Every commit is a crucial decision point

"What's the probability there is a bug in my code?" - SoLittleCode's Pull Request tool, ReviewML makes an unbiased empirical prediction based on an analysis of historic bugs in your repo, and recommends the reviewers most likely to reduce that risk.

Installation

Once you install the app on Github and select the repos you would like to analyse, a container will be created that clones your repo on which a Deep Learning Model is trained. Once the training is complete (usually about 15 minutes) the container and any copy of your code is removed keeping everything secure. The status of your model can be viewed on a dashboard, where the model can be trained as often as you like.

Pull Request Process

When a PR is raised on your repo, ReviewML will pass your change into the model and make a prediction on the probability of the code containing a bug. It will then look at the other contributors in your repo and make a prediction on the probability that there would be a bug had they submitted similar code. If you are concerned the probability is too high ask one of them for a review and help reduce the risk of creating a costly bug.

Please help us to make it better

ReviewML is currently in Beta and will stay free for those trying the service out now. We continue to develop and improve the tool so please send any issues or suggestions you have to tom@solittlecode.com