Installation

Radish is available on PyPi as radish-run and can be installed on most systems with pip:

$ pip install radish-run

How to use

An example invocation of radish:

$ radish command tests --from 19abc023 --to 2514ecb1
Changed paths:
  - extensions/cool-extension/
  - frontend/js/

Running tests for extensions/cool-extension/:
...........
OK

Running tests for frontend/js/:
..........................
OK

All commands ended successfully and ran in 9.75s.

Configuration

radish configuration is a yaml file named Radishfile, because I can.

paths:
  - extensions/*/  # Mark each subdirectory in extensions as a path
  - frontend/js/

commands:  # Runs from the directory denoted by paths above
  tests:
    default: bin/rspec spec
    frontend/js/: npm test

An example use case

Take that you’re building a single page web app, it consists of two parts: - The backend that delivers JSON - The frontend that holds all the clicky bits that end-users see

Because cross-functional teams this project is in one repository, which is a great win for productivity. But it has a downside: when there are only changes to the backend, then all the tests for the frontend is still run. Finally, after running all tests it can get deployed, but then both sites get deployed despite nothing changing on the frontend.

So this is a crazy situation. It shouldn’t be. So this is where radish comes in. You tell radish about test and deploy, and what that means for both the backend and the frontend, then on your CI server you run the commands with the last green commit. Radish then figures out what has changed from the current commit and the last one, and only runs the command for those projects.

Contributing

$ git clone https://github.com/gaqzi/radish.git
$ cd radish/
$ make develop
$ make test

License

Beerware License