Python Runner

Docker container that clones a GitHub repository containing a Python web application, installs dependencies, and runs the application.

Usage

From GitHub

docker run --rm \
  -e GITHUB_URL=https://github.com/<org>/<repo>/ \
  -e GITHUB_TOKEN=<token> \
  -p 8080:8080 \
  eyevinn/python-runner

From S3 / MinIO

First, zip your project directory:

cd my-python-app && zip -r ../my-app.zip .

Upload the zip file to your S3 bucket, then run:

docker run --rm \
  -e SOURCE_URL=s3://bucket/my-app.zip \
  -e S3_ENDPOINT_URL=http://minio:9000 \
  -e AWS_ACCESS_KEY_ID=<access-key> \
  -e AWS_SECRET_ACCESS_KEY=<secret-key> \
  -p 8080:8080 \
  eyevinn/python-runner

Environment Variables

Variable	Description
GITHUB_URL	GitHub repository URL (e.g., https://github.com/org/repo/)
GITHUB_TOKEN	GitHub personal access token for private repositories
SOURCE_URL	S3 URL to a zipped application (e.g., s3://bucket/app.zip)
S3_ENDPOINT_URL	Custom S3 endpoint URL (for MinIO or other S3-compatible services)
AWS_ACCESS_KEY_ID	AWS/S3 access key ID
AWS_SECRET_ACCESS_KEY	AWS/S3 secret access key
OSC_ACCESS_TOKEN	Access token for Eyevinn Open Source Cloud config service
CONFIG_SVC	URL to config service endpoint
PORT	Port to run the application on (default: 8080)

Supported Project Structures

The runner automatically detects and installs dependencies from:

• pyproject.toml - Modern Python packaging (PEP 517/518)
• requirements.txt - Traditional pip requirements file
• setup.py - Legacy setuptools configuration

If a setup.sh script is present in the repository root, it will be executed after dependency installation.

Automatic Framework Detection

The runner automatically detects your application type and starts it with the appropriate server:

Framework	Detection	Server Used
FastAPI/Starlette	fastapi or starlette in dependencies	uvicorn
Flask	flask in dependencies	gunicorn (if available) or flask run
Gunicorn	gunicorn in dependencies + config file	gunicorn with config
Plain Python	main.py or app.py present	python main.py

The runner looks for the app object in common entry point files (main.py, app.py, application.py, server.py, api.py) and detects common patterns like app = FastAPI() or app = Flask(__name__).

Custom Command

You can override the auto-detection by passing a custom command:

docker run --rm \
  -e GITHUB_URL=https://github.com/<org>/<repo>/ \
  -p 8080:8080 \
  eyevinn/python-runner \
  python app.py

Or for a specific Flask configuration:

docker run --rm \
  -e GITHUB_URL=https://github.com/<org>/<repo>/ \
  -p 8080:8080 \
  eyevinn/python-runner \
  flask run --host=0.0.0.0 --port=8080

Or for Gunicorn with specific options:

docker run --rm \
  -e GITHUB_URL=https://github.com/<org>/<repo>/ \
  -p 8080:8080 \
  eyevinn/python-runner \
  gunicorn -w 4 -b 0.0.0.0:8080 main:app

Building

docker build -t eyevinn/python-runner .

License

This project is licensed under the MIT License - see the LICENSE file for details.

About Eyevinn Technology

Eyevinn Technology is an independent consultant firm specialized in video and streaming. We assist companies with their streaming and cloud strategies, including architecture design and implementation.

In addition to consulting services, we offer cloud-based products like the Eyevinn Open Source Cloud platform - a fully managed solution to launch open-source products in AWS infrastructure.

Contact

• Web: https://www.eyevinntechnology.se
• Community: Join our Slack community here