I'm going to raise "#question" issues as I continue to read through the code. Many of these are just documenting minor observations during the read through that are likely already on someone's todo list if there ever was a clean up. They are not necessarily important or require immediate action.
It would be useful to extract data-independent visualization routines into their own file. For example, in https://github.com/scanner-research/esper/blob/master/app/esper/identity.py, the function tile_imgs is just a helper function to make an image montage. I'm sure something like this is used in many situations in Esper. This is likely better off somewhere else for reuse, e.g., we could consider app/esper/image_util.py for symmetry with spark_util.py.
I'm going to raise "#question" issues as I continue to read through the code. Many of these are just documenting minor observations during the read through that are likely already on someone's todo list if there ever was a clean up. They are not necessarily important or require immediate action.
It would be useful to extract data-independent visualization routines into their own file. For example, in https://github.com/scanner-research/esper/blob/master/app/esper/identity.py, the function
tile_imgs
is just a helper function to make an image montage. I'm sure something like this is used in many situations in Esper. This is likely better off somewhere else for reuse, e.g., we could consider app/esper/image_util.py for symmetry with spark_util.py.