tests/README.md

# Test Suite

Comprehensive test coverage for the LinguaCustodia Financial AI API.

## Running Tests

### Prerequisites
1. Start the API server: `python app_clean.py`
2. Install test dependencies: `pip install pytest requests`

### Run Tests
```bash
# Run all tests
pytest tests/

# Run specific test files
pytest tests/test_api.py
pytest tests/test_inference.py
pytest tests/test_openai_compatibility.py

# Run with verbose output
pytest tests/ -v

# Run with coverage
pytest tests/ --cov=app
```

## Test Categories

### `test_api.py`
- Health check endpoint
- Models listing endpoints
- Import diagnosis endpoint
- Basic API functionality

### `test_inference.py`
- Basic inference functionality
- Different model parameters
- Error handling
- Performance timing

### `test_openai_compatibility.py`
- Chat completions endpoint
- Text completions endpoint
- JSON response format
- Streaming responses
- RAG document analysis

## Test Configuration

Tests automatically skip if the API server is not running. The `conftest.py` file handles:
- API availability checking
- Automatic test skipping
- Common fixtures

## Expected Behavior

All tests should pass when the API is running properly. Tests cover:
- ✅ API endpoint availability
- ✅ Response format validation
- ✅ OpenAI compatibility
- ✅ Error handling
- ✅ Performance expectations