deepfake-detector/ │ ├── app/ │ ├── main.py # FastAPI entrypoint │ ├── config.py # Global settings / env vars │ ├── celery_config.py # Celery broker and backend config │ ├── tasks/ # Celery tasks (frame/model/LLM) │ │ ├── init.py │ │ ├── frame_processing.py # Optical flow, scene change │ │ ├── detection.py # FakeShield inference │ │ ├── llm_explanation.py # Explanation generation │ │ └── postprocess.py # Heatmaps, overlays, formatting │ │ │ ├── api/ │ │ ├── init.py │ │ ├── auth/ # Auth API logic │ │ │ ├── routes.py │ │ │ ├── schemas.py │ │ │ └── security.py │ │ ├── detection/ # Detection endpoint logic │ │ │ ├── routes.py │ │ │ ├── schemas.py │ │ │ └── utils.py │ │ └── response/ # Result formatting or downloads │ │ ├── routes.py │ │ └── utils.py │ │ │ ├── core/ │ │ ├── database.py # DB connection & session │ │ ├── models.py # SQLAlchemy models │ │ └── schemas.py # Shared Pydantic schemas │ │ │ ├── services/ │ │ ├── inference.py # Load FakeShield + model utils │ │ ├── extractor.py # Load video, extract frames │ │ ├── llm_client.py # Langchain / HuggingFace LLM interface │ │ └── media_utils.py # Encoding masks, base64 images │ ├── worker.py # Celery worker startup ├── Dockerfile # Backend Docker image ├── docker-compose.yml # Services (FastAPI, Redis, Postgres, etc.) ├── requirements.txt # Python dependencies ├── .env # Environment variables ├── README.md # Project overview └── frontend/ # (Optional) React / HTML frontend └── ...