A simple continuous integration server written in Python.
- Highly configurable
- REST API compatible with OpenAPI
- Per-repo fine tuning via .murdock.yml
- Live status of CI jobs via websockets
- Can comment results on pull-request
And many more not listed here.
Murdock is fully written in Python (requires >= 3.9) and is based on the FastAPI framework for the REST API. In order to keep the history of completed jobs, they are stored in a MongoDB database.
We recommend that you use docker-compose to deploy Murdock on a server.
Murdock uses Webhook to interact with GitHub, so we recommend that you read the Github WebHook documentation.
A default Docker based deployment can be performed using make:
Build the React application:
$ make
This command will copy the provided .env.example file to a default .env so that it contains good defaults, then clone the murdock-html repository, build the frontend web application and finally launch the Docker services (mongo database, web frontend and Murdock API server). Once deployed, by default, the UI is available at http://localhost:8000.
For local development, we recommend that you use a Python virtualenv.
Install all Murdock dependencies:
$ python3 -m pip install -r requirements.txt
Launch the MongoDB service:
$ docker-compose up mongo-dev
Launch the uvicorn server with the --reload option:
$ uvicorn murdock.main:app --reload --reload-dir murdock
The murdock application can be configured by setting environment variables:
- to the command line
- in a .env file (see .env.example for a settings overview)
- in a
.env.localand addENV_FILE=.env.localto the command line. Again, see the .env.example as example.
The .env files can also be used as-is when deploying with docker-compose using
the --env-file <.env filename> option.
To get a complete list of the configuration options, check the Murdock config code.
Unittests are available and performed using pytest. The full test suite can be launched using Tox:
Install Tox:
$ python3 -m pip install tox
Run the tests:
$ tox