Portfolio Assistant

This is a conversational assistant developed to answer questions about my portfolio and qualifications. It utilizes Azure Functions, Python, Azure AI Search, and Azure AI Foundry and is deployed via GitHub Actions. for a version without Azure AI Search, please see the chat-only branch.

Project Structure

├── host.json # Azure Functions host configuration
├── local.settings.json # Local environment variables (not included in repository, not for production)
├── requirements.txt # Dependencies
├── .github/workflows/ # GitHub Actions pipeline
│ └── deploy.yml
├── ChatFunction/ # Azure Function entrypoint
│ ├── init.py
│ └── function.json
├── src/portfolio_assistant/
│ ├── agent_client.py # Client for Azure AI Foundry chat agent
│ ├── config.py
│ ├── utils.py
│ ├── search_client.py # Client to support Azure AI Search │ └── init.py
└── tests/ # Unit tests (pytest)
├── test_agent_client.py
├── test_api.py
└── test_search_client.py

Architecture

Setup Instructions

git clone https://github.com/vmgaribay/portfolio-assistant.git
cd portfolio-assistant
python -m venv .venv
source .venv/bin/activate   # or for Windows: .venv\Scripts\activate
pip install -r requirements.txt

Running Locally

It is necessary to create your own...

• Azure OpenAI resource and deploy a chat model to obtain the endpoint and API key (I tested with 4.1-mini for a balance of cost and performance).
• Azure AI Search resource with semantic search capabilities.
• An index populated with relevant documents including metadata for the added fields "topics" and "source" for best results. (I indexed an Azure blob storage container with PDFs and markdown files related to my portfolio content).

Prerequisites:

• Azure Functions Core Tools
• Python 3.10
• An endpoint and key for an Azure OpenAI resource with a deployed chat model

Set up local environment variables in local.settings.json:

{  
  "IsEncrypted": false,  
  "Values": {  
    "FUNCTIONS_WORKER_RUNTIME": "python",  
    "AzureWebJobsStorage": "UseDevelopmentStorage=true",  
    "AZURE_CHAT_AGENT_ENDPOINT": "<your-endpoint>",  
    "AZURE_CHAT_API_KEY": "<your-api-key>",
    "AZURE_SEARCH_ENDPOINT": "<your-search-endpoint>",  
    "AZURE_SEARCH_API_KEY": "<your-search-api-key>",  
    "AZURE_SEARCH_INDEX_NAME": "<your-search-index-name>",
    "AZURE_SEARCH_API_VERSION": "<latest stable version>"
  }  
}

Run the Azure Function locally:

func start

Test the API endpoint:

curl -X POST http://localhost:7071/api/chat-function -H "Content-Type: application/json" -d '{"message": "Is this functioning?"}'

Deployment

The project is set up to deploy automatically to Azure Functions via GitHub Actions. It is first necessary to create an Azure Function App with Python runtime since its name and publishing profile are necessary for deployment. Ensure you have the following secrets in your GitHub repository settings:

• AZURE_FUNCTIONAPP_NAME
• AZURE_FUNCTIONAPP_PUBLISH_PROFILE Push changes to the main branch to trigger deployment. Add the environment variables from your working local.settings.json to the Function App Configuration in the Azure Portal.

Future Enhancements

• Easily interchangeable system prompt for different contexts

Acknowledgements

This project was made possible by resources such as GitHub Copilot (GPT-5, GPT-4.1, Claude Sonnet 3.7), Azure Learn, too many YouTube tutorials, patience, and willpower.

Contact

Victoria Garibay, Ph.D. - Contact Form | GitHub Profile