Load_func module #11
Load_func module #11
Conversation
|
nice! will have a look |
| Returns: | ||
| str: the path to the file | ||
| """ | ||
| path = os.path.join(DATASETS_PATH, 'metadata', '2 - Dictionary properties - field_type.csv') |
There was a problem hiding this comment.
not my favorite file name, but I won't insist.
There was a problem hiding this comment.
I agree, but this is consistent with the original file
I hope this is temporary until we can store it in a different location, such as Athena or another type of storage service
pheno_utils/bulk_data_loader.py
Outdated
| function_name = FIELD_TYPE_TO_FUNC.get(field_type) | ||
| if function_name is not None: | ||
| return globals().get(function_name) | ||
| else: | ||
| # Handle the case where the field type is not found | ||
| return pd.read_parquet |
There was a problem hiding this comment.
what should we do if the function name is not in the globals? it will return a None, and there will be a downstream error, but perhaps we should catch it here first ("field type load func is missing..." of sorts)
minor style suggestion (unless you object), instead of the if-else clause you may use:
function_name = FIELD_TYPE_TO_FUNC.get(field_type, 'read_parquet')
and have from pandas import read_parquet at the top of the py.
There was a problem hiding this comment.
Thanks! added to the new commit
Please review :)
I have added a new module, which is supposed to contain all future 'show_fundus' functions, now called 'load_image'.
Now, you don't need to specify the specific function you want to use in the code (though you still can, for now). Instead, you need to add your field_type and the corresponding function it should use in the dictionary properties sheet, which will be updated in the general metadata directory.