embedding error - Githubissues

langgenius / dify

Dify is an open-source LLM app development platform. Dify's intuitive interface combines AI workflow, RAG pipeline, agent capabilities, model management, observability features and more, letting you quickly go from prototype to production.

Other

50.36k stars 7.22k forks source link

worker-1 | Traceback (most recent call last): worker-1 | File "/app/api/core/indexing_runner.py", line 77, in run worker-1 | self._load( worker-1 | File "/app/api/core/indexing_runner.py", line 694, in _load worker-1 | tokens += future.result() worker-1 | File "/usr/local/lib/python3.10/concurrent/futures/_base.py", line 451, in result worker-1 | return self.__get_result() worker-1 | File "/usr/local/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result worker-1 | raise self._exception worker-1 | File "/usr/local/lib/python3.10/concurrent/futures/thread.py", line 58, in run worker-1 | result = self.fn(*self.args, **self.kwargs) worker-1 | File "/app/api/core/indexing_runner.py", line 749, in _process_chunk worker-1 | index_processor.load(dataset, chunk_documents, with_keywords=False) worker-1 | File "/app/api/core/rag/index_processor/processor/paragraph_index_processor.py", line 60, in load worker-1 | vector.create(documents) worker-1 | File "/app/api/core/rag/datasource/vdb/vector_factory.py", line 100, in create worker-1 | embeddings = self._embeddings.embed_documents([document.page_content for document in texts]) worker-1 | File "/app/api/core/embedding/cached_embedding.py", line 82, in embed_documents worker-1 | raise ex worker-1 | File "/app/api/core/embedding/cached_embedding.py", line 51, in embed_documents worker-1 | embedding_result = self._model_instance.invoke_text_embedding( worker-1 | File "/app/api/core/model_manager.py", line 172, in invoke_text_embedding worker-1 | return self._round_robin_invoke( worker-1 | File "/app/api/core/model_manager.py", line 303, in _round_robin_invoke worker-1 | return function(*args, **kwargs) worker-1 | File "/app/api/core/model_runtime/model_providers/__base/text_embedding_model.py", line 38, in invoke worker-1 | raise self._transform_invoke_error(e) worker-1 | core.model_runtime.errors.invoke.InvokeError: [tongyi] Error: 'NoneType' object is not subscriptable

@imaxwen Hello there! I'm here to help you with any bugs, questions, or contributions. Let me know how I can assist you while we wait for a human maintainer.

It looks like you're encountering a 'NoneType' object is not subscriptable error when using TONGYI's text-embedding-v2 in Dify version 0.7.2. This issue typically occurs when the response from the dashscope.TextEmbedding.call method does not contain the expected data.

To resolve this, you need to ensure that the response.output and response.usage are not None before attempting to subscript them. Here is a modified version of the embed_documents method with added checks to handle this scenario:

@staticmethod
def embed_documents(
    credentials_kwargs: dict, model: str, texts: list[str]
) -> tuple[list[list[float]], int]:
    """Call out to Tongyi's embedding endpoint.

    Args:
        credentials_kwargs: The credentials to use for the call.
        model: The model to use for embedding.
        texts: The list of texts to embed.

    Returns:
        List of embeddings, one for each text, and tokens usage.
    """
    embeddings = []
    embedding_used_tokens = 0
    for text in texts:
        response = dashscope.TextEmbedding.call(
            api_key=credentials_kwargs["dashscope_api_key"],
            model=model,
            input=text,
            text_type="document",
        )
        if response.output and "embeddings" in response.output and response.output["embeddings"]:
            data = response.output["embeddings"][0]
            if "embedding" in data:
                embeddings.append(data["embedding"])
            else:
                raise ValueError("Embedding data is missing in the response.")
        else:
            raise ValueError("Response output is missing or does not contain embeddings.")

        if response.usage and "total_tokens" in response.usage:
            embedding_used_tokens += response.usage["total_tokens"]
        else:
            raise ValueError("Response usage is missing or does not contain total tokens.")

    return [list(map(float, e)) for e in embeddings], embedding_used_tokens

This code adds checks to ensure that response.output and response.usage are not None and contain the expected keys before attempting to access them. If the expected data is missing, it raises a ValueError with an appropriate message ^[1]^[2].

To continue talking to Dosu, mention @dosu.

langgenius / dify

embedding error #7701

Self Checks

Dify version

Cloud or Self Hosted

Steps to reproduce

✔️ Expected Behavior

❌ Actual Behavior