With updates to the science behind the workflow, we need to update our tests to use this.
We have also discovered some areas we would like to cover with end to end tests.
As a result the end to end tests have had the following updates
Test | Action taken by the test | Intended result of the test | How to use |
---|---|---|---|
Automated submission wrapper restart | The submission wrapper used by the end to end test module is restarted multiple times during the test | Ensures that even if the submission wrapper is restarted a cybershake run can continue | Use the argument --test_restart when running any cybershake end 2 end test |
Database stress test | A cybershake consisting of 30000 realisations is simulated (Nothing is run using slurm) and 3000 tasks are marked as progressing each cycle | Ensures that the sqlite database is capable to performing under a heavy workload | Use the argmuent --test_queue and supply the stress test config. This is not compatible with the --test_restart argument |
Cybershake v18p9 run | Approximately 1000 realisations are run, ranging from 500 realisations of Hossack to 1 realisation of AlpineF2K | A large cybershake run to ensure that if significant changes are made to the code, the science comes out the same | Use the min cs 18p9 config file |
Version 18p5p3p4 run | Update the simple test configuration to use the 18p5p3p4 version | The new version uses an updated VS30 model | Use the simple 18p5p3p4p config file |