profQUOTES parser#

profquotes collects professor quotes from a Discord channel and stores them in a local SQLite database.

Messages are fetched directly from the Discord API. After an initial extraction, use Helix to clean up the parsed results before they are saved.

Messages are expected to be in the following format, "irrelevant" messages are skipped.

COURSE CODE: Professor Name
> [quote]

Config is all env vars:

• DISCORD_TOKEN, DISCORD_CHANNEL_ID — required for --fetch
• LLM_BASE_URL, LLM_API_KEY, LLM_MODEL — required for processing, if using LLM processing.
• DB_PATH — defaults to profquotes.db
• CHUNK_SIZE — messages per LLM batch, defaults to 30
• REVIEW_EDITOR or EDITOR — must point to hx or helix