def addEMatchEqTheorem (declName : Name) : MetaM Unit := do
ematchTheoremsExt.add ( mkEMatchEqTheorem declName)
/-- Returns the E-matching theorems registered in the environment. -/
def getEMatchTheorems : CoreM EMatchTheorems :=
return ematchTheoremsExt.getState ( getEnv)
private inductive TheoremKind where
| eqLhs | eqRhs | eqBoth | fwd | bwd | default
deriving Inhabited, BEq
private def TheoremKind.toAttribute : TheoremKind String
| .eqLhs => "[grind =]"
| .eqRhs => "[grind =_]"
| .eqBoth => "[grind _=_]"
| .fwd => "[grind →]"
| .bwd => "[grind ←]"
| .default => "[grind]"
private def TheoremKind.explainFailure : TheoremKind String
| .eqLhs => "failed to find pattern in the left-hand side of the theorem's conclusion"
| .eqRhs => "failed to find pattern in the right-hand side of the theorem's conclusion"
| .eqBoth => unreachable! -- eqBoth is a macro
| .fwd => "failed to find patterns in the antecedents of the theorem"
| .bwd => "failed to find patterns in the theorem's conclusion"
| .default => "failed to find patterns"
/-- Returns the types of `xs` that are propositions. -/
private def getPropTypes (xs : Array Expr) : MetaM (Array Expr) :=
xs.filterMapM fun x => do
let type inferType x
if ( isProp type) then return some type else return none
/-- State for the (pattern) `CollectorM` monad -/
private structure Collector.State where
/-- Pattern found so far. -/
patterns : Array Expr := #[]
done : Bool := false
private structure Collector.Context where
proof : Expr
xs : Array Expr
/-- Monad for collecting patterns for a theorem. -/
private abbrev CollectorM := ReaderT Collector.Context $ StateRefT Collector.State NormalizePattern.M
/-- Similar to `getPatternFn?`, but operates on expressions that do not contain loose de Bruijn variables. -/
private def isPatternFnCandidate (f : Expr) : CollectorM Bool := do
match f with
| .const declName _ => return !isForbidden declName
| .fvar .. => return !( read).xs.contains f
| _ => return false
private def addNewPattern (p : Expr) : CollectorM Unit := do
trace[grind.ematch.pattern.search] "found pattern: {ppPattern p}"
let bvarsFound := ( getThe NormalizePattern.State).bvarsFound
let done := ( checkCoverage ( read).proof ( read).xs.size bvarsFound) matches .ok
if done then
trace[grind.ematch.pattern.search] "found full coverage"
modify fun s => { s with patterns := s.patterns.push p, done }
private partial def collect (e : Expr) : CollectorM Unit := do
if ( get).done then return ()
match e with
| .app .. =>
let f := e.getAppFn
if ( isPatternFnCandidate f) then
let saved getThe NormalizePattern.State
try
trace[grind.ematch.pattern.search] "candidate: {e}"
let p := e.abstract ( read).xs
unless p.hasLooseBVars do
trace[grind.ematch.pattern.search] "skip, does not contain pattern variables"
return ()
let p NormalizePattern.normalizePattern p
if saved.bvarsFound.size < ( getThe NormalizePattern.State).bvarsFound.size then
addNewPattern p
return ()
trace[grind.ematch.pattern.search] "skip, no new variables covered"
-- restore state and continue search
set saved
catch _ =>
-- restore state and continue search
trace[grind.ematch.pattern.search] "skip, exception during normalization"
set saved
let args := e.getAppArgs
for arg in args, flag in ( NormalizePattern.getPatternSupportMask f args.size) do
unless flag do
collect arg
| .forallE _ d b _ =>
if ( pure e.isArrow <&&> isProp d <&&> isProp b) then
collect d
collect b
| _ => return ()
private def collectPatterns? (proof : Expr) (xs : Array Expr) (searchPlaces : Array Expr) : MetaM (Option (List Expr × List HeadIndex)) := do
let go : CollectorM (Option (List Expr)) := do
for place in searchPlaces do
let place preprocessPattern place
collect place
if ( get).done then
return some (( get).patterns.toList)
return none
let (some ps, s) go { proof, xs } |>.run' {} |>.run {}
