feat: add Solidity language extractor#1874
Open
babyblueviper1 wants to merge 1 commit into
Open
Conversation
Structural extraction over tree-sitter-solidity: contracts, interfaces, libraries, functions/constructors/modifiers, structs, enums, and events, plus inheritance (`is A, B`), imports (named/namespace/bare), and same-file call resolution -- including modifier invocations, which are Solidity- specific call sites with no explicit call syntax. Follows the existing extractor pattern (rust.py/verilog.py): add_node/ add_edge helpers, the sourceless-stub convention for cross-file references (ensure_named_node), and the label_to_nid same-file call resolution rust.py uses, extended to Solidity's `X.method()` / bare-call / modifier-invocation call sites. Scope, disclosed in the module docstring: call resolution is by same-file method-NAME matching, not receiver-type checking, so it can occasionally mismatch if two unrelated types in the same file share a method name (same tradeoff rust.py's label_to_nid already makes) and never resolves a call to something this file doesn't itself define (raw_calls carries those for a possible future cross-file pass). Full type-qualified receiver resolution, the kind the larger OOP-language blocks in extract.py do, is out of scope for this first cut. Tested against 16 real, non-vendored .sol files from a live project (contracts + Foundry tests) with zero extraction errors, correct intra-contract call graphs (verified a real internal helper-function call chain resolves correctly), and correct inheritance/import edges. - graphify/extractors/solidity.py: extract_solidity() - graphify/extract.py: import + `.sol` dispatch entry - graphify/detect.py: `.sol` added to CODE_EXTENSIONS (also covers watch.py's _WATCHED_EXTENSIONS, which derives from it) - pyproject.toml: tree-sitter-solidity>=1.0,<2.0 dependency - tests/fixtures/sample.sol + 14 new tests in tests/test_multilang.py covering nodes, inherits, method/contains, imports, same-file calls, modifier-invocation calls, confidence, dangling-edge safety, and the missing-file error path
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a Solidity extractor (
graphify/extractors/solidity.py) overtree-sitter-solidity, following the existing per-language extractor pattern in this package (closest analog:rust.py).Extracts: contracts, interfaces, libraries, functions/constructors/modifiers, structs, enums, events, inheritance (
is A, B), imports (named{A, B as C}, namespace* as X, and bare), and same-file call resolution — including modifier invocations, which are a Solidity-specific call site with no explicit call syntax.Scope (disclosed honestly, not oversold)
Call resolution here is by same-file method-name matching, not receiver-type checking — matches rust.py's
label_to_nidtradeoff exactly. This gets the common cases right (a storage variable typed as a locally-defined interface,Library.staticMethod(),InterfaceName(addr)cast-calls, modifier invocations) but:raw_callsfor a possible future cross-file pass, same as several other extractors already doFull type-qualified receiver resolution (the kind the larger OOP-language blocks in
extract.pydo for Java/C#/TS) is out of scope for this first cut — happy to follow up if that's wanted.Testing
tests/test_multilang.py+tests/fixtures/sample.sol, covering nodes, inherits, method/contains, import context, same-file call resolution, modifier-invocation calls, confidence, dangling-edge safety, and the missing-file error path. All pass.pytest tests/ -qon the full suite: no new failures introduced (confirmed by running the same set of failing tests against the cleanv8HEAD before my change — same ~24 pre-existing failures on unrelated modules with missing optional deps/fixtures, e.g.tree-sitter-hclnot installed,test_skillgen.pyaudit baselines, one pre-existing test-ordering flake intest_cache.pyunrelated to any language extractor)..solfiles from a live project (contracts + Foundry tests) — zero extraction errors, and a real internal helper-function call chain (validationRequest() → _authorized() → _agentOwner()) resolved correctly end to end.ruff check --select E9,F63,F7,F82clean on the new/touched files.Checklist (per ARCHITECTURE.md § Adding a new language extractor)
extract_solidity(path: Path) -> dictfollowing the existing pattern.solinextract.py's_DISPATCH.soltoCODE_EXTENSIONSindetect.py(coverswatch.py's_WATCHED_EXTENSIONS, which derives from it — no separate edit needed there)tree-sitter-solidity>=1.0,<2.0topyproject.tomltests/test_multilang.py