Joint Tabular/Semantic QA over Tesla 10K with Azure Chat Open AI

[Manikanta5112]

Hi,

I am trying to replicate the same code from https://docs.llamaindex.ai/en/latest/examples/query_engine/sec_tables/tesla_10q_table.html#joint-tabular-semantic-qa-over-tesla-10k.
But, using AzureChatOpenAI as below

from langchain.chat_models import AzureChatOpenAI
import os
from llama_index.llms.langchain import *
from dotenv import load_dotenv

load_dotenv()

model = AzureChatOpenAI(
temperature=0.2,
deployment_name=os.getenv("AZURE_OPENAI_DEPLOYMENT"),
verbose=True)

llm = LangChainLLM(
model
)

And, passing the llm to UnstructuredElementNodeParser

node_obj = UnstructuredElementNodeParser(llm=llm)

raw_nodes_2021 = node_obj.get_nodes_from_documents(docs_2021)

But, ending up with ValueError: Could not extract json string from output: The answer cannot be provided as there is no table mentioned in the context information

Any suggestions, How can I effectively user Azure chat open ai with this

[dosubot[bot]]

🤖

Hi,

The error you're encountering seems to be related to the output format of the AzureChatOpenAI model. The LlamaIndex expects the output to be a JSON string, and if it's not in this format, it could lead to the ValueError you're seeing.

To resolve this issue, you can modify the chat method of the OpenAIAgent to ensure the output is a JSON string. Here's an example:

import json

class OpenAIAgent:
    # ... other methods ...

    def chat(self, message):
        # ... code to process the message and call the function ...

        # Convert the output to a JSON string
        output_json = json.dumps(output)

        return output_json

This will ensure that the output of the function call is a JSON string, which should resolve the ValueError you're encountering.

Also, make sure that the environment variables required by the AzureOpenAI class (OPENAI_API_TYPE, OPENAI_API_VERSION, OPENAI_API_BASE, and OPENAI_API_KEY) are correctly set in your environment. The engine parameter should also be correctly set when initializing the AzureOpenAI class. This parameter corresponds to the custom name you chose for your deployment when you deployed a model on Azure OpenAI.

You can refer to the following resources for more information:

• Building a chatbot tutorial
• Chat engine context example
• AzureOpenAI customization example
• AzureOpenAI class source code

I hope this helps! Let me know if you have any other questions.

Sources

• azure openai compatibility
• docs/end_to_end_tutorials/chatbots/building_a_chatbot.md
• docs/examples/chat_engine/chat_engine_context.ipynb
• docs/examples/customization/llms/AzureOpenAI.ipynb
• llama_index/llms/azure_openai.py
• llama_index/chat_engine/types.py

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.