ragas evaluate with llama_index and langchain doesn't seem to work with Azure OpenAI

[Data-drone]

Azure OpenAI requires the special parameter deployment or deployment_id

The langchain wrappers seem to mostly have been updated to accomodate for this but it doesn't seem to work with ragas.

I ended up getting faithfulness working but updating the generate method with:

from

    elif isinstance(llm, BaseChatModel):
        ps = [p.format_messages() for p in prompts]
        result = llm.generate(ps, callbacks=callbacks)
        
        

to

    elif isinstance(llm, BaseChatModel):
        ps = [p.format_messages() for p in prompts]
        result = llm.generate(ps, callbacks=callbacks, deployment_id='<my_id>', api_version='<my_version>')

but with answer_relevancy I hit the same issue when it tries to run:

│    91 │   def calculate_similarity(                                                              │
│    92 │   │   self: t.Self, question: str, generated_questions: list[str]                        │
│    93 │   ):                                                                                     │
│ ❱  94 │   │   question_vec = np.asarray(self.embedding.embed_query(question)).reshape(1, -1)     │
│    95 │   │   gen_question_vec = np.asarray(                                                     │
│    96 │   │   │   self.embedding.embed_documents(generated_questions)                            │
│    97 │   │   )

Any ideas?

[shahules786]

Hi @Data-drone , It's possible to use AzureOpenAI with ragas. You have to init a new AzureChatOpenAI LLM as mentioned in this notebook

To know how to init AzureOpenAI using Langchain you can refer this guide

[Data-drone]

I have setup AzureOpenAI and Llama_index is working fine. But it seems like when you call the service context from ragas it is losing the Azure config somehow?

[jjmachan]

We did have a running thread where we fixed this one time but I would love to get on a call with you when you are available and help debug this. The issue is that we haven't gotten access to Azure Openai endpoints yet so can't debug from our end

the original thread: https://discord.com/channels/1119637219561451644/1123849273562386453/1148224350915407962

it would be great if @Data-drone could commend in that thread, We'll debug this in discord?

[Data-drone]

Sure should I try the latest master?

[jjmachan]

you could try that but we've released a 0.0.13 version too. Do try it out and let us know.
you might have to configure the AzureOpenAI separately for ragas metrics. You can use this guide to see how.