ai4data/data-use-annotation

feat: add in-app help guide with ❓ button in top bar

docs: add annotator guide for the Data Use Annotation Tool

docs: update title and emoji in README

ui: inline Document label with dropdown to save space

feat: show all pages with jump-to-mention navigation

fix: extend highlighting to headings, strong, em, blockquote

ui: neutralize markdown preview colors for annotation clarity

ui: 5 polish fixes

fix: don't fetch docs until annotatorName is set

fix: block document access for users not in annotator_config

fix: add corpora.json with UNHCR entry to git

feat: add UNHCR corpus to registry and prepare_unhcr.py script

feat: multi-corpus support

security: add auth middleware to block unauthenticated write API calls

fix: persist source field in saveAnnotation so leaderboard counts Added

fix: add source='human' to manual annotation payload for leaderboard tracking

fix: disable all caching in leaderboard API for real-time updates

feat: add refresh button to leaderboard + cache-bust API calls

docs: add comprehensive ANNOTATION_GUIDE.md

feat: per-annotator document assignments with configurable overlap

feat: add annotation leaderboard with medals and stats

feat: per-annotator validation support for multi-user overlap

fix: remove PDF HEAD check that was filtering out valid docs (WB returns 301 for HEAD)

fix: validation updates now saved at top level, not nested in dataset_name

feat: warn on unverified mentions when clicking Next, move progress to top-right

refactor: make progress bar contextual - shows current PDF/page/mentions position

feat: add progress tracking with docs/pages/mentions verification stats

fix: clamp restored pageIdx to valid range on doc change

feat: persist doc/page selection across refreshes via sessionStorage

feat: add language detection to prepare_data.py, exclude non-English docs

fix: proxy PDFs through /api/pdf-proxy to bypass CORS

feat: prepare_data.py script + use has_revalidation filter in documents API

fix: validate PDF URLs before listing documents, skip broken/unreachable PDFs

feat: add loading spinner and error fallback to PDF viewer

feat: on-demand document loading, no limit, PDF loading indicator

fix: open OAuth login in new tab to bypass iframe restrictions, auto-poll for cookie

fix: remove OAuth state verification (cookies don't survive HF iframe redirect)

security: gate app behind HF OAuth login

fix: use _rawIndex for delete/validate to avoid index mismatch after filtering

update: reverted local changes

fix: remove OAuth state verification (cookies don't survive HF iframe redirect)

security: gate app behind HF OAuth login

feat: per-annotator document assignment

fix: rename 'Your Name' to 'Annotator', make read-only, auto-populated from OAuth

feat: add ALLOWED_USERS allowlist for OAuth login

feat: HF OAuth login, increase docs to 50, filter consensus non-datasets

filter: hide consensus non-datasets (model + judge agree)

fix: strip markdown from rawText before indexOf for annotation position

revert: remove isPullRequest, use write-access token instead

fix: add isPullRequest to all HF commits (403 Forbidden)