| return none
return some (ps, s.symbols.toList)
private def mkEMatchTheoremWithKind? (origin : Origin) (levelParams : Array Name) (proof : Expr) (kind : TheoremKind) : MetaM (Option EMatchTheorem) := do
if kind == .eqLhs then
return ( mkEMatchEqTheoremCore origin levelParams proof (normalizePattern := false) (useLhs := true))
else if kind == .eqRhs then
return ( mkEMatchEqTheoremCore origin levelParams proof (normalizePattern := false) (useLhs := false))
let type inferType proof
forallTelescopeReducing type fun xs type => do
let searchPlaces match kind with
| .fwd =>
let ps getPropTypes xs
if ps.isEmpty then
throwError "invalid `grind` forward theorem, theorem `{← origin.pp}` does not have proposional hypotheses"
pure ps
| .bwd => pure #[type]
| .default => pure <| #[type] ++ ( getPropTypes xs)
| _ => unreachable!
go xs searchPlaces
where
go (xs : Array Expr) (searchPlaces : Array Expr) : MetaM (Option EMatchTheorem) := do
let some (patterns, symbols) collectPatterns? proof xs searchPlaces
| return none
let numParams := xs.size
trace[grind.ematch.pattern] "{← origin.pp}: {patterns.map ppPattern}"
return some {
proof, patterns, numParams, symbols
levelParams, origin
}
private def getKind (stx : Syntax) : TheoremKind :=
if stx[1].isNone then
.default
else if stx[1][0].getKind == ``Parser.Attr.grindEq then
.eqLhs
else if stx[1][0].getKind == ``Parser.Attr.grindFwd then
.fwd
else if stx[1][0].getKind == ``Parser.Attr.grindEqRhs then
.eqRhs
else if stx[1][0].getKind == ``Parser.Attr.grindEqBoth then
.eqBoth
else
.bwd
private def addGrindEqAttr (declName : Name) (attrKind : AttributeKind) (useLhs := true) : MetaM Unit := do
if ( getConstInfo declName).isTheorem then
ematchTheoremsExt.add ( mkEMatchEqTheorem declName (normalizePattern := true) (useLhs := useLhs)) attrKind
else if let some eqns getEqnsFor? declName then
unless useLhs do
throwError "`{declName}` is a definition, you must only use the left-hand side for extracting patterns"
for eqn in eqns do
ematchTheoremsExt.add ( mkEMatchEqTheorem eqn) attrKind
else
throwError "`[grind_eq]` attribute can only be applied to equational theorems or function definitions"
private def addGrindAttr (declName : Name) (attrKind : AttributeKind) (thmKind : TheoremKind) : MetaM Unit := do
if thmKind == .eqLhs then
addGrindEqAttr declName attrKind (useLhs := true)
else if thmKind == .eqRhs then
addGrindEqAttr declName attrKind (useLhs := false)
else if thmKind == .eqBoth then
addGrindEqAttr declName attrKind (useLhs := true)
addGrindEqAttr declName attrKind (useLhs := false)
else if !( getConstInfo declName).isTheorem then
addGrindEqAttr declName attrKind
else
let some thm mkEMatchTheoremWithKind? (.decl declName) #[] ( getProofFor declName) thmKind
| throwError "`@{thmKind.toAttribute} theorem {declName}` {thmKind.explainFailure}, consider using different options or the `grind_pattern` command"
ematchTheoremsExt.add thm attrKind
builtin_initialize
registerBuiltinAttribute {
name := `grind
descr :=
"The `[grind_eq]` attribute is used to annotate equational theorems and functions.\
When applied to an equational theorem, it marks the theorem for use in heuristic instantiations by the `grind` tactic.\
When applied to a function, it automatically annotates the equational theorems associated with that function.\
The `grind` tactic utilizes annotated theorems to add instances of matching patterns into the local context during proof search.\
For example, if a theorem `@[grind_eq] theorem foo_idempotent : foo (foo x) = foo x` is annotated,\
`grind` will add an instance of this theorem to the local context whenever it encounters the pattern `foo (foo x)`."
applicationTime := .afterCompilation
add := fun declName stx attrKind => do
addGrindAttr declName attrKind (getKind stx) |>.run' {}
}
end Lean.Meta.Grind

