Mostly written in Java, these functions build search indexes, compute dataset characteristics, generate tasks and evaluate the results of certain tasks.
The Java code is available in the 'OpenML' repository: https://github.com/openml/OpenML/tree/master/Java
General: - OpenML: Building Lucene search index and smaller tools, e.g. extracting documentation from WEKA source files and ARFF files - generateApiDocs: Generates API HTML Documentation - http_post_file: Example how to post files to the api using Java.
Support for tasks: - foldgeneration: Java code for generating cross-validation folds. Can be used from command line. - splitgeneration: Split generator for cross validation and holdout. Unsure what's the difference with the previous? - generate_predictions: Helper class to build prediction files based on WEKA output. Move to WEKA repository? - evaluate_predictions: The evaluation engine computing evaluation scores based on submitted predictions