How to extract the Vector from an OpenAI Embeddings Call?

February 5, 2024

I use nearly the same code as here in this Git Repo to get Embeddings from OpenAI:
https://gist.github.com/limcheekin/997de2ae0757cd46db796f162c3dd58c

oai = OpenAI(
# This is the default and can be omitted
api_key="sk-.....",
)

def get_embedding(text_to_embed, openai):
   
    response = openai.embeddings.create(
        model= "text-embedding-ada-002",
        input=[text_to_embed]
    )
    
    return response

embedding_raw = get_embedding(text,oai)

According to the Git Repo the Vector should be in response['data'][0]['embedding']. But it isn’t in my case.

When I print the response Variable, I got this:

print(embedding_raw)

Output:

CreateEmbeddingResponse(data=[Embedding(embedding=[0.009792150929570198, -0.01779201813042164, 0.011846082285046577, -0.0036859565880149603, -0.0013213189085945487, 0.00037509595858864486,..... -0.0121011883020401, -0.015751168131828308], index=0, object='embedding')], model='text-embedding-ada-002', object='list', usage=Usage(prompt_tokens=360, total_tokens=360))

Sorry, I’m new to python, but how can I access the vector data?

>Solution :

Simply return just the embedding vector as follows:

def get_embedding(text_to_embed, openai):
   
    response = openai.embeddings.create(
        model= "text-embedding-ada-002",
        input=[text_to_embed]
    )
    
    return response.data[0].embedding # Change this

embedding_raw = get_embedding(text,oai)