.gitignore

# ─── Generated pipeline outputs ──────────────────────────────────────────────
# All of these are regenerable from the source programs + pipeline scripts.



# Stage 1: extracted UTF-8 text per year
extracted/

# Stage 2: structured records (regenerated from extracted/)
# records.jsonl

# Stage 3: final QA report (regenerated each export)
qa_report.md

# Stage 4: diagnostic report
diagnostics.md

# Stage 5: review subset
review_needed.bib

# Stages 6/7: preview-only correction files (the canonical patches live in
# corrections.jsonl, which IS tracked)
corrections_from_diff.jsonl
corrections_from_splits.jsonl

# Working scratch files for the manual review-loop
reviewed_and_fixed.bib
reviewed_and_fixed_1.bib
reviewed_and_fixed*.bib

# Ad-hoc rollback backups
*.bak

# ─── Slide deck outputs ──────────────────────────────────────────────────────
# Generated by slides/analyze.py — derived from inhouse_conference.bib
slides/figures/01_*.png
slides/figures/02_*.png
slides/figures/03_*.png
slides/figures/04_*.png
slides/figures/05_*.png
slides/stats.json

# Generated by marp from slides.md
slides/slides.pdf

# ─── Older / duplicate source copies ────────────────────────────────────────
# Pre-2026-05 versions of programs, now superseded by the rescanned files in
# the project root.
_original full PDFs/

# ─── Tool / OS detritus ──────────────────────────────────────────────────────
__pycache__/
*.pyc
.DS_Store
.idea/
.vscode/
.claude/
*.swp