Add ability to fix some data quality issues.

[danielbeach]

Add the ability to "fix" some data quality issues and return a fixed dataframe. Example, removing duplicates or removing whitespace etc.

[tsterbak]

Sounds like a great idea!
How should this be implemented? For example as a get_fixed_data method with flags to indicate which fixes to apply. Or as a method for each type of fix?

[danielbeach]

@tsterbak interesting question, makes me wonder if the default_checks should really all be wrapped up into a single method, or if we should just provide direct access to the individual methods, with a flag in those methods for if you want the fixed dataframe back?

[jcorrado76]

Is "fixing" the DataFrame within the scope of this package? If this is a package around running validations, actually editing the source data seems a little like scope creep. Not to mention that the logic for how you do that might be really complicated, and Tiny Timmy might not necessarily want to own that beyond exposing some functionality for common imputations, or dropping nulls with some of the common methods. Maybe consider allowing users to run cleaning methods by leveraging the underlying compute platform. i.e. if there's a "Tiny Timmy frame", and it inherits from a Polars Lazy Frame but just defines new methods or a new namespace on it, then users can call whatever Polars methods they want while also having access to validation methods from Tiny Timmy. Great expectations does this too - you have a great expectations frame, which is a wrapper around a pandas dataframe, except it also defines "expect_this", "expect_that" on those frames. This sounds like a larger architectural question, so I defer to @danielbeach on where to take it.