🎛️ Circuit Analysis API

Welcome to the Circuit Analysis API — a cutting-edge FastAPI application designed to analyze electronic circuit components using the Llama vision model via the Groq API. This project was a standout at the Meta Llama Hackathon: Toronto, clinching the 1st Prize among ~45 teams and 150 attendees.

🎥 Demo Video

Experience the capabilities of our API by watching the demo video:

📋 Prerequisites

Ensure you have the following before getting started:

• Python: Version 3.9 or higher
• Poetry: For dependency management
• Groq API Key: For Llama 3.2 90B
• FastAPI: For building the API
• Python-dotenv: For managing environment variables
• Uvicorn: For running the ASGI server
• Pydantic: For data validation and settings management
• Loguru: For logging
• OpenCV: For image processing
• Pillow: For image handling
• Requests: For making HTTP requests
• Pytest: For testing
• litellm: Allows switching between different language model APIs easily.

🚀 Setup

Follow these steps to set up the project:

1. Install Poetry if you haven't already:

   curl -sSL https://install.python-poetry.org | python3 -

2. Clone the repository and navigate to the project directory:

   git clone <repository-url>
   cd circuit-analysis-api

3. Install dependencies using Poetry:

   poetry install

4. Create a .env file in the project root with the following content:

   GROQ_API_KEY=your_groq_api_key_here

5. Set up SSL certificates for HTTPS:

   • If you have existing certificates, place them in the certs directory as cert.pem and key.pem.
   • To generate self-signed certificates, use OpenSSL:

     openssl req -x509 -newkey rsa:4096 -keyout certs/key.pem -out certs/cert.pem -days 365 -nodes

Note: The litellm library is used to facilitate switching between different language model APIs.

🏃‍♂️ Running the Application

1. Activate the Poetry virtual environment:

   poetry shell

2. Start the Phoenix server:

   phoenix serve

3. Start the FastAPI server:

   python run.py

• The API will be available at http://localhost:8000
• The Phoenix server will be available at http://localhost:6006

📚 API Documentation

Once the server is running, access the following documentation:

• Interactive API documentation (Swagger UI): http://localhost:8000/docs
• Alternative API documentation (ReDoc): http://localhost:8000/redoc

🏗️ Architecture

The Circuit Analysis API is structured into several key components, each responsible for a specific aspect of the circuit analysis process:

1. Component Identifier: Detects and identifies various electronic components within a circuit diagram using machine learning models and predefined prompts.

2. Sheet Identifier: Detects and extracts the circuit diagram from an image using a two-stage approach.

3. Connection Identifier: Determines the connections between identified components in the circuit.

Each component works in tandem to provide a comprehensive analysis of circuit diagrams, leveraging advanced machine learning techniques and robust service orchestration to deliver accurate and reliable results.

🧪 Running Tests

Run the test suite and generate coverage reports with:

poetry run pytest

📊 Running Benchmarks

Evaluate the accuracy and performance of the API's detection capabilities with benchmark tests:

Running Component Detection Benchmarks

python -m tests.benchmarks.circuit_components_benchmark

Running Connection Identification Benchmarks

python -m tests.benchmarks.circuit_connections_benchmark

Running Sheet Detection Benchmarks

python -m tests.benchmarks.sheet_detector_benchmark

Each benchmark script will output detailed logs and statistics, including accuracy rates and any errors encountered during the tests.

📚 Lessons Learned

• Prompt Engineering: Not the optimal solution for this problem.
• Model Fine-Tuning: Extensive fine-tuning is required for models to perform well on these examples.
• Future Solutions: Consider using architectures like YOLOv5 for initial detection and Llama for reasoning between detected components.

📜 License

This project is licensed under the Apache License 2.0. See the LICENSE file for details.