Changes

• Run SQLite ANALYZE at Docker build time to bake query planner statistics into the image, improving cold-start query performance

Changes

• Lower gzip compression level from 9 to 4 for ~10x faster response times with negligible size increase (29% vs 28% ratio)

Changes

• Fix test for root endpoint to match new index route
• Fix all ruff lint issues across the project
• Disable auto-publish workflow (now manual-only)
• Fix mypy type ignore comments for starlette imports

What's Changed

• Startup warmup: Pre-warm DB connection and forms query cache during app startup, so the first user request skips the cold-query cost
• Forms endpoint logging: Added print-based logging to identify server-side bottlenecks (request received, query, JSON serialization, done)
• Increased _get_forms LRU cache from 1 to 10 entries

What's New

• Sense frequency counts on words endpoint: The /lexicons/{lexicon}/words endpoint now includes SemCor corpus frequency counts (count attribute) in each included synset, enabling clients to rank word senses by usage frequency.
• Merged upstream changes from goodmami/wn
• Added Wikidata lexemes extension for function words
• Docker and web configuration improvements

Performance Improvements

• 55% faster forms query - Added covering index on (lexicon_rowid, form) which eliminates temporary B-tree creation and allows queries to be satisfied entirely from the index
• Expected production improvement: ~3s → ~1.3s for forms endpoint
• Better scaling with multiple lexicons

Technical Details

• Added form_lexicon_form_covering_index to the forms table
• Query plan now uses covering index instead of creating temporary B-trees for DISTINCT operations
• Compatible with existing databases (schema hash compatibility maintained)

Migration

For existing databases, the index will be created automatically when downloading new lexicons. To add the index to existing databases immediately, run:

CREATE INDEX form_lexicon_form_covering_index ON forms (lexicon_rowid, form);

Performance Improvements

• Optimized forms query: Added database index (form_lexicon_index) to significantly improve performance of forms endpoint
• More efficient SQL: Updated query to use SQL DISTINCT and direct joins for faster results

Changes

• Added form_lexicon_index on forms table for better query performance
• Updated compatible schema hashes to support new database schema
• Optimized _get_forms function to leverage new index

🤖 Generated with Claude Code

Full Changelog: v0.1.0...v0.2.0

Initial release of docker image for wn as a web server