Run locally with Yadage

Sometimes it can be convenient to run the workflows locally with Yadage, for debugging purposes. To do this you do not need REANA, only Yadage, serving as the execution engine.

Yadage defines the workflow specification language, which the MadMiner workflow is written in, but also includes a Command Line Interface (CLI) tool that coordinates the execution of workflows locally.

Install Yadage

To install Yadage, simply run:

pip3 install yadage

Launch Docker

If you are running the workflow locally with Yadage, then you will also need to have Docker installed, as Docker is used to pull the necessary Docker images from DockerHub.

Run locally

All the combined workflow, as well as the Physics and Machine Learning sub-workflow, have a Makefile that defines runnable high-level commands for convenience. The main command that you would use in this case is called yadage-run.

make yadage-run

This command would run the specified workflow within the workflow folder, generating the results within the .yadage folder (an invisible folder only accessible via terminal or IDE).

cd .yadage
ls -l