[ty] Use `ThinVec` for sub segments in `PlaceExpr` #19470

MichaReiser · 2025-07-21T17:51:44Z

Summary

This replaces the SmallVec on PlaceExpr with a ThinVec. The main motivation is that many places (all symbols) have no sub segments and paying the extra cost of an empty vec (32 bytes) for each of those adds quiet a bit to overall memory usage

This reduces the place_table size by roughly 25%

MichaReiser · 2025-07-21T17:54:15Z

crates/ty_python_semantic/src/semantic_index/place.rs

+fn sub_segments_size(segments: &thin_vec::ThinVec<PlaceExprSubSegment>) -> usize {
+    segments.capacity() * std::mem::size_of::<PlaceExprSubSegment>()
+        + segments
+            .iter()
+            .map(|segment| segment.get_heap_size())
+            .sum::<usize>()
 }


I already PRd (and it's merged but not yet released) adding thin_vec support to the getsize crate

github-actions · 2025-07-21T17:55:27Z

`mypy_primer` results

No ecosystem changes detected ✅

Memory usage changes were detected when running on open source projects

sphinx (https://github.com/sphinx-doc/sphinx)
-     memo fields = ~236MB
+     memo fields = ~224MB

prefect (https://github.com/PrefectHQ/prefect)
- TOTAL MEMORY USAGE: ~626MB
+ TOTAL MEMORY USAGE: ~596MB

github-actions · 2025-07-21T18:02:43Z

`ruff-ecosystem` results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

ℹ️ ecosystem check encountered format errors. (no format changes; 1 project error)

openai/openai-cookbook (error)

warning: Detected debug build without --no-cache.
error: Failed to parse examples/mcp/databricks_mcp_cookbook.ipynb:12:1:8: Simple statements must be separated by newlines or semicolons

Formatter (preview)

ℹ️ ecosystem check encountered format errors. (no format changes; 1 project error)

openai/openai-cookbook (error)

ruff format --preview

warning: Detected debug build without --no-cache.
error: Failed to parse examples/mcp/databricks_mcp_cookbook.ipynb:12:1:8: Simple statements must be separated by newlines or semicolons

MichaReiser · 2025-07-21T18:14:41Z

crates/ty_python_semantic/src/semantic_index/place.rs

@@ -775,7 +786,7 @@ impl std::fmt::Debug for PlaceTable {
 pub(super) struct PlaceTableBuilder {
    table: PlaceTable,

-    associated_place_ids: IndexVec<ScopedPlaceId, Vec<ScopedPlaceId>>,
+    associated_place_ids: IndexVec<ScopedPlaceId, SmallVec<[ScopedPlaceId; 4]>>,


This SmallVec here is "free" as 4 ScopedPlaceIds + the tag fit into 24 bytes

AlexWaygood · 2025-07-21T18:16:58Z

Nice, looks like this yields some small speedups too! https://codspeed.io/astral-sh/ruff/branches/micha%2Fplace-table-thin-vec?runnerMode=Instrumentation

AlexWaygood · 2025-07-21T18:30:17Z

I think PlaceExprSubSegment::StringSubscript could also be relatively easily changed so that it wraps a Box<str> rather than a String -- not sure if that's worth doing or not

MichaReiser · 2025-07-21T19:30:54Z

I think PlaceExprSubSegment::StringSubscript could also be relatively easily changed so that it wraps a Box rather than a String -- not sure if that's worth doing or not

This, unfortunately, doesn't help because Name is as big as a String.

An alternative to what I've done here is to split PlaceTable and internally track:

symbol_map: HashTable<ScopedSymbolId>,
symbols: IndexVec<ScopedSymbolId, Symbol>
member_map: HashTable<ScopedMemberId>,
members: IndexVec<ScopedMemberId, Member>, Can use a SmallVec because each element has at least one item.
ScopedPlaceId becomes an enum over ScopedSymbolId and ScopedMemberId.

I suspect that this might be slightly larger refactor but it might also help to untangle some of the complexity in PlaceTable

carljm

Nice!

carljm · 2025-07-21T22:17:31Z

An alternative to what I've done here is to split PlaceTable

Yes, I thought about this when reviewing the PR that added PlaceTable; I think it might be a nice refactor.

