Update standardizer.py #944
Conversation
Adding a new class called ignore_errors to enable the standardizers to gracefully continue when encountering null key:value pairs
jonfritz
requested review from
mdwelsh,
mashah,
bsowell,
eric-anderson and
HenryL27
| @@ -283,3 +283,19 @@ def __init__( | |||
| **kwargs, | |||
| ): | |||
| super().__init__(child, f=standardizer.standardize, args=path, kwargs=kwargs) | |||
|
|
|||
| class IgnoreErrors(standardizer) | |||
There was a problem hiding this comment.
This isn't quite the right syntax, so I don't think this will work.
@bsowell My guess is that we would like to have this logic be more generic, perhaps a subclass of Map that swallows errors from the function being applied?
There was a problem hiding this comment.
It might be worth just making this a function rather than a full class - the main benefit you get from having classes with single static methods is that you can do interfaces and stuff. I don't imagine there will be any other things that follow the same contract as IgnoreErrors so it should probably just be a function (useable like docset.map(lambda d: ignore_errors(d, standardizer, key_path))).
If you want to be really fancy (and actually I think this would be a good interface) you can make this a higher-order function - i.e. have it return a function that calls the standardizer and swallows errors. Then the usage can become docset.map(ignore_errors(standardizer, key_path)).
Update with Henry's feedback
Adding unit tests for ignore_errors
Update helper text
Add types to function
Shorten line in imports
Reduce line length from helper text
Adding a new class called ignore_errors to enable the standardizers to gracefully continue when encountering null key:value pairs