Integrate database

[makkos-lilly]

Completed integrating the database with the OSM dashboard.

• Added a python overlay file (overlay.py) that builds a project-local overlay in data.db_overlay from the database. This overlay file filters to users in data/metadata/group_users.csv with optional extra filters. It writes changesets_subsets.csv, changesets_tags_subset.csv, contributions_summary_subset.csv, user_hashtag_pairs.csv and _overlay_log.txt to data/db_overlay. The file contains robust path resolution so the overlay can be called from the correct folder. 
• Added a new data bootstrap to the dashboard file that detects the data root and then reads either the database overlay (from data/db_overlay when use_db_overlay:true) or the raw/API path (from data/raw when use_db_overlay:false). It also creates robust bounding box validation for the changesets. 
• Hashtags and all user changeset information now come from the database. Changesets_tags_subset.csv is used to populate information about user hashtags. The word cloud is capped at the top 50 tokens. 
• In the dashboard, three extra boxes are added for more fine-grained information on user comments and hashtags. Previously, we just had the comment bigrams. These are: top words, per-user TF-IDF view, raw comments capped at 20 most recent. 
• The feature keys panel is only available when use_db_overlay:false as keys such as highway, building, etc are only available from the planet file, which is accessed via APIs. When use_db_overlay:true, these panels are not available.
• README file is updated for information on how to use the dashboard with the database (using use_db_overlay:true when rendering the dashboard with quarto). 
• To use the dashboard with the database the database must be named osm_changesets.duckdb and must be placed in the directory named database.

To run the dashboard using the database, when rendering the dashboard write "quarto render dashboard.qmd -P use_db_overlay: true"