View File

@@ -0,0 +1,34 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Lemmas
import Lean.Meta.Tactic.Grind.Types
import Lean.Meta.Tactic.Grind.Internalize
import Lean.Meta.Tactic.Grind.Simp
namespace Lean.Meta.Grind
/--
If `parent` is a projection-application `proj_i c`,
check whether the root of the equivalence class containing `c` is a constructor-application `ctor ... a_i ...`.
If so, internalize the term `proj_i (ctor ... a_i ...)` and add the equality `proj_i (ctor ... a_i ...) = a_i`.
-/
def propagateForallProp (parent : Expr) : GoalM Unit := do
let .forallE n p q bi := parent | return ()
trace[grind.debug.forallPropagator] "{parent}"
unless ( isEqTrue p) do return ()
trace[grind.debug.forallPropagator] "isEqTrue, {parent}"
let h₁ mkEqTrueProof p
let qh₁ := q.instantiate1 (mkApp2 (mkConst ``of_eq_true) p h₁)
let r simp qh₁
let q := mkLambda n bi p q
let q' := r.expr
internalize q' ( getGeneration parent)
trace[grind.debug.forallPropagator] "q': {q'} for{indentExpr parent}"
let h₂ r.getProof
let h := mkApp5 (mkConst ``Lean.Grind.forall_propagator) p q q' h₁ h₂
pushEq parent q' h
end Lean.Meta.Grind

View File

@@ -0,0 +1,139 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Util
import Lean.Meta.LitValues
import Lean.Meta.Match.MatcherInfo
import Lean.Meta.Match.MatchEqsExt
import Lean.Meta.Tactic.Grind.Types
import Lean.Meta.Tactic.Grind.Util
namespace Lean.Meta.Grind
/-- Adds `e` to congruence table. -/
def addCongrTable (e : Expr) : GoalM Unit := do
if let some { e := e' } := ( get).congrTable.find? { e } then
-- `f` and `g` must have the same type.
-- See paper: Congruence Closure in Intensional Type Theory
let f := e.getAppFn
let g := e'.getAppFn
unless isSameExpr f g do
unless ( hasSameType f g) do
trace[grind.issues] "found congruence between{indentExpr e}\nand{indentExpr e'}\nbut functions have different types"
return ()
trace[grind.debug.congr] "{e} = {e'}"
pushEqHEq e e' congrPlaceholderProof
let node getENode e
setENode e { node with congr := e' }
else
modify fun s => { s with congrTable := s.congrTable.insert { e } }
private def updateAppMap (e : Expr) : GoalM Unit := do
let key := e.toHeadIndex
modify fun s => { s with
appMap := if let some es := s.appMap.find? key then
s.appMap.insert key (e :: es)
else
s.appMap.insert key [e]
}
mutual
/-- Internalizes the nested ground terms in the given pattern. -/
private partial def internalizePattern (pattern : Expr) (generation : Nat) : GoalM Expr := do
if pattern.isBVar || isPatternDontCare pattern then
return pattern
else if let some e := groundPattern? pattern then
let e shareCommon ( canon ( normalizeLevels ( unfoldReducible e)))
internalize e generation
return mkGroundPattern e
else pattern.withApp fun f args => do
return mkAppN f ( args.mapM (internalizePattern · generation))
private partial def activateTheorem (thm : EMatchTheorem) (generation : Nat) : GoalM Unit := do
-- Recall that we use the proof as part of the key for a set of instances found so far.
