Fix: docstring_parser/numpydoc.py maps both 'Parameters' and 'Attributes'... - #114
Open
M001N wants to merge 2 commits into
Open
Fix: docstring_parser/numpydoc.py maps both 'Parameters' and 'Attributes'...#114M001N wants to merge 2 commits into
M001N wants to merge 2 commits into
Conversation
Docstring.params is meant to list function/constructor parameters, but for NUMPYDOC-style docstrings it also picked up entries from a NumPy- style Attributes section (class attributes), since numpydoc.py tags both kinds with the same DocstringParam type and only differs via item.args[0] (param vs attribute). Fix .params to exclude attribute-tagged items, but only when the docstring style is NUMPYDOC. Other styles are left untouched: Google-style intentionally folds its own Attributes section into .params (see test_google.py::test_attributes) and the attribute- docstring merging feature (attrdoc.py / parse_from_object) relies on attribute-tagged DocstringParam items appearing in .params too, for any style. Narrowing the filter unconditionally would have broken both. Also update numpydoc.compose() to build its Parameters/Attributes/ Receives/Other Parameters sections from docstring.meta instead of the now-narrower docstring.params, so composing a previously-parsed docstring still round-trips its Attributes/Receives/Other Parameters sections correctly. Add regression tests in test_numpydoc.py covering the reporter's Attributes-only reproducer and a docstring with both Parameters and Attributes sections. Update the existing test_attributes to assert the corrected .params behavior (verifying detailed field parsing via docstring.meta instead).
Docstring.params now excludes NumpydocStyle attribute-tagged entries (fix for rr-#21). attrdoc.add_attribute_docstrings() relied on docstring.params to know which attribute names were already documented, which meant an attribute already covered by an explicit numpydoc Attributes section was no longer recognized as such, causing a duplicate DocstringParam to be appended from the source-level docstring. Dedup now checks all DocstringParam items in docstring.meta regardless of style tag. Adds a regression test in test_parse_from_object.py reproducing parse_from_object() on a class with both a numpydoc Attributes section and a matching source-level attribute docstring.
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
Changed attrdoc.add_attribute_docstrings() to build its dedup set from all DocstringParam items in docstring.meta (regardless of style-specific args[0] tag) instead of the narrower docstring.params property, so attributes already documented via a numpydoc Attributes section are correctly recognized as already-covered and not duplicated when merging in source-level attribute docstrings. (The Docstring.params filter itself in common.py was already fixed in commit b4100c8, prior to this session's work.)
Problem
rr-/docstring_parser issue reference: rr-/docstring_parser#21
Root Cause
docstring_parser/numpydoc.py maps both 'Parameters' and 'Attributes' headers to DocstringParam via ParamSection, distinguished only by item.args[0] ('param' vs 'attribute'). Docstring.params (common.py) filtered purely by isinstance(item, DocstringParam), so Attributes-section entries leaked into params for NUMPYDOC-style docstrings. This part was already fixed in a prior commit (b4100c8) by narrowing Docstring.params to exclude NUMPYDOC items tagged 'attribute'. However, that narrowing had an unaddressed side effect: attrdoc.py's add_attribute_docstrings() built its 'already documented' name set from docstring.params (now narrower for NUMPYDOC), so an attribute already covered by an explicit Attributes section was no longer recognized as documented, and a duplicate DocstringParam was appended from the source-level docstring parse.
Testing
PASS - full suite 256 passed, including the new regression test and the previously-added test_numpydoc.py::test_params_excludes_attributes and related tests.
Related Issue
rr-/docstring_parser#21