Wait for batch process API job to update status from PROCESSING to something else

[mlubej]

A simple addition with more descriptions.

Another option would be to run the while loop as long as the status is PROCESSING, but perhaps the user would be confused why the job doesn't move forward as soon as the progress bars hit 100%. An additional reason is that the suggested implementation is perhaps more informative

[mlubej]

@zigaLuksic in the monitor_batch_job we have the monitor_batch_analysis function which waits until the batch analysis is finished and then moves on. Then we put the responsibility of starting the job on the user, but if the user doesn't start the batch job, the loop will run forever.

Wouldn't it make more sense if the monitor_batch_job finishes early if the last user action was not START, and if it was, it waits for the PROCESSING status?

Asking because some test cases don't make much sense

[zigaLuksic]

@zigaLuksic in the monitor_batch_job we have the monitor_batch_analysis function which waits until the batch analysis is finished and then moves on. Then we put the responsibility of starting the job on the user, but if the user doesn't start the batch job, the loop will run forever.

Wouldn't it make more sense if the monitor_batch_job finishes early if the last user action was not START, and if it was, it waits for the PROCESSING status?

Asking because some test cases don't make much sense

Because monitor_batch_job is meant to monitor a started batch job. I'm not sure we want to think about all the possible cases of it being used wrong, because there are too many. Or do you think it's crystal clear?

[mlubej]

Fixed the tests by additionally mocking the batch_client.get_request calls and updating the number of expected calls/loops

I'm not sure we want to think about all the possible cases of it being used wrong, because there are too many. Or do you think it's crystal clear?

It could very well be that I have a focus on a specific and this wouldn't hold in general, so I'm fine with leaving it like it is, I just wanted to point out that the tests simulate cases which don't seem to be realistic. I agree that it's hard to draw the line between the functionality and what is expected from the user. I guess that so far this has not been a problem, so let's leave it as it is.