lean4-htt/src/Lean/Meta/AbstractNestedProofs.lean
Kyle Miller 02c8c2f9e1
feat: use nondep flag in Expr.letE and LocalContext.ldecl (#8804)
This PR implements first-class support for nondependent let expressions
in the elaborator; recall that a let expression `let x : t := v; b` is
called *nondependent* if `fun x : t => b` typechecks, and the notation
for a nondependent let expression is `have x := v; b`. Previously we
encoded `have` using the `letFun` function, but now we make use of the
`nondep` flag in the `Expr.letE` constructor for the encoding. This has
been given full support throughout the metaprogramming interface and the
elaborator. Key changes to the metaprogramming interface:
- Local context `ldecl`s with `nondep := true` are generally treated as
`cdecl`s. This is because in the body of a `have` expression the
variable is opaque. Functions like `LocalDecl.isLet` by default return
`false` for nondependent `ldecl`s. In the rare case where it is needed,
they take an additional optional `allowNondep : Bool` flag (defaults to
`false`) if the variable is being processed in a context where the value
is relevant.
- Functions such as `mkLetFVars` by default generalize nondependent let
variables and create lambda expressions for them. The
`generalizeNondepLet` flag (default true) can be set to false if `have`
expressions should be produced instead. **Breaking change:** Uses of
`letLambdaTelescope`/`mkLetFVars` need to use `generalizeNondepLet :=
false`. See the next item.
- There are now some mapping functions to make telescoping operations
more convenient. See `mapLetTelescope` and `mapLambdaLetTelescope`.
There is also `mapLetDecl` as a counterpart to `withLetDecl` for
creating `let`/`have` expressions.
- Important note about the `generalizeNondepLet` flag: it should only be
used for variables in a local context that the metaprogram "owns". Since
nondependent let variables are treated as constants in most cases, the
`value` field might refer to variables that do not exist, if for example
those variables were cleared or reverted. Using `mapLetDecl` is always
fine.
- The simplifier will cache its let dependence calculations in the
nondep field of let expressions.
- The `intro` tactic still produces *dependent* local variables. Given
that the simplifier will transform lets into haves, it would be
surprising if that would prevent `intro` from creating a local variable
whose value cannot be used.

Note that nondependence of lets is not checked by the kernel. To
external checker authors: If the elaborator gets the nondep flag wrong,
we consider this to be an elaborator error. Feel free to typecheck `letE
n t v b true` as if it were `app (lam n t b default) v` and please
report issues.

This PR follows up from #8751, which made sure the nondep flag was
preserved in the C++ interface.
2025-06-22 21:54:57 +00:00

91 lines
3.8 KiB
Text

/-
Copyright (c) 2020 Microsoft Corporation. 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.Closure
import Lean.Meta.Transform
namespace Lean.Meta
/-- Abstracts the given proof into an auxiliary theorem, suitably pre-processing its type. -/
def abstractProof [Monad m] [MonadLiftT MetaM m] [MonadEnv m] [MonadOptions m] [MonadFinally m]
(proof : Expr) (cache := true) (postprocessType : Expr → m Expr := pure) : m Expr := do
let type ← withoutExporting do inferType proof
let type ← (Core.betaReduce type : MetaM _)
let type ← zetaReduce type
let type ← postprocessType type
/- We turn on zetaDelta-expansion to make sure we don't need to perform an expensive `check` step to
identify which let-decls can be abstracted. If we design a more efficient test, we can avoid the eager zetaDelta expansion step.
In a benchmark created by @selsam, The extra `check` step was a bottleneck. -/
mkAuxTheorem (cache := cache) type proof (zetaDelta := true)
namespace AbstractNestedProofs
def getLambdaBody (e : Expr) : Expr :=
match e with
| .lam _ _ b _ => getLambdaBody b
| _ => e
def isNonTrivialProof (e : Expr) : MetaM Bool := do
if !(← isProof e) then
return false
else if e.isAppOf ``Grind.nestedProof then
-- Grind.nestedProof is a gadget created by the `grind` tactic.
-- We want to avoid the situation where `grind` keeps creating them,
-- and this module, which is used by `grind`, keeps abstracting them.
return false
else
-- We consider proofs such as `fun x => f x a` as trivial.
-- For example, we don't want to abstract the body of `def rfl`
(getLambdaBody e).withApp fun f args =>
pure $ !f.isAtomic || args.any fun arg => !arg.isAtomic
structure Context where
cache : Bool
abbrev M := ReaderT Context $ MonadCacheT ExprStructEq Expr MetaM
partial def visit (e : Expr) : M Expr := do
if e.isAtomic then
pure e
else
let visitBinders (xs : Array Expr) (k : M Expr) : M Expr := do
let localInstances ← getLocalInstances
let mut lctx ← getLCtx
for x in xs do
let xFVarId := x.fvarId!
let localDecl ← xFVarId.getDecl
let type ← visit localDecl.type
let localDecl := localDecl.setType type
let localDecl ← match localDecl.value? (allowNondep := true) with
| some value => let value ← visit value; pure <| localDecl.setValue value
| none => pure localDecl
lctx := lctx.modifyLocalDecl xFVarId fun _ => localDecl
withLCtx lctx localInstances k
checkCache { val := e : ExprStructEq } fun _ => do
if (← withoutExporting do isNonTrivialProof e) then
/- Ensure proofs nested in type are also abstracted -/
abstractProof e (← read).cache visit
else match e with
| .lam ..
| .letE .. => lambdaLetTelescope e fun xs b => visitBinders xs do mkLambdaFVars xs (← visit b) (usedLetOnly := false) (generalizeNondepLet := false)
| .forallE .. => forallTelescope e fun xs b => visitBinders xs do mkForallFVars xs (← visit b)
| .mdata _ b => return e.updateMData! (← visit b)
| .proj _ _ b => return e.updateProj! (← visit b)
| .app .. => e.withApp fun f args => return mkAppN (← visit f) (← args.mapM visit)
| _ => pure e
end AbstractNestedProofs
/-- Replace proofs nested in `e` with new lemmas. The new lemmas are named using `getDeclNGen`. -/
def abstractNestedProofs (e : Expr) (cache := true) : MetaM Expr := do
if (← isProof e) then
-- `e` is a proof itself. So, we don't abstract nested proofs
return e
else
AbstractNestedProofs.visit e |>.run { cache } |>.run
end Lean.Meta