Chroma db Code changed thats why unable to access the vectorstore from ChromaDB for embeddings

[utkarshg1]

Checked other resources

• I added a very descriptive title to this issue.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

# Python code
from langchain_openai import OpenAIEmbeddings
from langchain_community.vectorstores import Chroma
vectorstore = Chroma.from_documents(documents=splits,
                                    embedding=OpenAIEmbeddings(model="text-embedding-3-large"), 
                                    persist_directory = "./chroma_db")

Error Message and Stack Trace (if applicable)

File /workspace/.pyenv_mirror/user/current/lib/python3.12/site-packages/chromadb/api/types.py:211, in validate_embedding_function(embedding_function)
208 protocol_signature = signature(EmbeddingFunction.call).parameters.keys()
210 if not function_signature == protocol_signature:
--> 211 raise ValueError(
212 f"Expected EmbeddingFunction.call to have the following signature: {protocol_signature}, got {function_signature}\n"
213 "Please see https://docs.trychroma.com/embeddings for details of the EmbeddingFunction interface.\n"
214 "Please note the recent change to the EmbeddingFunction interface: https://docs.trychroma.com/migration#migration-to-0416---november-7-2023 \n"
215 )

ValueError: Expected EmbeddingFunction.call to have the following signature: odict_keys(['self', 'input']), got odict_keys(['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

Description

I'm Trying to get Embeddings but looks like ChromaDB has changed functionality from their end.
Hence Langchain code also needs to be changed

https://docs.trychroma.com/embeddings/openai

System Info

langchain==0.1.14
langchain-community==0.0.30
langchain-core==0.1.37
langchain-openai==0.1.1
langchain-text-splitters==0.0.1
langchainhub==0.1.15
chromadb==0.4.24

[wanjianwei]

i also met this error, and is looking for a solution

[srnsrn120]

Same issue i am also facing .

[Thabet-Chaaouri]

Same issue here

[srnsrn120]

Same issue i am also facing .

If Any Work Around is there please help

[liugddx]

Let me see

[avoutsas67]

Same issue with me - it seems the version 0.4.16 of ChromaDB introduced a breaking change.

[liugddx]

I tried the code and it works for me, what's the problem?

[avoutsas67]

@liugddx see the top message in the thread by @utkarshg1. Describes the issue quite clearly.
Most likely you need to do a test on a clean venv where you have the versions mentioned above.

[liugddx]

If you want a quick fix I have a solution

openai_ef = embedding_functions.OpenAIEmbeddingFunction(model_name="text-embedding-3-large",api_key=os.environ["OPENAI_API_KEY"])
vectorstore = Chroma.from_documents(documents=all_splits,
                                    embedding=openai_ef,
                                    persist_directory = "./chroma_db")

And change line 279 of chroma.py to the following method

            embeddings = self._embedding_function(texts)

[utkarshg1]

Same issue i am also facing .

If Any Work Around is there please help

An easiest workaround for this is using the FAISS cpu as Vectorstore
Make sure you install faiss-cpu package as well

pip install faiss-cpu

Python code below:

# Index
from langchain_openai import OpenAIEmbeddings
from langchain_community.vectorstores import FAISS
vectorstore = FAISS.from_documents(documents=splits,
                                   embedding=OpenAIEmbeddings())

# Retriever
vectorstore.as_retriever()

[srnsrn120]

_embedding_function

No , Now i am using Multi vector retriever, i am not able to use Faiss or QdrantClient, only Chroma i am able to use
So thats why i wanted Chroma work around .

[baskaryan]

Should be fixed in langchain-community==0.0.31, please let me know if you're still seeing an issue