Currently, the test only tests the paths used for an RNA-seq run and a CITE-seq run. We should add more test cases by adding new entries to tests/stub-metadata.tsv along with the input files to support those runs (these can generally just be blank files created with touch)
Some of the data types we should add:
bulk RNA-seq
spatial (this may require setting a docker image for Space Ranger in config/profile_stub.config, but that image does not need to contain Space Ranger)
cellhashed with genetic demultiplexing
cellhashed without genetic demultiplexing
We also may want to consider whether we need to test any other particular parameter flags that we set. At the moment, the only one I can think of would be --skip-genetic-demux, but we would really only be testing that it ran successfully unless we added some extra logic.
We could also add some logic to test that the expected output files/directories are generated after each run, but that gets more complicated as we add more inputs, so I'm not sold on it.
In https://github.com/AlexsLemonade/scpca-nf/pull/332, I added a stub workflow that can automatically test that workflow logic is sound without requiring a full workflow run.
Currently, the test only tests the paths used for an RNA-seq run and a CITE-seq run. We should add more test cases by adding new entries to
tests/stub-metadata.tsv
along with the input files to support those runs (these can generally just be blank files created withtouch
)Some of the data types we should add:
config/profile_stub.config
, but that image does not need to contain Space Ranger)We also may want to consider whether we need to test any other particular parameter flags that we set. At the moment, the only one I can think of would be
--skip-genetic-demux
, but we would really only be testing that it ran successfully unless we added some extra logic.We could also add some logic to test that the expected output files/directories are generated after each run, but that gets more complicated as we add more inputs, so I'm not sold on it.