fix: disable Next.js caching on dynamic GET routes and fetches

chore: clean up unnecessary python scripts and notebooks

feat: add utility script to directly clean up team mentions

perf: skip extraction of already processed documents

feat: set data mention confidence threshold to 0.5

feat: rename corpus to UNHCR and remove login authorization note text

feat: format dropdown names and add light/dark theme toggle

feat: bypass document assignments to allow access to all PDFs

feat: update extraction script and colab notebook configuration

feat: add local gliner2 installation step to Colab notebook

feat: change clone repo to monitoring_of_datause in Colab notebook

feat: add Colab PDF extraction notebook and generator script

feat: re-run Helping Hands with post-processing and scale sample PDFs to 12 documents

fix: use whitespace-insensitive regex search when submitting annotations to support matching across line breaks

fix: copy public folder in production Dockerfile to prevent PDF renderer 404

Fix production crash: bundle corpora.json in Git and fetch annotations from HF Dataset

Remove ALLOWED_USERS and enable default document access for unconfigured users

Initial commit with Git LFS