Tex LaTeX Structure Parser

Overview

Use this skill for read-only structure extraction before any fixing step. Build a factual map of the paper, then hand that evidence to other skills.

Executable Entry Points

• •scripts/parse_latex_structure.py: extracts sections, citations, bibliography keys, figures, tables, references, and cross-check issues as JSON.

Example command:

python scripts/parse_latex_structure.py --project-root . --main-tex main.tex --pretty

Project Adaptation

1. •Identify TeX entry file(s) and inclusion style (\\input, \\include, generated files).
2. •Identify bibliography sources (.bib, inline bibliography, or mixed).
3. •Identify directories to ignore (build, cache, backups, checkpoints).

Workflow

1. •Run scripts/parse_latex_structure.py against the target project root.
2. •Extract core entities:
   • •section hierarchy
   • •citation keys used in TeX
   • •bibliography keys defined
   • •figure/table labels and captions
   • •figure/table references (\\ref, \\cref, equivalents)
3. •Run cross-checks from the JSON result:
   • •undefined citations
   • •uncited bibliography entries
   • •unreferenced figures/tables
   • •missing labels/captions
4. •Apply project-specific regex checks (TODO markers, style constraints, forbidden patterns).
5. •Produce a normalized issue list with file and line context.

Inputs and Outputs

• •Input: project root, TeX files, bibliography files, optional regex rules.
• •Output: structured facts plus actionable consistency findings.

Common Mistakes

• •Running fixes before building a full structure map.
• •Ignoring recursively included TeX files.
• •Treating regex matches as final truth without reference cross-checks.
• •Mixing parse evidence with subjective style feedback.