-- We don't want to use structural equality when comparing keys.
let proof shareCommon thm.proof
let thm := { thm with proof, patterns := ( thm.patterns.mapM (internalizePattern · generation)) }
trace[grind.ematch] "activated `{thm.origin.key}`, {thm.patterns.map ppPattern}"
modify fun s => { s with newThms := s.newThms.push thm }
/--
If `Config.matchEqs` is set to `true`, and `f` is `match`-auxiliary function,
adds its equations to `newThms`.
-/
private partial def addMatchEqns (f : Expr) (generation : Nat) : GoalM Unit := do
if !( getConfig).matchEqs then return ()
let .const declName _ := f | return ()
if !( isMatcher declName) then return ()
if ( get).matchEqNames.contains declName then return ()
modify fun s => { s with matchEqNames := s.matchEqNames.insert declName }
for eqn in ( Match.getEquationsFor declName).eqnNames do
-- We disable pattern normalization to prevent the `match`-expression to be reduced.
activateTheorem ( mkEMatchEqTheorem eqn (normalizePattern := false)) generation
private partial def activateTheoremPatterns (fName : Name) (generation : Nat) : GoalM Unit := do
if let some (thms, thmMap) := ( get).thmMap.retrieve? fName then
modify fun s => { s with thmMap }
let appMap := ( get).appMap
for thm in thms do
let symbols := thm.symbols.filter fun sym => !appMap.contains sym
let thm := { thm with symbols }
match symbols with
| [] => activateTheorem thm generation
| _ =>
trace[grind.ematch] "reinsert `{thm.origin.key}`"
modify fun s => { s with thmMap := s.thmMap.insert thm }
partial def internalize (e : Expr) (generation : Nat) : GoalM Unit := do
if ( alreadyInternalized e) then return ()
trace[grind.internalize] "{e}"
match e with
| .bvar .. => unreachable!
| .sort .. => return ()
| .fvar .. | .letE .. | .lam .. =>
mkENodeCore e (ctor := false) (interpreted := false) (generation := generation)
| .forallE _ d _ _ =>
mkENodeCore e (ctor := false) (interpreted := false) (generation := generation)
if ( isProp d <&&> isProp e) then
internalize d generation
registerParent e d
propagateUp e
| .lit .. | .const .. =>
mkENode e generation
| .mvar ..
| .mdata ..
| .proj .. =>
trace[grind.issues] "unexpected term during internalization{indentExpr e}"
mkENodeCore e (ctor := false) (interpreted := false) (generation := generation)
| .app .. =>
if ( isLitValue e) then
-- We do not want to internalize the components of a literal value.
mkENode e generation
else e.withApp fun f args => do
addMatchEqns f generation
if f.isConstOf ``Lean.Grind.nestedProof && args.size == 2 then
-- We only internalize the proposition. We can skip the proof because of
-- proof irrelevance
let c := args[0]!
internalize c generation
registerParent e c
else
if let .const fName _ := f then
activateTheoremPatterns fName generation
else
internalize f generation
registerParent e f
for h : i in [: args.size] do
let arg := args[i]
internalize arg generation
registerParent e arg
mkENode e generation
addCongrTable e
updateAppMap e
propagateUp e
end
end Lean.Meta.Grind

