Trying to fix this:

At a glance

The community member is trying to fix an issue with their code, which is raising an AttributeError related to the composable_graph.search() function. The comments suggest that the composable_graph is deprecated, and the community member should use composable_graph.as_query_engine().query() instead.

The community member is building a general Q&A tool that ingests data from various sources like Salesforce, Slack, and GraphQL, using a Chrome extension called Bloom. They are new to the "RAG" (Retrieval Augmented Generation) concept and are having trouble with the documentation issues.

One of the community members suggests that the community member should use an agent for their use case.

Useful resources

LLong Drink

Trying to fix this:

Plain Text

@app.post("/query")
def query(query: str = Body(..., embed=True)):
    response = composable_graph.search(query)
    return {"response": response.response}

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)

Getting this:

Plain Text

        if not self.create and original is DEFAULT:
>           raise AttributeError(
                "%s does not have the attribute %r" % (target, name)
            )
E           AttributeError: <llama_index.core.indices.composability.graph.ComposableGraph object at 0x7fce4d9bbd60> does not have the attribute 'search'

/usr/local/python/3.10.13/lib/python3.10/unittest/mock.py:1420: AttributeError
================================ warnings summary ================================
tests.py::test_ingest
  /home/codespace/.local/lib/python3.10/site-packages/httpx/_client.py:680: DeprecationWarning: The 'app' shortcut is now deprecated. Use the explicit style 'transport=WSGITransport(app=...)' instead.
    warnings.warn(message, DeprecationWarning)

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
============================ short test summary info =============================
FAILED tests.py::test_query - AttributeError: <llama_index.core.indices.composability.graph.ComposableGraph...
===================== 1 failed, 1 passed, 1 warning in 3.51s =====================

10 comments

LLogan M

composable graph is low-key deprecated lol (so docs got phased out)

It seems like you are calling composable_graph.search() which is not a function

Normally you'd do composable_graph.as_query_engine().query()

LLong Drink

Do you have any suggestions on what to use instead?

LLong Drink

I I'd rather use the up to date tooling haha

LLogan M

Depends I suppose. What are you using the graph for now?

LLong Drink

Welp. Uhhh because Mistral told me to :3.

LLong Drink

But

LLong Drink

The goal is to make a general Q&A tool on data from tools from LlamaHub such as Salesforce, Slack, and GraphQL. What I've got is a chrome extension called Bloom that I'm trying to set up to allow people to ingest data from any source, and ask questions for GTM research.

LLong Drink

https://bloom.getoasis.io

LLong Drink

Disclosure: I'm new to the whole RAG thing, but I have autistic pattern recog so I've been trying to get by with just putting things together haha. But the docs issue has been making it tough 😭 .

LLogan M

Sounds like you should use an agent then? 👀

Add a reply

Find answers from the community

Trying to fix this: