FIPS Chat

A FIPS 140-2 compliant multi-provider AI chat and image analysis platform designed for containerized deployment in OpenShift environments.

Features

• 💬 Chat Interface: Real-time conversations with Ollama models
• 🖼️ Image Analysis: Upload and analyze images with vision models
• 🎛️ Model Management: Dynamic model selection and information display
• 📊 Session Statistics: Track usage, response times, and performance
• 💾 Export Functionality: Export conversations and analyses
• ⚙️ Configurable Settings: Adjust temperature, tokens, and other parameters
• 🔒 FIPS Compliance: Built with FIPS 140-2 cryptographic standards

🚨 IMPORTANT: Production Deployment Requirements

This application MUST be deployed as a container for FIPS compliance. Direct Python execution is only supported for development.

Production Deployment

• ✅ Container deployment (Podman/Docker + OpenShift)
• ✅ FIPS-enabled environment
• ✅ Container registry access

Development Only

• ⚠️ Local Python execution (development/testing only)

Prerequisites

For Production (Container Deployment):

• Podman or Docker
• OpenShift 4.8+ cluster with FIPS mode enabled
• Access to container registry (Quay.io, Docker Hub, etc.)
• Ollama service deployed in OpenShift

For Development Only:

• Python 3.11+
• Ollama installed and running locally
• At least one Ollama model installed (see recommended models below)

Quick Start

Production Deployment (Container)

1. Build the FIPS-compliant container:

   ./scripts/build-podman.sh

2. Tag and push to registry:

   podman tag ollama-streamlit:latest quay.io/your-username/fips-chat:latest
   podman push quay.io/your-username/fips-chat:latest

3. Deploy to OpenShift:

   cd openshift/
   # Update image reference in deployment.yaml
   oc apply -k .

Development Setup (Local Only)

1. Install dependencies:

   pip install -r requirements.txt

2. Install Ollama models (recommended):

   # Vision models for image analysis
   ollama pull llava:7b
   ollama pull granite3.2-vision:latest
   
   # Chat models for conversations
   ollama pull granite3.3:8b
   ollama pull gemma3:latest
   ollama pull phi4-mini:3.8b
   
   # Code-focused model
   ollama pull qwen2.5-coder:7b

3. Run the application (development only):

   # Start Ollama service
   ollama serve
   
   # Run Streamlit application
   streamlit run app.py
   
   # Open browser to http://localhost:8501

Container Testing

Test the container locally before deploying:

# Test with provided script
./scripts/test-podman.sh

# Or manual testing
podman run -p 8080:8080 --rm ollama-streamlit:latest
# Open browser to http://localhost:8080

Model Management

After deployment, you'll need to install Ollama models. The application provides several ways to manage models:

🔗 Admin Route (Easiest)

Access the admin interface at: https://ollama-admin-{namespace}.apps.{cluster}/

# Deploy a test model
curl -X POST https://ollama-admin-ollama-platform.apps.your-cluster.com/api/pull \
  -d '{"name": "llama3.2:1b"}' -H "Content-Type: application/json"

# Check available models  
curl -s https://ollama-admin-ollama-platform.apps.your-cluster.com/api/tags

🔧 CLI Access (Advanced)

# Port forward to Ollama
oc port-forward service/ollama-service 11434:11434 &

# Deploy models
curl -X POST http://localhost:11434/api/pull \
  -d '{"name": "granite3.3:8b"}' -H "Content-Type: application/json"

See DEPLOYMENT.md for complete model management guide.

Recommended Models

Primary Models (Balanced Performance)

• llava:7b - Primary vision model for image description (4.7 GB)
• granite3.2-vision:latest - Alternative vision model (2.4 GB)
• granite3.3:8b - Primary chat model (4.9 GB)
• gemma3:latest - Lightweight chat alternative (3.3 GB)
• phi4-mini:3.8b - Fast response chat model (2.5 GB)

Secondary Models (Power Users)

• qwen2.5-coder:7b - Code-focused conversations (4.7 GB)
• mistral-small3.1:24b - High-quality responses (15 GB)

Application Structure

fips-chat/
├── app.py                    # Main Streamlit application
├── config.py                # Configuration management
├── ollama_client.py          # Ollama API wrapper
├── ui_components/            # Reusable UI components
│   ├── chat_interface.py     # Chat functionality
│   ├── image_interface.py    # Image analysis
│   └── model_selector.py     # Model management
├── utils/                    # Utility functions
│   ├── image_processing.py   # Image utilities
│   └── session_manager.py    # Session management
├── tests/                    # Test suite
├── requirements.txt          # Dependencies
└── README.md                # This file

Configuration

The application can be configured through environment variables:

export OLLAMA_HOST="http://localhost:11434"  # Ollama server URL
export DEFAULT_CHAT_MODEL="granite3.3:8b"   # Default chat model
export DEFAULT_VISION_MODEL="llava:7b"      # Default vision model
export TEMPERATURE="0.7"                     # Model temperature
export MAX_TOKENS="2048"                     # Max response length
export MAX_FILE_SIZE_MB="10"                 # Max image file size

Testing

Run the test suite:

# Run all tests
python -m pytest tests/ -v

# Run specific test modules
python -m pytest tests/test_config.py -v
python -m pytest tests/test_ollama_client.py -v
python -m pytest tests/test_image_processing.py -v

Performance Targets

• Model list refresh: < 2 seconds
• Chat response initiation: < 3 seconds
• Image analysis initiation: < 5 seconds
• UI responsiveness: No blocking operations

Troubleshooting

Common Issues

1. "Cannot connect to Ollama"

   • Ensure Ollama is running: ollama serve
   • Check if Ollama is accessible: ollama list

2. "No models available"

   • Install at least one model: ollama pull llava:7b
   • Verify models are installed: ollama list

3. "Model does not support image analysis"

   • Select a vision model (llava:7b, granite3.2-vision:latest)
   • Check model capabilities in the Models tab

4. High memory usage warning

   • Use the "Clean Up" button in the sidebar
   • Clear conversation history
   • Reduce image file sizes

Performance Tips

• Use smaller models (phi4-mini:3.8b, gemma3:latest) for faster responses
• Reduce image file sizes before upload
• Clear old conversations and images regularly
• Adjust temperature and max tokens for your use case

OpenShift Deployment

For FIPS-compliant deployment to OpenShift:

# Build with Podman
./scripts/build-podman.sh

# Test locally
./scripts/test-podman.sh

# Deploy to OpenShift
oc apply -k openshift/

See DEPLOYMENT.md for comprehensive OpenShift deployment instructions.

FIPS Compliance

✅ This application is FIPS 140-2 compliant

• No weak cryptographic functions (MD5, SHA1, etc.)
• Container runs with OPENSSL_FIPS=1
• Uses OCI-compliant Containerfile for Podman

License

This project is provided as-is for educational and development purposes.