lint-code

name	lint-code
description	Check code style against SUEWS conventions. Complements ruff/fprettify.

Lint Code

Check code against project conventions in .claude/rules/.

Workflow

1. Run standard tools: ruff check, fprettify --diff
2. For changed Python files anywhere in the repo, run ruff check --select D to surface numpy-style docstring violations (the global ruff check already enforces D, but this scoped call keeps docstring findings separated in the report).
   • File scope: git diff --name-only origin/master...HEAD -- '*.py'
   • Legacy debt is parked in [lint.per-file-ignores] in .ruff.toml (seeded from the whole repo, not just src/supy/); any finding reported here is new debt introduced on this branch.
3. Apply SUEWS-specific checks from .claude/rules/
4. Report issues by file and priority

File Types

Type	Rules
Fortran	.claude/rules/fortran/conventions.md
Python	.claude/rules/python/conventions.md
RST	.claude/rules/docs/conventions.md
CHANGELOG	.claude/rules/changelog/format.md

Output Format

[code-style] Analysis

=== Fortran ===
  file.f95:L12: issue description

=== Python ===
  file.py:L23: issue description

=== Docstrings ===
  file.py:L23: Dxxx description (numpy-style rule)

Summary: N files, M issues

If a changed file still shows docstring findings after a good-faith fix, check whether the file is in [lint.per-file-ignores] in .ruff.toml — if so, either (a) finish cleaning it and remove its entry from that table, or (b) leave it alone (existing debt).

References

• .claude/rules/ - Full conventions (auto-loaded)
• references/quick-checks.md - Common issues
• For non-ASCII characters in sphinx-gallery .py tutorials (comment -> RST, print -> terminal, matplotlib -> LaTeX), use the audit-docs skill, which carries the context-specific replacement rules.