Running unit tests

To run the unit tests, run the run-tests.sh script. You will need to set these environment variables:

  • SPARK_HOME: your local copy of Apache Spark. Look at .travis.yml and download_spark_dependencies.sh for details on how to download Apache Spark.
  • HYPEROPT_FMIN_SEED: the random seed. You need to get its value from .travis.yml.

For example:

hyperopt$ HYPEROPT_FMIN_SEED=3 SPARK_HOME=/usr/local/lib/spark-2.4.4-bin-hadoop2.7 ./run_tests.sh

To run the unit test for one file, you can add the file name as the parameter, e.g:

hyperopt$ HYPEROPT_FMIN_SEED=3 SPARK_HOME=/usr/local/lib/spark-2.4.4-bin-hadoop2.7 ./run_tests.sh hyperopt/tests/test_spark.py

To run all unit tests except test_spark.py, add the --no-spark flag, e.g:

hyperopt$ HYPEROPT_FMIN_SEED=3 ./run_tests.sh --no-spark

To run the unit test for one file other than test_spark.py, add the file name as the parameter after the --no-spark flag, e.g:

hyperopt$ HYPEROPT_FMIN_SEED=3 ./run_tests.sh --no-spark test_base.py