Home

The Issue Labeler comprises several GitHub workflows and backing applications that orchestrate the process of training a repository's model, testing the model against existing data, and automatically performing predictions.

Downloader

Download issue and pull request data from GitHub, creating tab-separated (.tsv) data files to be consumed by the Trainer.

Trainer

Load the tab-separated issue and pull request data that has already been downloaded, and train an ML.NET model over the data to prepare for making label predictions.

Tester

Perform a comparison test run over GitHub data, predicting labels and comparing the predictions against the actual values. This can be performed either by downloading issue and pull request data from GitHub or loading a tab-separated (.tsv) file created by the Downloader.

Predictor

Consume the ML.NET model and make predictions for issues and pull requests.

Cache Retention

Forces cache restores of the Predictor app and the prediction models, preventing cache eviction that would disrupt predictions.

Build Predictor

Builds the Predictor app and caches it. This can be run manually if the Predictor does get evicted from cache and prediction jobs begin failing.