Chatbot Issues with Postgres Chat Memory and Supabase Vector Store
Problem 1: Empty Chat Memory Causing Errors
Hi everyone,
I'm encountering an issue with my chatbot, which receives webhook input from Voiceflow. The problem arises when I integrate Postgres Chat Memory to store the dialogue history and context. It appears the chat model encounters errors when it processes an empty segment within the chat memory, leading to a [GoogleGenerativeAI Error]: 400 Bad Request - contents.parts must not be empty
.
I found a relevant discussion on the callin.io community forum: GoogleGenerativeAI Error: 400 Bad Request. The thread includes a screenshot that seems to offer a solution, but I'm unsure how to implement it.
Problem 2: Chat Memory Overriding Supabase Vector Store
When the Chat Memory is active and functioning, it seems to take precedence as the knowledge source, overriding the Supabase Vector Store. This is not the desired behavior. My prompts are configured to use chat memory solely for tracking past interactions and responding to repeated questions. However, I haven't been able to achieve this specific functionality.
Information on your callin.io setup
- callin.io version: 1.1
- Database (default: SQLite): SQLite
- callin.io EXECUTIONS_PROCESS setting (default: own, main): own
- Running callin.io via (Docker, npm, callin.io cloud, desktop app): Docker Self Hosted
- Operating system: Windows 11
I also require chat memory functionality to effectively track user intents, especially for responses like a simple "Yes".
Hi there.
For the second problem, you need to clearly define your vector tool. Instead of naming it “Supabase Vector Store2”, you could name it Vector DB Tool
or any other name you prefer. Following that, you can add more detail to your instructions for the Vector DB Tool
, for instance: You are a helpful assistant. You have the capability to search for anything related to [topic] using the "Vector DB Tool", which will help you get more precise answers.
If this is helpful, please mark my message as the answer. Thanks:)