BasicVids Storage

Video storage microservice for BasicVids.

The service stores uploaded videos on local disk, tracks metadata in the database, and offloads transcoding and HLS generation to Celery workers.

Stack

• Gunicorn
• FastAPI
• SQLModel
• Redis
• Celery
• ffmpeg / ffprobe

Development

Use a virtual environment:

virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

Run tests:

pytest

Run locally:

uvicorn basicvids_storage.main:app --reload

Container

mkdir -p data
cp .env.example data/.env
docker compose up -d --build

The service is available through the shared gateway at:

http://localhost:8080/api/v1/videos/
http://localhost:8080/api/v1/categories/

Configuration

Project environment is loaded from:

./data/.env

Start from:

./.env.example

Database examples:

# SQLite default
# DATABASE_URL=sqlite:///./data/database.db

# PostgreSQL example
DATABASE_URL=postgresql://basicvids_storage_user:change_me@host.docker.internal:5432/basicvids_storage

Important variables:

Variable	Default	Description
DATA_PATH	./data	Data directory mounted in container
DATABASE_URL	sqlite:///./data/database.db	Metadata database URL
REDIS_URL	redis://localhost:6379/2	Redis connection for rate limiting and app state
CELERY_BROKER_URL	redis://basicvids_redis:6379/0	Celery broker
CELERY_RESULT_BACKEND	redis://basicvids_redis:6379/1	Celery result backend
AUTH_CURRENT_USER_URL	http://basicvids_auth:8000/api/v1/users/detail/	Auth service current-user endpoint
VIDEO_TRANSCODE_WORKER_CONCURRENCY	1	Celery worker concurrency
VIDEO_TRANSCODE_THREADS	2	ffmpeg thread count per transcode

Runtime Notes

Local development requires more than Python packages:

• Redis
• Celery worker
• ffmpeg
• ffprobe

The docker compose file starts both API and worker containers.

Healthcheck

http://localhost:8080/storage/health