We are already able to collect coverage data for subprocesses (we collect while doing make import_local_datahttps://github.com/commonsearch/cosr-back/blob/master/Makefile#L144), but pyspark seems to copy the job file in a temporary folder, which breaks coverage collection, probably because the filenames change and/or the pytest-cov.pth file is not executed?
Currently our test coverage score is lower than it should be.
We are not collecting coverage information for the code running under
spark-submit
: https://github.com/commonsearch/cosr-back/blob/master/tests/sparktests/test_index.py#L79We are already able to collect coverage data for subprocesses (we collect while doing
make import_local_data
https://github.com/commonsearch/cosr-back/blob/master/Makefile#L144), but pyspark seems to copy the job file in a temporary folder, which breaks coverage collection, probably because the filenames change and/or the pytest-cov.pth file is not executed?