Changed arg scale to data_scale in run_deepcell_task BatchProcessingJob #328
Conversation
|
What happens if you pass a string instead of a float? |
|
It gives the correct output if you input a string. In other words, if you pass something like |
There was a problem hiding this comment.
Okay, in that case, can you add a step within create_deepcell_output that attempts to cast scale to a float if it isn't already one, and raises an informative value error if the input can't be cast to a float? And add a check to the test function that the appropriate value error is raised. You can take a look at some of our other functions where we do input validation and pytest.raises() checks in the test functions as an example
…cell_service_utils, added a corresponding test in deepcell_service_utils_test
|
I added a try/except clause to check whether or not the scale arg can be cast in |
…ob (#328) * changed arg scale to data_scale in run_deepcell_task BatchProcessingJobManager call * added a check to see that scale arg can be converted to float in deepcell_service_utils, added a corresponding test in deepcell_service_utils_test * style fixes * second round of style changes
…ob (#328) * changed arg scale to data_scale in run_deepcell_task BatchProcessingJobManager call * added a check to see that scale arg can be converted to float in deepcell_service_utils, added a corresponding test in deepcell_service_utils_test * style fixes * second round of style changes
This PR will fix and close #327 . The deepcell ```BatchProcessingJobManager```` is expecting the keyword "data_scale" and not "scale". It is also expecting a float and not a string. This PR changes both of these issues.
How did you implement your changes
I changed the argument name scale to data_scale in the
run_deepcell_taskmethod inutils.deepcell_service_utils.py. I tested the changes on two different images by changing and comparing the scale to 0.5, 1 and 2.0. Not sure why I saw differences when I originally added this functionality -- when I re-ran the same code I got different results (in other words, the deepcell output wasn't changing even if I changed the scale).