Behave testing harness for RTX

As part of the larger translator-testing-framework, we've created a testing harness for RTX using Behave, a Python-based behavior-driven development testing framework. These tests ask RTX questions through the RTX OpenAPI (which currently uses KG1), and compare its answers to ground truths.

To get set up

First, create and start up a Python virtual environment:

python3 -m venv behave-env

source behave-env/bin/activate

Then clone the RTX fork of translator-testing-framework:

git clone --recursive https://github.com/RTXteam/translator-testing-framework.git

Then navigate into your newly created directory and sync the fork to the original repository:

cd translator-testing-framework

git remote add upstream https://github.com/NCATS-Tangerine/translator-testing-framework.git

git pull upstream master

And run this to install packages:

pip install -r requirements.txt

To run the tests

Use this command to run all of the RTX tests:

behave -i rtx-tests.feature

Use behave -n "[scenario name]" to run only a particular test, e.g.:

behave -n "Fanconi anemia is associated with expected genes"

NOTE: In addition to RTX, translator-testing-framework contains tests for various other Translator projects; you can run all of the tests (across all projects) using the command behave, but beware that other projects' tests may have further dependencies not detailed here, and thus may not run properly.

Contributing code

The going protocol for contributing is to push changes to our RTX fork, and then make pull requests from our fork to the original repository.

Currently, all of the RTX-specific tests live in one feature file: rtx-tests.feature. The Python implementations of the steps these tests use are currently in the main steps.py file (along with various other projects' step implementations).