Bug after the openai updated in Embedding #13051

SGA-Yuze-Li · 2023-11-08T07:56:36Z

dosubot · 2023-11-08T07:57:01Z

🤖

I'm Dosu, a bot designed to assist you with the LangChain repository. I'm here to help you navigate through bugs, answer your questions, and guide you if you'd like to contribute to the project. While we wait for a human maintainer, I'm here to provide initial assistance.

I'm looking at your issue right now and will provide a more comprehensive response shortly.

95 · 2023-11-08T07:59:18Z

I've had this exact issue as well. #13050
Hoping for a resolution

dosubot · 2023-11-08T08:09:00Z

🤖

The error you're encountering is due to a recent change in the EmbeddingFunction interface. The EmbeddingFunction.__call__ method now expects two arguments: self and input. However, it seems that in your code, the EmbeddingFunction.__call__ method is being called with self, args, and kwargs.

To resolve this issue, you need to update the way you're calling the EmbeddingFunction.__call__ method. Instead of passing args and kwargs, you should pass a single input argument.

Here's how you can do it:

class SagemakerEndpointEmbeddings(Embeddings):
    # ...

    def embed_query(self, query: str) -> List[float]:
        input = self.prepare_input(query)
        return self.embedding_function(input)

    def embed_documents(self, documents: List[str]) -> List[List[float]]:
        input = self.prepare_input(documents)
        return self.embedding_function(input)

In the prepare_input method, you should prepare the input argument in a way that is compatible with the new EmbeddingFunction.__call__ interface.

Also, you might need to adjust the predict_fn() function within the custom inference.py script to handle batched requests. Change the return line from return {"vectors": sentence_embeddings[0].tolist()} to return {"vectors": sentence_embeddings.tolist()}.

Please refer to the following sources for more information:

I hope this helps! If you have any further questions or if something is unclear, please let me know.

Sources

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.

thundo · 2023-11-08T09:24:54Z

Same error on Colab with openai pinning to 0.28.1

95 · 2023-11-08T09:27:50Z

Same error on Colab with openai pinning to 0.28.1

Yes. I pinned to 0.28.1 as well, didn't help.

sjoerd222888 · 2023-11-08T09:59:02Z

Problem seems to appear with chroma update to 0.4.16

Pinning to previous version seems to help.

chromadb==0.4.15

Langchain still implements the old interface. See: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023

pa1kumar9 · 2023-11-08T10:02:56Z

Problem seems to appear with chroma update to 0.4.16

Pinning to previous version seems to help.
chromadb==0.4.15
This worked.,, Thank you

95 · 2023-11-08T10:07:31Z

Problem seems to appear with chroma update to 0.4.16

Pinning to previous version seems to help.
chromadb==0.4.15
Langchain still implements the old interface. See: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023

Worked perfectly. Thank you

itsvaibhav01 · 2023-11-08T10:32:14Z

Problem seems to appear with chroma update to 0.4.16

Pinning to previous version seems to help.
chromadb==0.4.15
Langchain still implements the old interface. See: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023

Thanks a lot!!!

SGA-Yuze-Li · 2023-11-08T10:44:41Z

Problem seems to appear with chroma update to 0.4.16

Pinning to previous version seems to help.
chromadb==0.4.15
Langchain still implements the old interface. See: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023

Thanks a lot! Worked!

Kozato01 · 2023-11-08T14:14:12Z

O problema parece aparecer com a atualização do chroma para 0.4.16

Fixar na versão anterior parece ajudar.
chromadb==0.4.15
Langchain ainda implementa a interface antiga. Consulte: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023

My hero!

efriis · 2023-11-08T21:39:20Z

Langchain 0.0.332 released with the chroma team's fix for compatibility with chromadb>=0.4.16

Can now use latest of both pip install -U langchain chromadb

ScottXiao233 · 2023-11-09T08:39:08Z

Langchain 0.0.332 released with the chroma team's fix for compatibility with chromadb>=0.4.16

Can now use latest of both pip install -U langchain chromadb

I still met the problem, all the library are the latest version, here are my code:

from langchain.document_loaders import TextLoader
loader = TextLoader('/tanglg/lx/code/Qwen/forest_file.txt')
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
texts = text_splitter.split_documents(documents)
 
embeddings = OpenAIEmbeddings(
    openai_api_key="EMPTY",
    openai_api_base="http://localhost:8000/v1", 
)

forest_file_store = Chroma.from_documents(documents=texts, embedding=embeddings, collection_name="forest_file")