MichaReiser · 2025-07-22T05:51:27Z

Yes, I thought about this when reviewing the PR that added PlaceTable; I think it might be a nice refactor.

I'll try to find time to look into this because it might allow us to reduce memory usage even more by changing the Member definition from Name, Segments to ScopedSymbolId, Segments or just a Segments portion

MichaReiser · 2025-07-22T18:17:51Z

Yes, I thought about this when reviewing the PR that added PlaceTable; I think it might be a nice refactor.

I think this would be a very nice refactor but it's also not trivial. I started and I've currently 188 errors... and many of them aren't very trivial to fix.

I do think that it would help clarify many things and having separate types for Symbol and Member even allows stricter typing in many cases. But, as it often is with these kind of refactors, introducing those abstractions later is sort of a pain

MichaReiser · 2025-07-22T18:49:10Z

The main blocker here is that UseDefMap uses a bunch of IndexVec<ScopedPlaceId>. That prevents us from splitting the places IndexVec into two.

We could split the hash map but there's little value in that because the main motivation is to reduce the size of PlaceExpr which is stored in the IndexVec.

* main: [ty] Use `ThinVec` for sub segments in `PlaceExpr` (#19470) [ty] Splat variadic arguments into parameter list (#18996) [`flake8-pyi`] Skip fix if all `Union` members are `None` (`PYI016`) (#19416) Skip notebook with errors in ecosystem check (#19491) [ty] Consistent use of American english (in rules) (#19488) [ty] Support iterating over enums (#19486) Fix panic for illegal `Literal[…]` annotations with inner subscript expressions (#19489) Move fix suggestion to subdiagnostic (#19464) [ty] Implement non-stdlib stub mapping for classes and functions (#19471) [ty] Disallow illegal uses of `ClassVar` (#19483) [ty] Disallow `Final` in function parameter/return-type annotations (#19480) [ty] Extend `Final` test suite (#19476) [ty] Minor change to diagnostic message for invalid Literal uses (#19482) [ty] Detect illegal non-enum attribute accesses in Literal annotation (#19477) [ty] Reduce size of `TypeInference` (#19435) Run MD tests for Markdown-only changes (#19479) Revert "[ty] Detect illegal non-enum attribute accesses in Literal annotation" [ty] Detect illegal non-enum attribute accesses in Literal annotation [ty] Added semantic token support for more identifiers (#19473) [ty] Make tuple subclass constructors sound (#19469)

* main: (28 commits) [ty] highlight the argument in `static_assert` error messages (astral-sh#19426) [ty] Infer single-valuedness for enums based on `int`/`str` (astral-sh#19510) [ty] Restructure submodule query around `File` dependency [ty] Make `Module` a Salsa ingredient [ty] Reachability analysis for `isinstance(…)` branches (astral-sh#19503) [ty] Normalize single-member enums to their instance type (astral-sh#19502) [ty] Invert `ty_ide` and `ty_project` dependency (astral-sh#19501) [ty] Implement mock language server for testing (astral-sh#19391) [ty] Detect enums if metaclass is a subtype of EnumType/EnumMeta (astral-sh#19481) [ty] perform type narrowing for places marked `global` too (astral-sh#19381) [ty] Use `ThinVec` for sub segments in `PlaceExpr` (astral-sh#19470) [ty] Splat variadic arguments into parameter list (astral-sh#18996) [`flake8-pyi`] Skip fix if all `Union` members are `None` (`PYI016`) (astral-sh#19416) Skip notebook with errors in ecosystem check (astral-sh#19491) [ty] Consistent use of American english (in rules) (astral-sh#19488) [ty] Support iterating over enums (astral-sh#19486) Fix panic for illegal `Literal[…]` annotations with inner subscript expressions (astral-sh#19489) Move fix suggestion to subdiagnostic (astral-sh#19464) [ty] Implement non-stdlib stub mapping for classes and functions (astral-sh#19471) [ty] Disallow illegal uses of `ClassVar` (astral-sh#19483) ... # Conflicts: # crates/ty_ide/src/goto.rs

