Learn how to get involved in the A2ML project.


Documents are generated using sphinx. To generate new docs locally you will want to navigate to the /docs directory located in the root of this project.

There is currently an index.rst file which is the entry point. This file loads all other document files.

There are currently two top level directories with documentation.

  • /dev

  • /community

To edit the files in there directly using restructured text syntax or you can add files for new sections. Any new sections will need to have the path added in index.rst file.

Building Documentation

First, install documentation-related dependencies:

$ make develop-docs

Then, from inside the /docs directory run:

$ make html

This will create new files inside of the /build directory. If you notice that your changes aren't showing up try deleting all the contents inside build/ and running make html to force rebuild.

Viewing Documentation in the Browser

open ./build/html/index.html

Implementing A2ML for Another AutoML Provider

The A2ML Model class in A2ML.PY abstracts out the PREDIT (ITEDPR) pipeline.

Implementations are provided for

  • Auger.AI (Auger)

  • Azure AutoML (AZModel)

  • Google Cloud AutoML Tables (GCModel)

If you want to add support for another AutoML provider of your choice, implement a child class of Model as shown below (replacing each "pass" with your own code.

class AnotherAutoMLModel(Model):
    def __init__(self):
    def predict(self,filepath,score_threshold):
    def review(self):
    def evaluate(self):
    def deploy(self):
    def import_data(self):
    def train(self):