This pull request introduces two main sets of changes: enhancements to the LLM (Large Language Model) records for improved searchability, and a significant cleanup of the prospects table by removing unused columns.

[goldlabelapps]

Additionally, the application version is incremented to reflect these updates.

LLM Table Enhancements:

• Added a new search_vector column (type tsvector) to the llm table to support full-text search, along with a GIN index for faster querying. Migration scripts and a Python runner are included. [1] [2]
• Updated LLM record insertion logic to automatically generate the search_vector from the prompt and completion fields.
• Modified API endpoints and queries to include the search_vector field in LLM record retrievals, ensuring it is returned in API responses where appropriate. [1] [2] [3] [4] [5]

Prospects Table Cleanup:

• Dropped multiple unused columns from the prospects table, including: cleaned, do_not_call, email_confidence, home_phone, primary_email_catchall_status, primary_intent_score, primary_intent_topic, qualify_contact, secondary_intent_score, secondary_intent_topic, tertiary_email, tertiary_email_source, tertiary_email_status, tertiary_email_verification_source, and work_direct_phone. Each column removal is handled by its own migration script. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

Other:

• Bumped the application version from 2.1.8 to 2.1.9.