Import PHP include and require targets

[yzxcj797]

Summary

PHP use declarations produced import edges, but procedural PHP files wired together with include, include_once, require, and require_once had no IMPORTS_FROM edges. Impact radius and importer queries therefore saw no cross-file structure for non-Composer codebases.

The parser now treats all four include/require expression forms as import nodes, extracts their string argument, and resolves relative paths ending in .php without appending a second extension.

Fixes #819.

Testing

• Added end-to-end regressions for:
  • require_once 'includes/conexao.php';
  • include_once 'helper.php'.
• Both tests fail on clean main with zero import edges and pass with the fix.
• pytest tests/test_pr819_php_includes.py -q — 2 passed
• pytest tests/test_php_scoped_calls.py tests/test_php_laravel.py tests/test_pr819_php_includes.py -q — 56 passed, 2 skipped, plus one pre-existing Windows path-separator assertion unchanged by this patch
• ruff check on changed files — passes
• git diff --check — passes

[github-actions]

code-review-graph review

Overall risk: 0.30 (LOW) — 5 changed function(s)/class(es), 0 affected flow(s), 2 test gap(s)

Risk-scored changes

Risk	Level	Symbol	Location	Tested
0.30	low	code_review_graph/parser.py::CodeParser	code_review_graph/parser.py:2411	yes
0.30	low	code_review_graph/parser.py::CodeParser._extract_import	code_review_graph/parser.py:15499	no
0.10	low	code_review_graph/parser.py::CodeParser._do_resolve_module	code_review_graph/parser.py:13602	no
0.05	low	tests/test_pr819_php_includes.py::test_require_once_resolves_relative_php_file	tests/test_pr819_php_includes.py:8	(test)
0.05	low	tests/test_pr819_php_includes.py::test_include_once_resolves_relative_php_file	tests/test_pr819_php_includes.py:26	(test)

Test gaps

• code_review_graph/parser.py::CodeParser._do_resolve_module (code_review_graph/parser.py:13602)
• code_review_graph/parser.py::CodeParser._extract_import (code_review_graph/parser.py:15499)

Token savings: this graph-backed report used ~163,666 fewer tokens (~99%) than reading every changed file in full (estimated, chars/4 approximation).

---

Powered by code-review-graph — local-first analysis; no code leaves the CI runner.