[ty] AST garbage collection (#18482)

## Summary

Garbage collect ASTs once we are done checking a given file. Queries
with a cross-file dependency on the AST will reparse the file on demand.
This reduces ty's peak memory usage by ~20-30%.

The primary change of this PR is adding a `node_index` field to every
AST node, that is assigned by the parser. `ParsedModule` can use this to
create a flat index of AST nodes any time the file is parsed (or
reparsed). This allows `AstNodeRef` to simply index into the current
instance of the `ParsedModule`, instead of storing a pointer directly.

The indices are somewhat hackily (using an atomic integer) assigned by
the `parsed_module` query instead of by the parser directly. Assigning
the indices in source-order in the (recursive) parser turns out to be
difficult, and collecting the nodes during semantic indexing is
impossible as `SemanticIndex` does not hold onto a specific
`ParsedModuleRef`, which the pointers in the flat AST are tied to. This
means that we have to do an extra AST traversal to assign and collect
the nodes into a flat index, but the small performance impact (~3% on
cold runs) seems worth it for the memory savings.

Part of https://github.com/astral-sh/ty/issues/214.
This commit is contained in:
Ibraheem Ahmed 2025-06-13 08:40:11 -04:00 committed by GitHub
parent 76d9009a6e
commit c9dff5c7d5
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
824 changed files with 25243 additions and 804 deletions

View file

@ -1,20 +1,22 @@
---
source: crates/ruff_python_parser/tests/fixtures.rs
input_file: crates/ruff_python_parser/resources/valid/expressions/tuple.py
snapshot_kind: text
---
## AST
```
Module(
ModModule {
node_index: AtomicNodeIndex(..),
range: 0..276,
body: [
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 19..21,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 19..21,
elts: [],
ctx: Load,
@ -25,9 +27,11 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 22..26,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 23..25,
elts: [],
ctx: Load,
@ -38,13 +42,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 27..37,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 27..37,
elts: [
Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 29..31,
elts: [],
ctx: Load,
@ -53,6 +60,7 @@ Module(
),
Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 34..36,
elts: [],
ctx: Load,
@ -68,13 +76,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 38..42,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 38..42,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 39..40,
id: Name("a"),
ctx: Load,
@ -89,13 +100,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 43..49,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 43..49,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 44..45,
id: Name("a"),
ctx: Load,
@ -103,6 +117,7 @@ Module(
),
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 47..48,
id: Name("b"),
ctx: Load,
@ -117,13 +132,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 50..57,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 50..57,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 51..52,
id: Name("a"),
ctx: Load,
@ -131,6 +149,7 @@ Module(
),
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 54..55,
id: Name("b"),
ctx: Load,
@ -145,13 +164,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 58..66,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 59..65,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 60..61,
id: Name("a"),
ctx: Load,
@ -159,6 +181,7 @@ Module(
),
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 63..64,
id: Name("b"),
ctx: Load,
@ -173,13 +196,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 90..92,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 90..92,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 90..91,
id: Name("a"),
ctx: Load,
@ -194,13 +220,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 93..97,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 93..97,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 93..94,
id: Name("a"),
ctx: Load,
@ -208,6 +237,7 @@ Module(
),
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 96..97,
id: Name("b"),
ctx: Load,
@ -222,13 +252,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 98..103,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 98..103,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 98..99,
id: Name("a"),
ctx: Load,
@ -236,6 +269,7 @@ Module(
),
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 101..102,
id: Name("b"),
ctx: Load,
@ -250,16 +284,20 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 126..129,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 126..129,
elts: [
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 126..128,
value: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 127..128,
id: Name("a"),
ctx: Load,
@ -277,13 +315,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 130..135,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 130..135,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 130..131,
id: Name("a"),
ctx: Load,
@ -291,9 +332,11 @@ Module(
),
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 133..135,
value: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 134..135,
id: Name("b"),
ctx: Load,
@ -311,19 +354,24 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 136..161,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 136..161,
elts: [
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 136..142,
value: BinOp(
ExprBinOp {
node_index: AtomicNodeIndex(..),
range: 137..142,
left: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 137..138,
id: Name("a"),
ctx: Load,
@ -332,6 +380,7 @@ Module(
op: BitOr,
right: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 141..142,
id: Name("b"),
ctx: Load,
@ -344,12 +393,15 @@ Module(
),
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 144..152,
value: Await(
ExprAwait {
node_index: AtomicNodeIndex(..),
range: 145..152,
value: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 151..152,
id: Name("x"),
ctx: Load,
@ -362,6 +414,7 @@ Module(
),
Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 154..156,
elts: [],
ctx: Load,
@ -370,9 +423,11 @@ Module(
),
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 158..161,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 159..161,
elts: [],
ctx: Load,
@ -391,16 +446,20 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 162..167,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 162..167,
elts: [
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 163..165,
value: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 164..165,
id: Name("a"),
ctx: Load,
@ -418,13 +477,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 168..175,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 168..175,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 169..170,
id: Name("a"),
ctx: Load,
@ -432,9 +494,11 @@ Module(
),
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 172..174,
value: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 173..174,
id: Name("b"),
ctx: Load,
@ -452,19 +516,24 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 176..203,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 176..203,
elts: [
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 177..183,
value: BinOp(
ExprBinOp {
node_index: AtomicNodeIndex(..),
range: 178..183,
left: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 178..179,
id: Name("a"),
ctx: Load,
@ -473,6 +542,7 @@ Module(
op: BitOr,
right: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 182..183,
id: Name("b"),
ctx: Load,
@ -485,12 +555,15 @@ Module(
),
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 185..193,
value: Await(
ExprAwait {
node_index: AtomicNodeIndex(..),
range: 186..193,
value: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 192..193,
id: Name("x"),
ctx: Load,
@ -503,6 +576,7 @@ Module(
),
Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 195..197,
elts: [],
ctx: Load,
@ -511,9 +585,11 @@ Module(
),
Starred(
ExprStarred {
node_index: AtomicNodeIndex(..),
range: 199..202,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 200..202,
elts: [],
ctx: Load,
@ -532,16 +608,20 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 224..233,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 224..233,
elts: [
Named(
ExprNamed {
node_index: AtomicNodeIndex(..),
range: 225..231,
target: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 225..226,
id: Name("x"),
ctx: Store,
@ -549,6 +629,7 @@ Module(
),
value: NumberLiteral(
ExprNumberLiteral {
node_index: AtomicNodeIndex(..),
range: 230..231,
value: Int(
1,
@ -566,13 +647,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 234..245,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 234..245,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 235..236,
id: Name("x"),
ctx: Load,
@ -580,9 +664,11 @@ Module(
),
Named(
ExprNamed {
node_index: AtomicNodeIndex(..),
range: 238..244,
target: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 238..239,
id: Name("y"),
ctx: Store,
@ -590,6 +676,7 @@ Module(
),
value: NumberLiteral(
ExprNumberLiteral {
node_index: AtomicNodeIndex(..),
range: 243..244,
value: Int(
2,
@ -607,13 +694,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 246..260,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 246..260,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 247..248,
id: Name("x"),
ctx: Load,
@ -621,9 +711,11 @@ Module(
),
Named(
ExprNamed {
node_index: AtomicNodeIndex(..),
range: 250..256,
target: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 250..251,
id: Name("y"),
ctx: Store,
@ -631,6 +723,7 @@ Module(
),
value: NumberLiteral(
ExprNumberLiteral {
node_index: AtomicNodeIndex(..),
range: 255..256,
value: Int(
2,
@ -641,6 +734,7 @@ Module(
),
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 258..259,
id: Name("z"),
ctx: Load,
@ -655,13 +749,16 @@ Module(
),
Expr(
StmtExpr {
node_index: AtomicNodeIndex(..),
range: 261..275,
value: Tuple(
ExprTuple {
node_index: AtomicNodeIndex(..),
range: 261..275,
elts: [
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 261..262,
id: Name("x"),
ctx: Load,
@ -669,9 +766,11 @@ Module(
),
Named(
ExprNamed {
node_index: AtomicNodeIndex(..),
range: 265..271,
target: Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 265..266,
id: Name("y"),
ctx: Store,
@ -679,6 +778,7 @@ Module(
),
value: NumberLiteral(
ExprNumberLiteral {
node_index: AtomicNodeIndex(..),
range: 270..271,
value: Int(
2,
@ -689,6 +789,7 @@ Module(
),
Name(
ExprName {
node_index: AtomicNodeIndex(..),
range: 274..275,
id: Name("z"),
ctx: Load,