Streamlit Search App

A Streamlit-based intelligent chat application that integrates Azure OpenAI and Azure Cognitive Search services to provide document-based question answering capabilities.

Features

• Direct Chat Interface: Direct conversation with Azure OpenAI GPT models
• Document Search & Chat: Question-answering based on indexed documents using Azure Cognitive Search
• Multi-Purpose Templates: Pre-configured for antenna parts selection and IT helpdesk scenarios
• Configuration Management: Web-based settings panel for Azure service configuration
• Index Management: Built-in tools to create and manage Azure Cognitive Search indexes

Application Structure

Main Components

1. Home: Basic chat interface using Azure OpenAI directly
2. Settings: Configuration panel for Azure services and index management
3. Upload File: File upload functionality (placeholder)
4. Chat: Advanced chat with document search capabilities using Cognitive Search

Core Files

• webapp.py: Main Streamlit application with user interface
• cog_search.py: Azure Cognitive Search integration and OpenAI helpers
• credentials.py: Configuration settings and service credentials
• requirements.txt: Python dependencies

Prerequisites

Azure Services Required

1. Azure OpenAI Service

   • Deployed GPT-35-Turbo model
   • Text-Davinci-003 model
   • API endpoint and access keys

2. Azure Cognitive Search

   • Search service instance
   • API keys with admin permissions

3. Azure Storage Account

   • Blob storage container for documents
   • Connection string and access keys

4. Azure Cognitive Services

   • Multi-service cognitive services resource
   • API keys for text processing

Installation

1. Clone the repository

   git clone https://github.com/memasanz/streamlitsearchapp.git
   cd streamlitsearchapp

2. Install dependencies

   pip install streamlit
   pip install streamlit-chat
   pip install streamlit-option-menu
   pip install openai
   pip install langchain
   pip install faiss-cpu
   pip install transformers
   pip install pandas
   pip install numpy
   pip install requests

3. Configure Azure services

   Edit credentials.py with your Azure service details:

   # Azure OpenAI Configuration
   AZURE_OPENAI_ENDPOINT = "https://your-openai-service.openai.azure.com/"
   AZURE_OPENAI_KEY = "your-openai-api-key"
   
   # Azure Cognitive Search Configuration
   COG_SEARCH_RESOURCE = "your-search-service-name"
   COG_SEARCH_KEY = "your-search-api-key"
   COG_SEARCH_INDEX = "your-index-name"
   
   # Azure Storage Configuration
   STORAGE_CONNECTION_STRING = "your-storage-connection-string"
   STORAGE_ACCOUNT = "your-storage-account-name"
   STORAGE_KEY = "your-storage-key"

Usage

Running the Application

streamlit run webapp.py

The application will start on http://localhost:8501

Initial Setup

1. Configure Services: Navigate to the "Settings" tab and enter your Azure service credentials
2. Create Search Index: Use the "Create Index" button to set up your Cognitive Search index
3. Upload Documents: Upload documents to your Azure Storage container for indexing

Using the Chat Features

Direct Chat (Home Tab)

• Simple conversation interface with Azure OpenAI
• Uses GPT-35-Turbo model for responses
• No document context - general purpose chat

Document-Based Chat (Chat Tab)

• Ask questions about your uploaded documents
• Uses Cognitive Search to find relevant content
• Combines search results with OpenAI for comprehensive answers

Configuration Options

Question Templates

The application includes pre-configured templates for specific use cases:

Antenna Parts Selector

XTG_TEMPPLATE_ORIG = """
You are an antenna part selector chat bot.  
You are given sections of a catalog of antennas.
Each document section includes the part number at the beginning...
"""

IT Helpdesk

TTEC_questiontemplate = """
You are an IT HelpDesk Chat Bot.  
You are to answer questions based on the context provided...
"""

Model Configuration

• GPT-35-Turbo: For chat conversations
• Text-Davinci-003: For document-based Q&A
• Text-Search-Curie: For embeddings and similarity search

Architecture

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   Streamlit UI  │────│   Azure OpenAI   │────│  Azure Cognitive│
│                 │    │                  │    │     Search      │
│ - Chat Interface│    │ - GPT Models     │    │                 │
│ - Settings      │    │ - Embeddings     │    │ - Document Index│
│ - File Upload   │    │ - Text Generation│    │ - Search & Rank │
└─────────────────┘    └──────────────────┘    └─────────────────┘
         │                        │                        │
         └────────────────────────┼────────────────────────┘
                                  │
                      ┌──────────────────┐
                      │  Azure Storage   │
                      │                  │
                      │ - Document Store │
                      │ - Blob Container │
                      └──────────────────┘

Advanced Features

Index Management

The application provides built-in tools to:

• Create data sources pointing to your Azure Storage
• Set up cognitive skillsets for document processing
• Create and configure search indexes
• Run indexers to process and index documents

Session Management

• Chat history is maintained in Streamlit session state
• Settings can be updated dynamically
• Message reset functionality available

Troubleshooting

Common Issues

1. Authentication Errors

   • Verify Azure service credentials in credentials.py
   • Ensure API keys have proper permissions
   • Check endpoint URLs are correct

2. Search Not Working

   • Confirm Cognitive Search index exists
   • Verify documents are uploaded to storage container
   • Check indexer has run successfully

3. Chat Responses Incomplete

   • Increase max_tokens in OpenAI configuration
   • Verify model deployment names match configuration

Debug Mode

Set DEBUG = "1" in credentials.py to enable additional logging.

Security Considerations

• Never commit actual API keys to version control
• Use environment variables for production deployments
• Regularly rotate API keys and connection strings
• Implement proper access controls for Azure resources

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

License

This project is provided as-is for demonstration purposes. Please ensure compliance with Azure service terms and conditions.

Support

For issues and questions:

1. Check the troubleshooting section above
2. Review Azure service documentation
3. Open an issue in the GitHub repository