Embeddings.create() got an unexpected keyword argument 'engine', when using ChromaDB

[pedrurrego]

Context ℹ️ 💁🏼‍♂️

I have a VectorDB with Chroma using OpenAI's Embeddings and created a retriever_model in dspy:

from dspy.retrieve.chromadb_rm import ChromadbRM
retriever_model = ChromadbRM(collection_name=COLLECTION_NAME, persist_directory=MY_PATH)

The following are my dependencies in the requirements.txt file. I am using Python 3.10.12.

chromadb==0.4.22
dspy-ai==2.1.9
openai==1.12.0
tiktoken==0.5.1

Problem 💩

When running

dspy.settings.rm("Who is Chuck Norris?")

I get the following error:

File [~/Library/CloudStorage/OneDrive-Personal/Code/AI/LLM/dspy/venv/lib/python3.10/site-packages/dspy/retrieve/chromadb_rm.py:139](139), in ChromadbRM._get_embeddings(self, queries)
    [125] """Return query vector after creating embedding using OpenAI
    [126]
    [127] Args:
   (...)
    [131]     List[List[float]]: List of embeddings corresponding to each query.
    [132] """
    [134] model_arg = {"engine": self.model_name,
    [135]     "deployment_id": self.model_name,
    [136]     "api_version": self.api_version,
    [137]     "api_base": self.api_base,
    [138] }
--> [139] embedding = self.openai_ef._client.create(
    [140]     input=queries, model=self._openai_embed_model, **model_arg, api_provider=self.openai_api_type
    [141] )
    [143] return [embedding.embedding for embedding in embedding.data]

TypeError: Embeddings.create() got an unexpected keyword argument 'engine'

Possible solution 👍🏼 ✅

I was able to solve it by using only input and model arguments in self.openai_ef._client.create() method, as shown below:

embedding = self.openai_ef._client.create(input=queries, model=self._openai_embed_model)

What do you guys think about this? Thanks in advance for your help!

[okhat]

Ok solved(?) by merging #400. Let us know if all is good now. Thank you @lukaskuhn-lku and @animtel

[okhat]

Will be pushed to pypi on the next release