Cogniware IMS End-to-End Tests¶

This directory contains end-to-end (E2E) tests for the OPEA Cogniware Inventory Management System.

Test Coverage¶

Docker Compose Tests¶

test_compose_on_xeon.sh - Tests Docker Compose deployment on Intel Xeon processors
test_gmc_on_xeon.sh - Tests GMC deployment on Kubernetes with Xeon

Prerequisites¶

For Docker Compose Tests¶

Docker and Docker Compose installed
16GB+ RAM (32GB recommended)
HuggingFace API token
Intel Xeon processor (recommended)

For Kubernetes Tests¶

Kubernetes cluster (v1.24+)
kubectl configured
Helm 3.0+ or GMC installed

Running Tests¶

Docker Compose Test on Intel Xeon¶

export HUGGINGFACEHUB_API_TOKEN=your_token_here
cd tests
./test_compose_on_xeon.sh

Kubernetes GMC Test¶

cd tests
./test_gmc_on_xeon.sh

Test Workflow¶

Each test follows this workflow:

Build: Build Docker images from source
Deploy: Start all services with docker compose
Validate Microservices: Check health of all OPEA microservices
- Redis vector database
- PostgreSQL database
- TGI service
- LLM microservice
- Embedding microservice
- Retriever microservice
- Reranking microservice
- Data prep microservice
Validate Backend: Test Cogniware IMS backend endpoints
- Health check
- Chat completion
- Knowledge base stats
Validate Frontend: Verify UI accessibility
Cleanup: Stop and remove all containers

Test Validation Points¶

Microservices¶

✅ Redis vector database connectivity
✅ PostgreSQL database ready
✅ TGI service health
✅ LLM microservice endpoint
✅ Embedding microservice endpoint
✅ Retriever microservice endpoint
✅ Reranking microservice endpoint
✅ Data preparation microservice

Application¶

✅ Backend health check
✅ Chat completion endpoint
✅ Inventory query endpoint
✅ Knowledge base statistics
✅ Frontend UI accessibility

Environment Variables¶

Required¶

HUGGINGFACEHUB_API_TOKEN - Your HuggingFace API token

Optional¶

LLM_MODEL_ID - LLM model (default: Intel/neural-chat-7b-v3-3)
EMBEDDING_MODEL_ID - Embedding model (default: BAAI/bge-base-en-v1.5)
RERANK_MODEL_ID - Reranking model (default: BAAI/bge-reranker-base)
POSTGRES_PASSWORD - PostgreSQL password (default: postgres)

Troubleshooting¶

Services Not Starting¶

# Check logs
cd docker_compose/intel/xeon
docker compose logs <service-name>

# Check Docker resources
docker stats

Port Conflicts¶

# Check if ports are in use
netstat -tulpn | grep -E '3000|6000|6379|7000|8000|9000'

# Stop conflicting services
docker ps -a
docker stop <container-id>

Out of Memory¶

Ensure at least 16GB RAM available
Increase Docker memory limit in Docker Desktop
Use smaller models or reduce batch sizes

Model Download Issues¶

# Verify HuggingFace token
echo $HUGGINGFACEHUB_API_TOKEN

# Check TGI logs
docker compose logs tgi-service

# Manual download if needed
docker exec -it tgi-service bash
huggingface-cli download Intel/neural-chat-7b-v3-3

Test Duration¶

Build Phase: 5-10 minutes (first time)
Startup Phase: 2-5 minutes
Validation Phase: 2-3 minutes
Total: ~10-20 minutes

CI/CD Integration¶

These tests are designed for GitHub Actions integration:

name: E2E Tests
on: [push, pull_request]

jobs:
  test-xeon:
    runs-on: [self-hosted, xeon]
    steps:
      - uses: actions/checkout@v3
      - name: Run Intel Xeon tests
        env:
          HUGGINGFACEHUB_API_TOKEN: ${{ secrets.HF_TOKEN }}
        run: ./tests/test_compose_on_xeon.sh

Contributing¶

When adding new tests:

Follow naming convention: test_<type>_on_<hardware>.sh
Include proper error handling with set -e
Add cleanup functions
Document prerequisites
Update this README

Support¶

For issues:

Check logs: docker compose logs -f
Review README files in parent directories
Open an issue on GitHub