View File

@@ -0,0 +1,154 @@
/-
Copyright (c) 2025 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Lemmas
import Lean.Meta.Tactic.Assert
import Lean.Meta.Tactic.Grind.Simp
import Lean.Meta.Tactic.Grind.Types
import Lean.Meta.Tactic.Grind.Cases
import Lean.Meta.Tactic.Grind.Injection
import Lean.Meta.Tactic.Grind.Core
namespace Lean.Meta.Grind
private inductive IntroResult where
| done
| newHyp (fvarId : FVarId) (goal : Goal)
| newDepHyp (goal : Goal)
| newLocal (fvarId : FVarId) (goal : Goal)
deriving Inhabited
private def introNext (goal : Goal) (generation : Nat) : GrindM IntroResult := do
let target goal.mvarId.getType
if target.isArrow then
goal.mvarId.withContext do
let p := target.bindingDomain!
if !( isProp p) then
let (fvarId, mvarId) goal.mvarId.intro1P
return .newLocal fvarId { goal with mvarId }
else
let tag goal.mvarId.getTag
let q := target.bindingBody!
-- TODO: keep applying simp/eraseIrrelevantMData/canon/shareCommon until no progress
let r simp p
let fvarId mkFreshFVarId
let lctx := ( getLCtx).mkLocalDecl fvarId target.bindingName! r.expr target.bindingInfo!
let mvarNew mkFreshExprMVarAt lctx ( getLocalInstances) q .syntheticOpaque tag
let mvarIdNew := mvarNew.mvarId!
mvarIdNew.withContext do
let h mkLambdaFVars #[mkFVar fvarId] mvarNew
match r.proof? with
| some he =>
let hNew := mkAppN (mkConst ``Lean.Grind.intro_with_eq) #[p, r.expr, q, he, h]
goal.mvarId.assign hNew
return .newHyp fvarId { goal with mvarId := mvarIdNew }
| none =>
-- `p` and `p'` are definitionally equal
goal.mvarId.assign h
return .newHyp fvarId { goal with mvarId := mvarIdNew }
else if target.isLet || target.isForall || target.isLetFun then
let (fvarId, mvarId) goal.mvarId.intro1P
mvarId.withContext do
let localDecl fvarId.getDecl
if ( isProp localDecl.type) then
-- Add a non-dependent copy
let mvarId mvarId.assert ( mkFreshUserName localDecl.userName) localDecl.type (mkFVar fvarId)
return .newDepHyp { goal with mvarId }
else
let goal := { goal with mvarId }
if target.isLet || target.isLetFun then
let v := ( fvarId.getDecl).value
let r simp v
let x shareCommon (mkFVar fvarId)
let goal GoalM.run' goal <| addNewEq x r.expr ( r.getProof) generation
return .newLocal fvarId goal
else
return .newLocal fvarId goal
else
return .done
private def isCasesCandidate (type : Expr) : MetaM Bool := do
let .const declName _ := type.getAppFn | return false
isGrindCasesTarget declName
private def applyCases? (goal : Goal) (fvarId : FVarId) : MetaM (Option (List Goal)) := goal.mvarId.withContext do
if ( isCasesCandidate ( fvarId.getType)) then
let mvarIds cases goal.mvarId (mkFVar fvarId)
return mvarIds.map fun mvarId => { goal with mvarId }
else
return none
private def applyInjection? (goal : Goal) (fvarId : FVarId) : MetaM (Option Goal) := do
if let some mvarId injection? goal.mvarId fvarId then
return some { goal with mvarId }
else
return none
/-- Introduce new hypotheses (and apply `by_contra`) until goal is of the form `... ⊢ False` -/
partial def intros (goal : Goal) (generation : Nat) : GrindM (List Goal) := do
let rec go (goal : Goal) : StateRefT (Array Goal) GrindM Unit := do
if goal.inconsistent then
return ()
match ( introNext goal generation) with
| .done =>
if let some mvarId goal.mvarId.byContra? then
go { goal with mvarId }
else
modify fun s => s.push goal
| .newHyp fvarId goal =>
if let some goals applyCases? goal fvarId then
goals.forM go
else if let some goal applyInjection? goal fvarId then
go goal
else
go ( GoalM.run' goal <| addHypothesis fvarId generation)
| .newDepHyp goal =>
go goal
| .newLocal fvarId goal =>
if let some goals applyCases? goal fvarId then
goals.forM go
else
go goal
let (_, goals) (go goal).run #[]
return goals.toList
/-- Asserts a new fact `prop` with proof `proof` to the given `goal`. -/
def assertAt (goal : Goal) (proof : Expr) (prop : Expr) (generation : Nat) : GrindM (List Goal) := do
if ( isCasesCandidate prop) then
let mvarId goal.mvarId.assert ( mkFreshUserName `h) prop proof
let goal := { goal with mvarId }
intros goal generation
else
let goal GoalM.run' goal do
let r simp prop
let prop' := r.expr
let proof' mkEqMP ( r.getProof) proof
add prop' proof' generation
if goal.inconsistent then return [] else return [goal]
/-- Asserts next fact in the `goal` fact queue. -/
def assertNext? (goal : Goal) : GrindM (Option (List Goal)) := do
let some (fact, newFacts) := goal.newFacts.dequeue?
| return none
assertAt { goal with newFacts } fact.proof fact.prop fact.generation
partial def iterate (goal : Goal) (f : Goal GrindM (Option (List Goal))) : GrindM (List Goal) := do
go [goal] []
where
go (todo : List Goal) (result : List Goal) : GrindM (List Goal) := do
match todo with
| [] => return result
| goal :: todo =>
if let some goalsNew f goal then
go (goalsNew ++ todo) result
else
go todo (goal :: result)
/-- Asserts all facts in the `goal` fact queue. -/
partial def assertAll (goal : Goal) : GrindM (List Goal) := do
iterate goal assertNext?
end Lean.Meta.Grind

View File

@@ -0,0 +1,109 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Lean.Meta.Tactic.Grind.Types
import Lean.Meta.Tactic.Grind.Proof
namespace Lean.Meta.Grind
/-!
Debugging support code for checking basic invariants.
-/
private def checkEqc (root : ENode) : GoalM Unit := do
let mut size := 0
let mut curr := root.self
repeat
size := size + 1
-- The root of `curr` must be `root`
assert! isSameExpr ( getRoot curr) root.self
-- Check congruence root
if curr.isApp then
if let some { e } := ( get).congrTable.find? { e := curr } then
if ( hasSameType e.getAppFn curr.getAppFn) then
assert! isSameExpr e ( getCongrRoot curr)
else
assert! ( isCongrRoot curr)
-- If the equivalence class does not have HEq proofs, then the types must be definitionally equal.
unless root.heqProofs do
assert! ( hasSameType curr root.self)
-- Starting at `curr`, following the `target?` field leads to `root`.
let mut n := curr
repeat
if let some target getTarget? n then
n := target
else
break
assert! isSameExpr n root.self
-- Go to next element
curr getNext curr
if isSameExpr root.self curr then
break
-- The size of the equivalence class is correct.
assert! root.size == size
private def checkParents (e : Expr) : GoalM Unit := do
if ( isRoot e) then
for parent in ( getParents e) do
let mut found := false
let checkChild (child : Expr) : GoalM Bool := do
let some childRoot getRoot? child | return false
return isSameExpr childRoot e
-- There is an argument `arg` s.t. root of `arg` is `e`.
for arg in parent.getAppArgs do
if ( checkChild arg) then
found := true
break
-- Recall that we have support for `Expr.forallE` propagation. See `ForallProp.lean`.
if let .forallE _ d _ _ := parent then
if ( checkChild d) then
found := true
unless found do
assert! ( checkChild parent.getAppFn)
else
-- All the parents are stored in the root of the equivalence class.
assert! ( getParents e).isEmpty
private def checkPtrEqImpliesStructEq : GoalM Unit := do
let nodes getENodes
for h₁ : i in [: nodes.size] do
let n₁ := nodes[i]
for h₂ : j in [i+1 : nodes.size] do
let n₂ := nodes[j]
-- We don't have multiple nodes for the same expression
assert! !isSameExpr n₁.self n₂.self
-- and the two expressions must not be structurally equal
assert! !Expr.equal n₁.self n₂.self
private def checkProofs : GoalM Unit := do
let eqcs getEqcs
for eqc in eqcs do
for a in eqc do
for b in eqc do
unless isSameExpr a b do
let p mkEqHEqProof a b
trace[grind.debug.proofs] "{a} = {b}"
check p
trace[grind.debug.proofs] "checked: {← inferType p}"
/--
Checks basic invariants if `grind.debug` is enabled.
-/
def checkInvariants (expensive := false) : GoalM Unit := do
if grind.debug.get ( getOptions) then
for (_, node) in ( get).enodes do
checkParents node.self
if isSameExpr node.self node.root then
checkEqc node
if expensive then
checkPtrEqImpliesStructEq
if expensive && grind.debug.proofs.get ( getOptions) then
checkProofs
def Goal.checkInvariants (goal : Goal) (expensive := false) : GrindM Unit :=
discard <| GoalM.run' goal <| Grind.checkInvariants expensive
end Lean.Meta.Grind

