OpenML Python Evaluation Engine

Python port of the OpenML Evaluation Engine

Installation

Preferably you want to setup a virtual environment first to prevent the package from being installed to your global python installation. You can install the package with the CLI interface by using the provided setuptools.

python setup.py install

After installation, the scripts can be ran with the following command: pyevaluationengine. For more specific information about parameters, add the -h flag. The entrypoint for scripts is configured to be the cli.py file.

To see more specific instructions on how to install this package as a development dependency, we recommend looking the the CONTRIBURING file and follow some of the steps under the "Code Contributions" header.

Usage

The CLI has the following modes:

config

Used to set the API key and URL. This command needs to be run before you can use any of the other scripts.

all

Processes and analyzes all of the unprocessed datasets once.

print

Prints all of the unprocessed datasets to the terminal.

singular

Processes the specified dataset by name.

amount

Processes a specified amount of datasets

repeat

Processes all unprocessed datasets and repeats this after a specified timeout

Further information

• OpenML documentation
• OpenML client APIs
• OpenML developer guide
• Contact information
• Citation request
• OpenML blog
• OpenML twitter account