This is a demo script for an Airtable to Weaviate ChatVectorDBChain QA Agent with Chat History

AirtableToWeaviateChatVectorDBChain.py

#!/usr/local/bin/python

import logging
import weaviate
import langchain
from langchain.document_loaders import AirtableLoader
from langchain.vectorstores.weaviate import Weaviate
from langchain.llms import OpenAI
from langchain.chains import ChatVectorDBChain
from llama_index.langchain_helpers.agents import IndexToolConfig, LlamaIndexTool, create_llama_chat_agent
from llama_index.langchain_helpers.memory import GPTIndexMemory

# Set up logging
logging.basicConfig(filename='logfile.log', level=logging.DEBUG)
logger = logging.getLogger()

# Weaviate setup
WEAVIATE_URL = "http://0.0.0.0:8082"
# WEAVIATE_API_KEY = "your_weaviate_api_key"
client = weaviate.Client(url=WEAVIATE_URL, auth_client_secret=weaviate.AuthApiKey(WEAVIATE_API_KEY))

# Airtable loader
airtable_loader = AirtableLoader("base_id", "table_name", "api_key")

# LangChain setup
vectorstore = Weaviate(client, "YourClass", "content")
MyOpenAI = OpenAI(temperature=0.2, openai_api_key="your_openai_api_key")
qa = ChatVectorDBChain.from_llm(MyOpenAI, vectorstore)

# LLAMAIndex setup
tool_config = IndexToolConfig(query_engine=airtable_loader, name="Airtable Index", description="Index for Airtable data", tool_kwargs={"return_direct": True})
tool = LlamaIndexTool.from_tool_config(tool_config)
memory = GPTIndexMemory(index=airtable_loader, human_prefix="Human", ai_prefix="AI", memory_key="history", input_key="question", output_key="answer")

# Create agent
agent_chain = create_llama_chat_agent(tool, MyOpenAI, memory=memory, verbose=True)

# Run chat
chat_history = []
print("Welcome to the Weaviate ChatVectorDBChain Demo!")
print("Please enter a question or dialogue to get started!")
while True:
    query = input("")
    result = agent_chain.run(input=query)
    print(result["answer"])
    chat_history.append((query, result["answer"]))

    # Log memory to a local file
    logger.debug(f'Chat history: {chat_history}')