and here is the error info:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/tanglg/lx/code/Qwen/xly_vectordb_test.ipynb 单元格 3 line 1
      [5](vscode-notebook-cell://ssh-remote%2B172.30.254.229/tanglg/lx/code/Qwen/xly_vectordb_test.ipynb#W1sdnNjb2RlLXJlbW90ZQ%3D%3D?line=4) texts = text_splitter.split_documents(documents)
      [7](vscode-notebook-cell://ssh-remote%2B172.30.254.229/tanglg/lx/code/Qwen/xly_vectordb_test.ipynb#W1sdnNjb2RlLXJlbW90ZQ%3D%3D?line=6) embeddings = OpenAIEmbeddings(
      [8](vscode-notebook-cell://ssh-remote%2B172.30.254.229/tanglg/lx/code/Qwen/xly_vectordb_test.ipynb#W1sdnNjb2RlLXJlbW90ZQ%3D%3D?line=7)     openai_api_key="EMPTY",
      [9](vscode-notebook-cell://ssh-remote%2B172.30.254.229/tanglg/lx/code/Qwen/xly_vectordb_test.ipynb#W1sdnNjb2RlLXJlbW90ZQ%3D%3D?line=8)     openai_api_base="http://localhost:8000/v1", 
     [10](vscode-notebook-cell://ssh-remote%2B172.30.254.229/tanglg/lx/code/Qwen/xly_vectordb_test.ipynb#W1sdnNjb2RlLXJlbW90ZQ%3D%3D?line=9) )
---> [12](vscode-notebook-cell://ssh-remote%2B172.30.254.229/tanglg/lx/code/Qwen/xly_vectordb_test.ipynb#W1sdnNjb2RlLXJlbW90ZQ%3D%3D?line=11) forest_file_store = Chroma.from_documents(documents=texts, embedding=embeddings, collection_name="forest_file")

File /tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:684, in Chroma.from_documents(cls, documents, embedding, ids, collection_name, persist_directory, client_settings, client, collection_metadata, **kwargs)
    [682](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:682) texts = [doc.page_content for doc in documents]
    [683](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:683) metadatas = [doc.metadata for doc in documents]
--> [684](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:684) return cls.from_texts(
    [685](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:685)     texts=texts,
    [686](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:686)     embedding=embedding,
    [687](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:687)     metadatas=metadatas,
    [688](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:688)     ids=ids,
    [689](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:689)     collection_name=collection_name,
    [690](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:690)     persist_directory=persist_directory,
    [691](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:691)     client_settings=client_settings,
    [692](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:692)     client=client,
    [693](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:693)     collection_metadata=collection_metadata,
    [694](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:694)     **kwargs,
    [695](https://vscode-remote+ssh-002dremote-002b172-002e30-002e254-002e229.vscode-resource.vscode-cdn.net/tanglg/commonsoft/miniconda3/envs/qwen/lib/python3.10/site-packages/langchain/vectorstores/chroma.py:695) )
...

ValueError: Expected EmbeddingFunction.__call__ to have the following signature: odict_keys(['self', 'input']), got odict_keys(['self', 'args', 'kwargs'])
Please see https://docs.trychroma.com/embeddings for details of the EmbeddingFunction interface.
Please note the recent change to the EmbeddingFunction interface: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023

TimCoates · 2023-11-09T13:14:11Z

I'm getting this bug with:
langchain V 0.0.332
chromadb V 0.4.16

persist_directory='./db'
from langchain.embeddings import GPT4AllEmbeddings
embedding=GPT4AllEmbeddings()
from langchain.vectorstores import Chroma
vectordb = Chroma.from_documents(documents=texts,
                                 embedding=embedding,
                                 persist_directory=persist_directory)

efriis · 2023-11-09T18:32:31Z

@ScottXiao233 - I tried reproducing, but your code works for me. If you're using a custom openai embeddings endpoint, maybe make sure it works with the new sdk alone?

@TimCoates - This also might be something with your local embeddings setup. Maybe submit as a new issue specific to GPT4AllEmbeddings? It's unclear to me why you'd see this using something other than OpenAI

afedotov-align · 2023-11-14T15:23:49Z

@efriis , also doesn't work for me with
langchain==0.0.335
chromadb==0.4.17

code example

embeddings = AzureOpenAIEmbeddings(
    model="text-embedding-ada-002",
    deployment=os.getenv("EMBEDDING_DEPLOYMENT_NAME"),
    azure_endpoint=os.getenv("AZURE_ENDPOINT"),
    api_key=os.getenv("AZURE_OPENAI_KEY"),
    openai_api_version=os.getenv("AZURE_API_VERSION"),
)

persistent_client = chromadb.PersistentClient("./test")

collection = persistent_client.get_or_create_collection(
    name="test", embedding_function=embeddings.embed_documents
)

efriis · 2023-11-14T17:35:11Z

@afedotov-align - looks like you might be mixing LangChain langchain.vectorstores.Chroma with directly accessing chromadb. Could you try again using the LangChain vectorstore with the LangChain embedding function? https://python.langchain.com/docs/integrations/vectorstores/chroma#basic-example

TimCoates · 2023-11-14T20:03:14Z

@ScottXiao233 - I tried reproducing, but your code works for me. If you're using a custom openai embeddings endpoint, maybe make sure it works with the new sdk alone?

