Is your feature request related to a problem? Please describe.
The current Pixie notebook tests assume linear workflows. In reality, users may run cells out of order, or begin at certain cells in certain states. We should think about what the common use-case scenarios are and explicitly test these.
Describe the solution you'd like
Pretty similar to the existing workflows, except we can simulate running the cells out of order in a separate test class.
Some common scenarios could be:
Restarting from a particular step, with and without some FOV data already generated for that process
Redoing the metacluster remapping
Running a step multiple times with an explicit overwrite flag.
Is your feature request related to a problem? Please describe.
The current Pixie notebook tests assume linear workflows. In reality, users may run cells out of order, or begin at certain cells in certain states. We should think about what the common use-case scenarios are and explicitly test these.
Describe the solution you'd like
Pretty similar to the existing workflows, except we can simulate running the cells out of order in a separate test class.
Some common scenarios could be:
overwrite
flag.