Skip to content

Assistants with Azure Cognitive Search and Azure Identity#

Open In Colab Open on GitHub

This notebook demonstrates the use of Assistant Agents in conjunction with Azure Cognitive Search and Azure Identity. Assistant Agents use tools that interact with Azure Cognitive Search to extract pertinent data.

Prerequisites#

Before running this notebook, please ensure the following prerequisites are met:

Dependencies#

  1. Autogen
  2. Azure SDK
  3. Cognitive Search/AI Search

If you have AI search enabled in your Azure Portal, you can use the following code to create an assistant agent that can search Azure Cognitive Search.

AI search setup details: - Documentation:
- Create search service: https://learn.microsoft.com/en-us/azure/search/search-create-service-portal - Search index: https://learn.microsoft.com/en-us/azure/search/search-how-to-create-search-index?tabs=portal hybrid search: https://learn.microsoft.com/en-us/azure/search/hybrid-search-how-to-query

  • Youtube walkthrough: https://www.youtube.com/watch?v=6Zfuw-UJZ7k

Install Azure CLI#

This notebook requires the Azure CLI for authentication purposes. Follow these steps to install and configure it:

  1. Download and Install Azure CLI:
    • Visit the Azure CLI installation page and follow the instructions for your operating system.
    • Mac users can install Azure CLI using Homebrew with the command brew install azure-cli
  2. Verify Installation:
    • In the below cell execute az --version to check if Azure CLI is installed correctly.
  3. Login to Azure:
    • In the below cell execute az login to log into your Azure account. This step is necessary as the notebook uses AzureCliCredential which retrieves the token based on the Azure account currently logged in.

Check Azure CLI Installation#

Run the cell below to check if Azure CLI is installed and properly configured on your system.

Check Azure CLI Installation and Login Status#

# Check Azure CLI installation and login status
# !az --version
# !az login

Install required packages#

Run the cell below to install the required packages for this notebook.

!pip3 install pyautogen==0.2.16
!pip3 install python-dotenv==1.0.1
!pip3 install pyautogen[graph]>=0.2.11
!pip3 install azure-search-documents==11.4.0b8
!pip3 install azure-identity==1.12.0

Next you will import the required packages for this notebook.

import json
import os

from azure.identity import DefaultAzureCredential
from azure.search.documents import SearchClient
from dotenv import load_dotenv

import autogen
from autogen import AssistantAgent, UserProxyAgent, register_function
from autogen.cache import Cache

load_dotenv()

# Import Cognitive Search index ENV
AZURE_SEARCH_SERVICE = os.getenv("AZURE_SEARCH_SERVICE")
AZURE_SEARCH_INDEX = os.getenv("AZURE_SEARCH_INDEX")
AZURE_SEARCH_KEY = os.getenv("AZURE_SEARCH_KEY")
AZURE_SEARCH_API_VERSION = os.getenv("AZURE_SEARCH_API_VERSION")
AZURE_SEARCH_SEMANTIC_SEARCH_CONFIG = os.getenv("AZURE_SEARCH_SEMANTIC_SEARCH_CONFIG")
AZURE_SEARCH_SERVICE_ENDPOINT = os.getenv("AZURE_SEARCH_SERVICE_ENDPOINT")

Next, you need to authenticate and create a SearchClient instance.

credential = DefaultAzureCredential()
endpoint = AZURE_SEARCH_SERVICE_ENDPOINT

from azure.identity import AzureCliCredential

credential = AzureCliCredential()
token = credential.get_token("https://cognitiveservices.azure.com/.default")

print("TOKEN", token.token)

client = SearchClient(endpoint=endpoint, index_name="test-index", credential=credential)

Then, load the configuration list and define the configuration for the AssistantAgent.

config_list = autogen.config_list_from_json(
    env_or_file="OAI_CONFIG_LIST",
)

gpt4_config = {
    "cache_seed": 42,
    "temperature": 0,
    "config_list": config_list,
    "timeout": 120,
}

Define your tool function search that will interact with the Azure Cognitive Search service.

def search(query: str):
    payload = json.dumps({
        "search": query,
        "vectorQueries": [{"kind": "text", "text": query, "k": 5, "fields": "vector"}],
        "queryType": "semantic",
        "semanticConfiguration": AZURE_SEARCH_SEMANTIC_SEARCH_CONFIG,
        "captions": "extractive",
        "answers": "extractive|count-3",
        "queryLanguage": "en-US",
    })

    response = list(client.search(payload))

    output = []
    for result in response:
        result.pop("titleVector")
        result.pop("contentVector")
        output.append(result)

    return output

Define the AssistantAgent and UserProxyAgent instances, and register the search function to them.

cog_search = AssistantAgent(
    name="COGSearch",
    system_message="You are a helpful AI assistant. "
    "You can help with Azure Cognitive Search."
    "Return 'TERMINATE' when the task is done.",
    llm_config=gpt4_config,
)

user_proxy = UserProxyAgent(
    name="User",
    llm_config=False,
    is_termination_msg=lambda msg: msg.get("content") is not None and "TERMINATE" in msg["content"],
    human_input_mode="NEVER",
)

register_function(
    search,
    caller=cog_search,
    executor=user_proxy,
    name="search",
    description="A tool for searching the Cognitive Search index",
)

Finally, initiate a chat.

if __name__ == "__main__":

    async def main():
        with Cache.disk() as cache:
            await user_proxy.a_initiate_chat(
                cog_search,
                message="Search for 'What is Azure?' in the 'test-index' index",
                cache=cache,
            )

    await main()