Summary by CodeRabbit

• New Features

  • Added support for semantic search using message embeddings, with configurable feature flags and token limits.
  • Introduced new API options to control semantic search behavior in peer and session endpoints.
  • Added batch embedding for efficient processing of multiple messages.
  • Created a dedicated table for storing message embeddings to enable fast vector similarity search.
  • Enhanced chat functionality with a new target parameter for specifying chat interaction targets.
  • Introduced a new script to generate embeddings for existing messages lacking embeddings.

• Configuration

  • Added new environment variables and settings for embedding token limits and feature enablement.

• Bug Fixes

  • Corrected parameter order in collection retrieval functions for consistency.

• Tests

  • Added comprehensive tests for message embedding creation, batch embedding, and semantic search feature flag handling.
  • Updated existing tests to use new structured search request format.

• Documentation

  • Improved environment variable documentation for embedding configuration.