Imported from XXXws/pytorch-openpose (AGENTS.md). Install upstream with npx skills add XXXws/pytorch-openpose. Copyright stays with the author.

Repository Guide: PyTorch OpenPose

Overview

This project provides a web-based pose detection system using PyTorch. The repository contains a Python backend powered by FastAPI and a Vue.js frontend for user interaction. Core OpenPose logic resides in src/ and is wrapped into API services under app/.

├── app/        # FastAPI backend (routes & services)
├── src/        # PyTorch OpenPose models and utilities
├── frontend/   # Vue 3 web client
├── model/      # Pretrained model files (not included)
├── images/     # Sample images for demos
├── results/    # Output files (ignored by git)

Backend

• Entry point: python -m app.main (uses FastAPI and uvicorn)
• Configuration: see app/config.py for paths and log level.
• API modules: located in app/api/ (detection, video, realtime, health).
• Core services: app/core/ implements detection service, video tasks, performance monitoring and FFmpeg utilities.
• Pose models: defined in src/model.py with wrappers in src/body.py and src/hand.py.

Coding style

• Follow PEP8 conventions; use four spaces for indentation.
• Keep line length under 120 characters.
• Prefer explicit logging via the logging module (see app/logger.py).
• Public APIs should return JSON-friendly structures or Pydantic models.

Frontend

• Located under frontend/ using Vue 3 and Element Plus.
• Development: npm run dev (default port 3000).
• Build for production: npm run build.

Installation

1. Install Python dependencies:

   pip install -r requirements.txt

2. Download pretrained models into the model/ directory:
   • body_pose_model.pth
   • hand_pose_model.pth
3. (Optional) install front‑end dependencies:

   cd frontend && npm install

Running

• Backend: python -m uvicorn app.main:app --reload --host 0.0.0.0 --port 8001
• Frontend: npm run dev from frontend/.
• API documentation available at http://localhost:8001/docs.

Notes for Contributors

• Tests are not provided; manual verification via starting the server is recommended after changes.
• When adding new Python modules, place them within the appropriate package (app for API/service code, src for model logic).
• Keep dependencies minimal and update requirements.txt when new packages are required.
• Avoid committing large model files or result media; these paths are already listed in .gitignore.