.gitignore

# --- Bring-your-own corpora: NEVER commit licensed data (e.g. Web of Science) ---
# The repo ships no .bib data. Drop your own into assetsbib/ (see assetsbib/README.md).
assetsbib/*.bib
assetsbib/*.txt
# corpus cache RDS that load_corpus() writes next to the .bib (contains the data!)
assetsbib/*.rds
*.corpus.rds

# --- Generated outputs: NOT shipped (they would derive from licensed WoS data) ---
runs/
ui_metadata/

# --- Local, corpus-specific regression baseline (bootstrapped per corpus, not shipped) ---
tests/fixtures/snapshots_*.json

# --- Personal / internal Spanish design & retrospective notes (kept local, not published) ---
docs/CONCLUSION.md
docs/EXECUTION_PLAN.md
docs/EXHAUSTIVE_ANALYSIS_PLAN.md
docs/UI_METADATA_METHOD.md
docs/PARITY_CHECK.md
docs/ACID_TEST.md

# --- renv: commit the lockfile, not the library ---
renv/library/
renv/local/
renv/cellar/
renv/staging/
renv/python/
renv/sandbox/

# --- R / RStudio artifacts ---
.Rproj.user/
.Rhistory
.Rapp.history
.RData
.Ruserdata
*.Rproj
*.Rcheck/
*.tar.gz
.httr-oauth

# --- metadata extractor scratch output ---
metadata_extractor/out/

# --- OS / editors ---
.DS_Store
Thumbs.db
*~
.vscode/
.idea/