Skip to content

Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.

docs(core): tighten Python group rationale to reflect measured shape

thejesh23e0630056d3
diff --git a/understand-anything-plugin/packages/core/src/ignore-generator.ts b/understand-anything-plugin/packages/core/src/ignore-generator.ts
index be33ea8f..cf90a3b4 100644
--- a/understand-anything-plugin/packages/core/src/ignore-generator.ts
+++ b/understand-anything-plugin/packages/core/src/ignore-generator.ts
@@ -95,10 +95,14 @@ const TEST_PATTERN_GROUPS: Array<{ label: string; patterns: string[] }> = [
],
},
{
- // pytest / unittest discovery is filename-based, so large Python
- // codebases (pandas, numpy, scikit-learn, tensorflow) commonly ship
- // test_*.py or *_test.py files co-located with the module under test
- // rather than clustered in a single tests/ tree.
+ // Python testing conventions are bimodal. Most projects (django,
+ // flask, pandas, numpy) cluster tests inside a top-level tests/ dir,
+ // where the existing directory rules already catch them. But Google-
+ // style codebases (tensorflow, jax, some Meta libs) interleave
+ // *_test.py directly alongside the module under test — e.g. tensor-
+ // flow/python/ops/array_ops.py + array_ops_test.py — so file-pattern
+ // rules add the majority of the token savings for that half of the
+ // ecosystem.
label: "Python",
patterns: [
"**/test_*.py",