🚀 Front-End Project

A modern front-end setup using React, Vite, TypeScript, and Tailwind CSS.

---

✅ Requirements

Before you begin, ensure you have the following installed:

Node.js (v16+ recommended)
Check with: node -v npm (v8+ recommended)
Check with: npm -v

---

📦 Installation

Clone the repository and install dependencies:

bash git clone https://github.com/your-username/your-repo.git cd frontend npm install

⸻

▶️ Start the App

Run the development server with:

npm start

Backend

Panamia API

A FastAPI-based API for automated surveillance video analysis, summary generation, and smart alert management.

---

🚀 Main Features

• Upload and full analysis of videos (frame extraction, summary, contextual analysis…)
• Trigger a decision agent to automatically contact the appropriate emergency services (Gendarmerie, SAMU, Firefighters, Police) via SMS (Twilio)
• Modular architecture (video, audio, contextual analysis, decision agent)
• Easy deployment with Docker & Docker Compose
• Security: API keys managed via environment variables

---

📁 Project Structure

.
├── API/
│   └── api.py                # FastAPI entry point (main routes)
├── Modules/
│   ├── analyse_video.py      # Video frame analysis
│   ├── analyse_audio.py      # (Optional) Audio transcription
│   ├── analyse_contextuelle.py # Contextual report generation
│   ├── agent_decisionel.py   # Decision agent (alerts, LLM, Twilio)
│   ├── run_analyses.py       # Orchestrates the analysis pipeline
│   └── alert.py              # SMS alert sending via Twilio
├── data/                     # Output folder for analysis results
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
├── deploy.sh                 # Automated install & deployment script
├── test_agent.py             # Test script for the decision agent
└── .gitignore

---

🛠️ Installation & Launch

1. Clone the repository

git clone <repo_url>
cd <repo_name>

2. Configure your environment variables

• Create a .env file at the root of the project and fill it with your API keys and phone numbers (see env_example.txt for reference).

3. Start with Docker Compose

sudo docker-compose up --build -d

The API will be available at:
http://<SERVER_IP>:8000

---

🧩 API Usage

1. POST /api/v1/envoyer-flux

• Description: Upload a video, run the full analysis pipeline, and generate analysis files.
• Body: multipart/form-data with a video field (video file)
• Response:

  {
    "message": "Analyse vidéo terminée avec succès",
    "fichiers_generes": [
      "data/descriptions_par_image.txt",
      "data/resume_video.txt",
      "data/transcription_audio.txt",
      "data/analyse_contextuelle.txt",
      "data/metadata.json"
    ],
    "taille_analyse": 1234
  }

2. POST /api/v1/gerer-alerte

• Description: Triggers the decision agent based on the video summary.
  The agent analyzes the summary, decides which services to contact, and sends alerts via SMS.
• Response:

  {
    "message": "Agent décisionnel exécuté",
    "taille_resume": 456,
    "sortie_agent": "Appel des services : ['Pompiers']\n[Pompiers]  Message envoyé avec succès."
  }

Interactive documentation

Access the interactive docs at:
http://<SERVER_IP>:8000/docs

---

⚙️ Main Dependencies

• Python 3.9
• FastAPI, Uvicorn
• OpenCV, numpy
• Langchain, Groq, LangGraph
• Twilio (SMS sending)
• python-dotenv

See requirements.txt for the full list.

---

🔒 Security & Best Practices

• Never commit your .env file (already in .gitignore)
• Clean your Git history if secrets were ever committed by mistake
• Restrict access to port 8000 (firewall already configured)

---

👨‍💻 Development & Testing

• To run locally without Docker:

  pip install -r requirements.txt
  uvicorn API.api:app --reload

• To test modules individually, see scripts in Modules/ and test_agent.py.

---

🆘 Support

For questions or bug reports, open an issue on GitHub or contact the author.

---

Project developed by Yassine Seidou.
License: MIT (or adapt as needed).

Let me know if you want to add or change any section!