Non Deployment API call support for AzureOpenAI class

[akshayjadiya]

Hi all,
I have access to OpenAI LLMs using a gateway from my organization. I am able to call the Azure OpenAI LLM using the below LangChain code -

from langchain.schema import HumanMessage, SystemMessage
import ssl
import httpx
from typing import Dict
import os 

    
# If you are using API Key mode authentication use this headers instead
api_key = <api_key>
headers: Dict[str, str] = { "X-Api-Key": api_key }

WMT_CA_PATH = "/Users/ca-bundle.crt" 
os.environ['SSL_CERT_FILE'] = WMT_CA_PATH
os.environ['REQUESTS_CA_BUNDLE'] = WMT_CA_PATH

# Build a custom client with Walmart CAs and Headers built into it
context = ssl.create_default_context()
context.load_verify_locations(WMT_CA_PATH);
client = httpx.Client(verify=context, headers=headers)


# Chat with OpenAI
chat = AzureChatOpenAI(openai_api_key=api_key,
                       model="gpt-4",
                       api_version="2023-05-15",
                       azure_endpoint="<endpoint_provided_by_my_employer>",
                       http_client=client)

messages = [
    SystemMessage(content="You are a helpful assistant that translates English to Spanish"),
    HumanMessage(content="Good morning, Vietnam!")
]
response = chat(messages)

However, using llamaindex AzureOpenAI class requires me to pass deployment_name argument as well that I DO NOT KNOW. Is there a workaround?

from llama_index.llms.azure_openai import AzureOpenAI

llm = AzureOpenAI(
    deployment_name="gpt-4o", 
    model_name="gpt-4o",
    azure_endpoint="https://your_resource_name.openai.azure.com/",
    api_version="2023-07-01-preview",
    temperature=0.0,
    api_key="YOUR_AZURE_OPENAI_API_KEY"
)

[dosubot[bot]]

Hey @akshayjadiya! I'm here to help you with any bugs, questions, or contributions. Let's tackle this issue together!

I found a similar discussion regarding the use of the AzureOpenAI class without specifying a deployment_name. The deployment_name parameter is crucial for identifying your specific deployment on Azure OpenAI, and it acts as an alias for the engine parameter, which is necessary to specify the custom name you assigned to your model deployment on Azure ^[1][2].

Unfortunately, there are no known workarounds or alternative methods for using the AzureOpenAI class when the deployment_name is unknown ^[1]. The engine parameter is mandatory and cannot be omitted, as it identifies your deployment ^[1].

If you have access to the deployment through a gateway or another method, you might need to consult with your organization's IT or Azure administrator to obtain the necessary deployment_name or engine identifier.

To continue talking to Dosu, mention @dosu.