@TimCoates - This also might be something with your local embeddings setup. Maybe submit as a new issue specific to GPT4AllEmbeddings? It's unclear to me why you'd see this using something other than OpenAI

Thanks, rather than using .from_documents, doing as so:
vectorstore = Chroma("langchain_store", embedding_function=embeddings, persist_directory="./db")
Then:
Chroma.add_documents(vectorstore. texts)
worked for me

shreyahegde18 · 2023-11-21T07:33:51Z

Problem seems to appear with chroma update to 0.4.16

Pinning to previous version seems to help.
chromadb==0.4.15
Langchain still implements the old interface. See: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023

Thank you :')

deeeed · 2023-12-04T14:30:25Z

issue should still be open

amanlai · 2024-02-25T22:47:30Z

@varayush007 Try defining a custom class that sub-classes OpenAIEmbeddings and use it as the embedding function. This discussion may help.

BlackSiao · 2024-03-01T09:47:30Z

chromadb==0.4.15
pip install chromaab=0.4.15 really solve my promblem, thank you

kaisersoser · 2024-04-02T12:17:33Z

Backdating my chromadb version to chromadb==0.4.15 worked for me.
Just run the command: pip install --upgrade chromadb==0.4.15

dosubot bot added Ɑ: embeddings Related to text embedding models module 🤖:bug Related to a bug, vulnerability, unexpected error with an existing feature 🔌: chroma Primarily related to ChromaDB integrations labels Nov 8, 2023

dennisdegreef added a commit to dennisdegreef/restai that referenced this issue Nov 8, 2023

Fix chromadb also pin. See langchain-ai/langchain#13051 (comment)

b5faeb5

This was referenced Nov 8, 2023

OpenAIEmbeddings() does not work because of these bugs #12943

Closed

Don't pass EF to chroma #13085

Merged

efriis closed this as completed Nov 8, 2023

dosubot bot mentioned this issue Nov 10, 2023

langchain & chroma - Basic Example #13191

Closed

dosubot bot mentioned this issue Nov 16, 2023

langchain.vectorstores.Chroma support for EmbeddingFunction.__call__ update of ChromaDB #13441

Closed

14 tasks

dosubot bot mentioned this issue Feb 9, 2024

unable to use embed_documents function for ChromaDB #17295

Closed

varayush007 mentioned this issue Feb 25, 2024

Chromadb error #18098

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Bug after the openai updated in Embedding #13051

Bug after the openai updated in Embedding #13051

SGA-Yuze-Li commented Nov 8, 2023

dosubot bot commented Nov 8, 2023

95 commented Nov 8, 2023

dosubot bot commented Nov 8, 2023

thundo commented Nov 8, 2023

95 commented Nov 8, 2023

sjoerd222888 commented Nov 8, 2023

pa1kumar9 commented Nov 8, 2023

95 commented Nov 8, 2023

itsvaibhav01 commented Nov 8, 2023

SGA-Yuze-Li commented Nov 8, 2023

Kozato01 commented Nov 8, 2023

efriis commented Nov 8, 2023

ScottXiao233 commented Nov 9, 2023

TimCoates commented Nov 9, 2023

efriis commented Nov 9, 2023

afedotov-align commented Nov 14, 2023

efriis commented Nov 14, 2023

TimCoates commented Nov 14, 2023

shreyahegde18 commented Nov 21, 2023

deeeed commented Dec 4, 2023

amanlai commented Feb 25, 2024

BlackSiao commented Mar 1, 2024

kaisersoser commented Apr 2, 2024

Bug after the openai updated in Embedding #13051

Bug after the openai updated in Embedding #13051

Comments

SGA-Yuze-Li commented Nov 8, 2023

System Info

Who can help?

Information

Related Components

Reproduction

Expected behavior

dosubot bot commented Nov 8, 2023

95 commented Nov 8, 2023

dosubot bot commented Nov 8, 2023

Sources

thundo commented Nov 8, 2023

95 commented Nov 8, 2023

sjoerd222888 commented Nov 8, 2023

pa1kumar9 commented Nov 8, 2023

95 commented Nov 8, 2023

itsvaibhav01 commented Nov 8, 2023

SGA-Yuze-Li commented Nov 8, 2023

Kozato01 commented Nov 8, 2023

efriis commented Nov 8, 2023

ScottXiao233 commented Nov 9, 2023

TimCoates commented Nov 9, 2023

efriis commented Nov 9, 2023

afedotov-align commented Nov 14, 2023

efriis commented Nov 14, 2023

TimCoates commented Nov 14, 2023

shreyahegde18 commented Nov 21, 2023

deeeed commented Dec 4, 2023

amanlai commented Feb 25, 2024

BlackSiao commented Mar 1, 2024

kaisersoser commented Apr 2, 2024