View File

@@ -0,0 +1,86 @@
/-
Copyright (c) 2025 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Lemmas
import Lean.Meta.Tactic.Util
import Lean.Meta.Tactic.Grind.RevertAll
import Lean.Meta.Tactic.Grind.PropagatorAttr
import Lean.Meta.Tactic.Grind.Proj
import Lean.Meta.Tactic.Grind.ForallProp
import Lean.Meta.Tactic.Grind.Util
import Lean.Meta.Tactic.Grind.Inv
import Lean.Meta.Tactic.Grind.Intro
import Lean.Meta.Tactic.Grind.EMatch
import Lean.Meta.Tactic.Grind.SimpUtil
namespace Lean.Meta.Grind
def mkMethods (fallback : Fallback) : CoreM Methods := do
let builtinPropagators builtinPropagatorsRef.get
return {
fallback
propagateUp := fun e => do
propagateForallProp e
let .const declName _ := e.getAppFn | return ()
propagateProjEq e
if let some prop := builtinPropagators.up[declName]? then
prop e
propagateDown := fun e => do
let .const declName _ := e.getAppFn | return ()
if let some prop := builtinPropagators.down[declName]? then
prop e
}
def GrindM.run (x : GrindM α) (mainDeclName : Name) (config : Grind.Config) (fallback : Fallback) : MetaM α := do
let scState := ShareCommon.State.mk _
let (falseExpr, scState) := ShareCommon.State.shareCommon scState (mkConst ``False)
let (trueExpr, scState) := ShareCommon.State.shareCommon scState (mkConst ``True)
let simprocs Grind.getSimprocs
let simp Grind.getSimpContext
x ( mkMethods fallback).toMethodsRef { mainDeclName, config, simprocs, simp } |>.run' { scState, trueExpr, falseExpr }
private def mkGoal (mvarId : MVarId) : GrindM Goal := do
let trueExpr getTrueExpr
let falseExpr getFalseExpr
let thmMap getEMatchTheorems
GoalM.run' { mvarId, thmMap } do
mkENodeCore falseExpr (interpreted := true) (ctor := false) (generation := 0)
mkENodeCore trueExpr (interpreted := true) (ctor := false) (generation := 0)
private def initCore (mvarId : MVarId) : GrindM (List Goal) := do
mvarId.ensureProp
-- TODO: abstract metavars
mvarId.ensureNoMVar
let mvarId mvarId.clearAuxDecls
let mvarId mvarId.revertAll
let mvarId mvarId.unfoldReducible
let mvarId mvarId.betaReduce
let goals intros ( mkGoal mvarId) (generation := 0)
goals.forM (·.checkInvariants (expensive := true))
return goals.filter fun goal => !goal.inconsistent
def all (goals : List Goal) (f : Goal GrindM (List Goal)) : GrindM (List Goal) := do
goals.foldlM (init := []) fun acc goal => return acc ++ ( f goal)
/-- A very simple strategy -/
private def simple (goals : List Goal) : GrindM (List Goal) := do
all goals ematchStar
def main (mvarId : MVarId) (config : Grind.Config) (mainDeclName : Name) (fallback : Fallback) : MetaM (List MVarId) := do
let go : GrindM (List MVarId) := do
let goals initCore mvarId
let goals simple goals
let goals goals.filterMapM fun goal => do
if goal.inconsistent then return none
let goal GoalM.run' goal fallback
if goal.inconsistent then return none
if ( goal.mvarId.isAssigned) then return none
return some goal
trace[grind.debug.final] "{← ppGoals goals}"
return goals.map (·.mvarId)
go.run mainDeclName config fallback
end Lean.Meta.Grind

View File

@@ -0,0 +1,66 @@
/-
Copyright (c) 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Leonardo de Moura
-/
prelude
import Init.Grind.Util
import Lean.Util.PtrSet
import Lean.Meta.Basic
import Lean.Meta.InferType
namespace Lean.Meta.Grind
unsafe def markNestedProofsImpl (e : Expr) : MetaM Expr := do
visit e |>.run' mkPtrMap
where
visit (e : Expr) : StateRefT (PtrMap Expr Expr) MetaM Expr := do
if ( isProof e) then
if e.isAppOf ``Lean.Grind.nestedProof then
return e -- `e` is already marked
if let some r := ( get).find? e then
return r
let prop inferType e
let e' := mkApp2 (mkConst ``Lean.Grind.nestedProof) prop e
modify fun s => s.insert e e'
return e'
-- Remark: we have to process `Expr.proj` since we only
-- fold projections later during term internalization
unless e.isApp || e.isForall || e.isProj do
return e
-- Check whether it is cached
if let some r := ( get).find? e then
return r
let e' match e with
| .app .. => e.withApp fun f args => do
let mut modified := false
let mut args := args
for i in [:args.size] do
let arg := args[i]!
let arg' visit arg
unless ptrEq arg arg' do
args := args.set! i arg'
modified := true
if modified then
pure <| mkAppN f args
else
pure e
| .proj _ _ b =>
pure <| e.updateProj! ( visit b)
| .forallE _ d b _ =>
-- Recall that we have `ForallProp.lean`.
let d' visit d
let b' if b.hasLooseBVars then pure b else visit b
pure <| e.updateForallE! d' b'
| _ => unreachable!
modify fun s => s.insert e e'
return e'
/--
Wrap nested proofs `e` with `Lean.Grind.nestedProof`-applications.
Recall that the congruence closure module has special support for `Lean.Grind.nestedProof`.
-/
def markNestedProofs (e : Expr) : MetaM Expr :=
unsafe markNestedProofsImpl e
end Lean.Meta.Grind

Some files were not shown because too many files have changed in this diff Show More