[ty] Use ThinVec for sub segments in PlaceExpr

035a692

MichaReiser added internal An internal refactor or improvement ty Multi-file analysis & type inference labels Jul 21, 2025

MichaReiser requested review from carljm, AlexWaygood, sharkdp and dcreager as code owners July 21, 2025 17:51

MichaReiser added internal An internal refactor or improvement ty Multi-file analysis & type inference labels Jul 21, 2025

MichaReiser self-assigned this Jul 21, 2025

MichaReiser marked this pull request as draft July 21, 2025 17:51

MichaReiser commented Jul 21, 2025

View reviewed changes

Use SmallVec for associated_place_ids

08926a4

MichaReiser marked this pull request as ready for review July 21, 2025 18:13

MichaReiser commented Jul 21, 2025

View reviewed changes

AlexWaygood approved these changes Jul 21, 2025

View reviewed changes

carljm approved these changes Jul 21, 2025

View reviewed changes

MichaReiser merged commit dc10ab8 into main Jul 22, 2025
37 checks passed

MichaReiser deleted the micha/place-table-thin-vec branch July 22, 2025 18:39

MichaReiser mentioned this pull request Jul 24, 2025

[ty] Split ScopedPlaceId into ScopedSymbolId and ScopedMemberId #19497

Merged

AlexWaygood pushed a commit that referenced this pull request Jul 25, 2025

[ty] Use ThinVec for sub segments in PlaceExpr (#19470)

f165687

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[ty] Use `ThinVec` for sub segments in `PlaceExpr` #19470

[ty] Use `ThinVec` for sub segments in `PlaceExpr` #19470

Uh oh!

MichaReiser commented Jul 21, 2025 •

edited

Loading

Uh oh!

MichaReiser Jul 21, 2025

Uh oh!

github-actions bot commented Jul 21, 2025 •

edited

Loading

Uh oh!

github-actions bot commented Jul 21, 2025 •

edited

Loading

Uh oh!

MichaReiser Jul 21, 2025

Uh oh!

AlexWaygood commented Jul 21, 2025

Uh oh!

AlexWaygood commented Jul 21, 2025

Uh oh!

MichaReiser commented Jul 21, 2025

Uh oh!

carljm left a comment

Uh oh!

carljm commented Jul 21, 2025

Uh oh!

MichaReiser commented Jul 22, 2025

Uh oh!

MichaReiser commented Jul 22, 2025

Uh oh!

Uh oh!

MichaReiser commented Jul 22, 2025

Uh oh!

Uh oh!

[ty] Use ThinVec for sub segments in PlaceExpr #19470

[ty] Use ThinVec for sub segments in PlaceExpr #19470

Uh oh!

Conversation

MichaReiser commented Jul 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Uh oh!

MichaReiser Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Jul 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

mypy_primer results

Uh oh!

github-actions bot commented Jul 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

ruff-ecosystem results

Linter (stable)

Linter (preview)

Formatter (stable)

Formatter (preview)

Uh oh!

MichaReiser Jul 21, 2025

Choose a reason for hiding this comment

Uh oh!

AlexWaygood commented Jul 21, 2025

Uh oh!

AlexWaygood commented Jul 21, 2025

Uh oh!

MichaReiser commented Jul 21, 2025

Uh oh!

carljm left a comment

Choose a reason for hiding this comment

Uh oh!

carljm commented Jul 21, 2025

Uh oh!

MichaReiser commented Jul 22, 2025

Uh oh!

MichaReiser commented Jul 22, 2025

Uh oh!

Uh oh!

MichaReiser commented Jul 22, 2025

Uh oh!

Uh oh!

[ty] Use `ThinVec` for sub segments in `PlaceExpr` #19470

[ty] Use `ThinVec` for sub segments in `PlaceExpr` #19470

MichaReiser commented Jul 21, 2025 •

edited

Loading

github-actions bot commented Jul 21, 2025 •

edited

Loading

`mypy_primer` results

github-actions bot commented Jul 21, 2025 •

edited

Loading

`ruff-ecosystem` results