langchain-integration-health 0.1.5

Comprehensive testing framework and dashboard for LangChain integration compatibility

LangChain Integration Health Dashboard & Testing Framework

A comprehensive testing framework and dashboard for monitoring LangChain integration compatibility, addressing critical pain points in the LangChain ecosystem.

Problem Statement

LangChain has major integration compatibility issues:

• Missing .bind_tools() methods in integrations like MLXPipeline
• Inconsistent API support across different model providers
• Poor integration testing leading to breaking changes
• No visibility into which integrations support which features
• Developers waste time discovering integration limitations at runtime

Solution

This framework provides:

1. Integration Compatibility Testing Framework - Automated testing suite that validates LangChain integrations
2. Real-time Integration Health Dashboard - Web-based dashboard showing integration compatibility status
3. Integration Template Generator - Standardized templates for new integrations

Quick Start

Installation

From PyPI (Recommended)

pip install langchain-integration-health

From Source

git clone https://github.com/sadiqkhzn/langchain-integration-health.git
cd langchain-integration-health
pip install -e .

Run the Dashboard

langchain-health dashboard

Or run Streamlit directly:

streamlit run -m langchain_integration_health.dashboard.app

CLI Usage

# Discover available integrations
langchain-health discover

# Run integration tests
langchain-health test

# Launch the dashboard
langchain-health dashboard

# Generate compatibility report
langchain-health report

# Clean old test results
langchain-health clean

Programmatic Usage

import asyncio
from langchain_integration_health.testers import LLMIntegrationTester
from langchain_integration_health.utils.config import Config

# Test a specific integration
async def test_integration():
    config = Config.from_env()
    
    # Example: Test OpenAI integration
    from langchain_openai import ChatOpenAI
    
    tester = LLMIntegrationTester(ChatOpenAI, config.get_integration_config("ChatOpenAI"))
    result = await tester.run_all_tests()
    
    print(f"Compatibility Score: {result.compatibility_score}")
    print(f"bind_tools Support: {result.bind_tools_support}")
    print(f"Streaming Support: {result.streaming_support}")

asyncio.run(test_integration())

Features

Comprehensive Testing

• Tests for required methods: bind_tools(), stream(), with_structured_output(), etc.
• Compatibility matrix generation
• Performance benchmarking
• Error handling validation

Real-time Dashboard

• Visual compatibility matrix with color-coded status indicators
• Detailed test results and error reporting
• Performance metrics and benchmarking data
• Historical trend analysis
• Export capabilities (JSON, CSV, Markdown)

Automatic Discovery

• Scans installed packages for LangChain integrations
• Supports main langchain, langchain-community, and third-party packages
• Parallel testing for faster results

Integration Fixes

• Example implementations for common issues (e.g., MLXPipeline bind_tools fix)
• Wrapper patterns for adding missing functionality
• Best practices for LangChain compatibility

Architecture

langchain-integration-health/
├── src/
│   ├── testers/           # Testing framework
│   │   ├── base_tester.py
│   │   ├── llm_tester.py
│   │   ├── chat_model_tester.py
│   │   └── embeddings_tester.py
│   ├── dashboard/         # Streamlit dashboard
│   │   ├── app.py
│   │   ├── components.py
│   │   └── data_loader.py
│   ├── utils/             # Utilities
│   │   ├── config.py
│   │   ├── reporters.py
│   │   └── discovery.py
│   └── examples/          # Example implementations
│       └── mlx_pipeline_fix.py
├── .github/workflows/     # CI/CD integration
└── tests/                 # Test suite

Testing Framework

Base Classes

• BaseIntegrationTester: Abstract base for all integration testers
• LLMIntegrationTester: Specialized for LLM integrations
• ChatModelTester: Specialized for chat models
• EmbeddingsTester: Specialized for embedding models

Test Types

1. Method Existence Tests: Verify required methods are present
2. Functionality Tests: Test actual method execution
3. Error Handling Tests: Validate proper error handling
4. Performance Tests: Measure latency and throughput
5. Compatibility Tests: Check version compatibility

