tesserakit-docs

Measure Python docstring coverage for public symbols.

tessera-docs parses Python source with the standard-library ast module (it never imports or runs the code), inventories every documentable symbol (modules, classes, functions, methods), and reports which public ones lack docstrings.

Coverage check

tessera docs coverage --input . --output ./out/docs_pack
tessera docs coverage --input . --include-tests   # also scan test files

Test files are excluded by default; pass --include-tests to include them.

Artifacts written:

symbols.jsonl            one DocSymbol per symbol (kind, public, has_docstring, line)
index.md                 coverage headline
validation_report.md     missing-docstring findings + low-coverage warning
coverage_report.md       coverage by kind and lowest-coverage files
undocumented.md          every undocumented public symbol with file:line

What counts

• Public = name does not start with _ (so _private and __dunder__ are excluded).
• Symbol kinds: module, class, function, method.
• A symbol is documented if ast.get_docstring returns a value.

Findings

• missing_module_docstring (info)
• missing_class_docstring, missing_function_docstring, missing_method_docstring (warning)
• low_doc_coverage — overall public coverage below 80%
• parse_error — a file could not be parsed
• no_public_symbols — nothing public found