Note: This tool is a work-in-progress, and may not work for all ML API python repos.

Currently, python directory has test and sample which fails one out of ten times randomly to produce the flakiness.

You will need to set up the local development environment by [installing the Google Cloud SDK][cloud-sdk] and running the following commands in command line: gcloud auth login and gcloud config set project [YOUR PROJECT ID].

Install python dependencies

pip3 install -r requirements.txt

Python >= 3.6

1. pip3 install pytest
2. pip3 install -r requirements.txt (if it is ML APIs repo)

• make bash script file executable

  chmod +x flaky_report_tool.sh
  

• execute the script in context to make available for usage

  source flaky_report_tool.sh 
  

• call bash function with flaky_tester python_test_file_test.py output_file_path.txt

1. Python test file path (required)

2. Output text file (optional)

   source flaky_report_tool.sh python/fake_quickstart_test.py output_text.txt