loomlike commented 4 months ago

Checked other resources

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

Example Code

Environment variables:

# for langchain
OPENAI_API_VERSION='...'
AZURE_OPENAI_API_KEY='my-key'
AZURE_OPENAI_ENDPOINT='my-azure-openai-endpoint'

# for mlflow -- this conflicts
OPENAI_API_BASE='my-azure-openai-endpoint'

Code to cause the error:

from langchain_openai import AzureOpenAIEmbeddings, AzureChatOpenAI

model = AzureChatOpenAI(
    azure_deployment="my-deployment-name",
)

# or 
model = AzureOpenAIEmbeddings(
    azure_deployment="my-deployment-name",
)

Error Message and Stack Trace (if applicable)

ValidationError: 1 validation error for AzureChatOpenAI
__root__
  As of openai>=1.0.0, Azure endpoints should be specified via the `azure_endpoint` param not `openai_api_base` (or alias `base_url`). (type=value_error)

Description

AzureOpenAIEmbeddings and AzureChatOpenAI automatically pick up the env vars and throwing errors complaining about having both azure_endpoint param and openai_api_base.

OPENAI_API_BASE env var is being used in other packages, like MLFlow.

System Info

#pip freeze | grep 'langchain\|mlflow'
langchain==0.1.20
langchain-community==0.0.38
langchain-core==0.1.52
langchain-openai==0.1.6
langchain-text-splitters==0.0.1
mlflow==2.12.2

chip-davis commented 3 months ago

Have you tried specifying everything when you create the model? For example:

model = AzureChatOpenAI(
            openai_api_version="2023-05-15",
            azure_deployment=os.getenv("AZURE_OPENAI_DEPLOYMENT_NAME"),
            temperature=0,
            azure_endpoint=os.getenv("AZURE_OPENAI_ENDPOINT"),
            api_key=os.getenv("AZURE_OPENAI_API_KEY"),
        )

Not sure if this would work but give it a try.

Zuber139 commented 3 months ago

I have the exact same problem as OP, and no, that does not work.

Example code

Environment variables

# azure openai (embedding API)
AZURE_OPENAI_ENDPOINT=my_endpoint
AZURE_OPENAI_API_KEY=my_key

# OPENAI_API_BASE=my_endpoint
# OPENAI_API_KEY=my_key

AZURE_OPENAI_EMBEDDINGS=my_embedding
OPENAI_API_VERSION=my_version
OPENAI_API_TYPE=azure

Code

import os
from langchain_openai import AzureOpenAIEmbeddings

from dotenv import load_dotenv

load_dotenv('azure.env')

embeddings = AzureOpenAIEmbeddings(
    azure_endpoint=os.environ["AZURE_OPENAI_ENDPOINT"],
    api_key=os.environ["AZURE_OPENAI_API_KEY"],
    azure_deployment=os.environ["AZURE_OPENAI_EMBEDDINGS"])

text = "Hello, world!"
query_result = embeddings.embed_query(text)             
print(query_result)

In azure.env, if comment out OPEN_API_BASE, the code works. If I uncomment it, I get the same exact error that OP sees.

Error message

ValidationError: 1 validation error for AzureOpenAIEmbeddings
__root__
  As of openai>=1.0.0, Azure endpoints should be specified via the `azure_endpoint` param not `openai_api_base` (or alias `base_url`).  (type=value_error)

So no, specifying the endpoint as an argument does not fix this problem.

#pip freeze | grep 'langchain\|openai'
langchain==0.2.0
langchain-cli==0.0.22
langchain-community==0.2.0
langchain-core==0.2.1
langchain-experimental==0.0.59
langchain-openai==0.1.7
langchain-text-splitters==0.2.0
openai==1.30.1
openai-core==0.19.1
openai-schema-validator==0.6.2
openai-spec-validator==0.7.1