Dashboard Features

Compatibility Matrix

Color-coded grid showing which integrations support which features:

• Green: High compatibility (>=0.8)
• Yellow: Medium compatibility (0.5-0.8)
• Red: Low compatibility (<0.5)

Integration Details

Expandable sections showing:

• Full test results
• Error and warning details
• Performance metrics
• Historical data

Export Options

• JSON: Structured data for programmatic use
• CSV: Spreadsheet-compatible format
• Markdown: Human-readable reports

Configuration

Environment Variables

# Database
LIH_DATABASE_URL=sqlite:///integration_health.db

# Testing
LIH_TEST_TIMEOUT=30
LIH_PARALLEL_TESTS=true
LIH_MOCK_MODE=false

# Dashboard
LIH_DASHBOARD_HOST=localhost
LIH_DASHBOARD_PORT=8501

# API Keys (optional, for real testing)
OPENAI_API_KEY=your_key_here
ANTHROPIC_API_KEY=your_key_here
GOOGLE_API_KEY=your_key_here

Configuration File

{
  "database_url": "sqlite:///integration_health.db",
  "test_timeout": 30,
  "parallel_tests": true,
  "mock_mode": false,
  "api_keys": {
    "openai": "your_key_here",
    "anthropic": "your_key_here"
  }
}

CI/CD Integration

The framework includes GitHub Actions workflow for automated testing:

# .github/workflows/integration-tests.yml
# Runs daily and on PRs to test all integrations

Features:

• Automatic integration discovery
• Parallel testing across integrations
• Report generation and artifact upload
• PR comment integration

MLXPipeline Fix Example

The framework includes a complete example of how to fix the MLXPipeline bind_tools issue:

from langchain_integration_health.examples.mlx_pipeline_fix import create_mlx_wrapper

# Original MLXPipeline (missing bind_tools)
mlx = MLXPipeline(model_name="mlx-community/Llama-3.2-1B-Instruct-4bit")

# Wrap for LangChain compatibility
langchain_mlx = create_mlx_wrapper(mlx)

# Now you can use bind_tools!
from langchain.tools import tool

@tool
def calculator(a: int, b: int) -> int:
    """Add two numbers"""
    return a + b

mlx_with_tools = langchain_mlx.bind_tools([calculator])
response = mlx_with_tools.invoke("What is 5 + 3?")

API Reference

Testing Classes

BaseIntegrationTester

Base class for all integration testers.

class BaseIntegrationTester:
    def __init__(self, integration_class: Type, config: Optional[Dict] = None)
    async def run_all_tests(self) -> IntegrationTestResult
    def _test_required_methods(self) -> Dict[str, bool]

LLMIntegrationTester

Specialized tester for LLM integrations.

class LLMIntegrationTester(BaseIntegrationTester):
    REQUIRED_METHODS = ['invoke', 'ainvoke', 'stream', 'astream', 'bind_tools', 'with_structured_output']
    
    async def _test_bind_tools_support(self) -> bool
    async def _test_streaming_support(self) -> bool

Data Models

IntegrationTestResult

Stores test results for an integration.

@dataclass
class IntegrationTestResult:
    integration_name: str
    integration_version: str
    test_timestamp: datetime
    bind_tools_support: bool
    streaming_support: bool
    structured_output_support: bool
    async_support: bool
    errors: List[str]
    warnings: List[str]
    performance_metrics: Dict[str, float]
    compatibility_score: float

Contributing

1. Fork the repository
2. Create a feature branch
3. Add tests for your changes
4. Ensure all tests pass
5. Submit a pull request

Development Setup

git clone https://github.com/sadiqkhzn/langchain-integration-health.git
cd langchain-integration-health
pip install -e ".[dev]"
pytest tests/

License

MIT License - see LICENSE file for details.

Support

For issues and feature requests, please use the GitHub issue tracker.