Re-embeds all chunks with the current embedding configuration.
Use this after switching embedding models or updating to a new version.
$ mix arcana.reembed_chunks
Options
--collection- Only re-embed chunks in the specified collection--batch-size- Number of chunks to process at once (default: 50)--concurrency N- Number of parallel embedding requests (default: 5)--skip N- Skip first N chunks (for resuming interrupted runs)--quiet- Suppress progress output
Examples
# Default usage (all collections)
mix arcana.reembed_chunks
# Re-embed only a specific collection
mix arcana.reembed_chunks --collection my-docs
# With larger batch size and higher concurrency
mix arcana.reembed_chunks --batch-size 100 --concurrency 10
# Resume from chunk 500
mix arcana.reembed_chunks --skip 500
# Quiet mode (no progress bar)
mix arcana.reembed_chunks --quiet