Diagnosing errors

In the Streamer.bot directory there is a logs directory that has a daily log file. When submitting issues, please upload your log file as well, and let me know what time exactly the event occurred. Each transaction to GPT logs the following information:

Combined Context: This is the Context File, plus any found keyword data that we have passed along as context before the prompt API Key: This is the API key that was used for OpenAI Moderation Response: This is the raw JSON response from the Moderation API that lists out the flags and scores per category Flagged Categories: These are the categories the content was flagged for Response: This is the raw response we received back from ChatGPT

If we receive an invalid response back from GPT, the entire debug error will print instead of the expected data.

It only keeps the last 20 messages, and deletes older ones with each new message The messages are passed into the prompt automatically so that the bot has context of what is being said in chat