Skip to content

Clean up stale docstrings, history comments, and leftover CI/test cache files #76

Closed

Description

@iskandr

These are small, and each item is a deletion or a one-line docstring fix. I left them out of #74 so that PR stays docs-only.

Docstrings and comments that are wrong or narrate history

  • gtfparse/read_gtf.py:279: read_gtf says it returns "a dictionary mapping column names to sequences of values". It returns a Polars DataFrame by default.
  • gtfparse/read_gtf.py:183: parse_gtf_and_expand_attributes documents a chunksize parameter it doesn't take.
  • gtfparse/read_gtf.py:419-426: an unrecognized result_type silently returns None. It should raise ValueError.
  • gtfparse/write_gtf.py:61-63: the docstring refers to "the naive if value: check in earlier drafts".
  • gtfparse/attribute_parsing.py:55-56: the comment explains intern(str(...)) (:81) as a Python 2 unicode workaround. On Python 3 the str() call does nothing.
  • tests/test_gencode_gtf.py:17,102,160: section dividers labeled with issue numbers (read_gtf attribute_aliases option for GENCODE biotype column names #63, Optional integer-typed version columns (gene_version, transcript_version, protein_version, exon_version) #64). tests/test_expand_attribute_column_false.py:1-4 describes the old NameError instead of the behavior under test.
  • pyproject.toml:74-80: ruff ignore comments like "out of scope for config PR".

Stale infrastructure

  • test.sh:3-4 points to ~/code/trufflepig/test.sh for its rationale. That path exists only on a local machine.
  • .github/workflows/tests.yml:4-8: the TODO block (cache ~/.cache/pyensembl, conda binaries, add a tests badge, update coveralls) doesn't apply. The badge is already in the README and Coveralls already uses coverallsapp/github-action@v2.2.3.
  • .hypothesis/unicode_data/13.0.0/charmap.json.gz and tests/.hypothesis/... are committed Hypothesis caches. Nothing imports Hypothesis. Delete them and add .hypothesis/ to .gitignore.

format.sh (apply formatting), lint.sh (ruff check plus format check), and lint-and-test.sh (lint, then test) don't overlap, so I'd keep all three.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions