Feature/return df

[kevindougherty-noaa]

Addresses #75

[kevindougherty-noaa]

@CoryMartin-NOAA still need to go through the other files for pycodestyle to pass, however diags.py passes which is the only file I changed.

[CoryMartin-NOAA]

@kevindougherty-noaa can you explain to @ADCollard and me why this PR is needed? I'm sure it is a method vs attribute thing but it's not clear to me at the moment.

[kevindougherty-noaa]

@CoryMartin-NOAA @ADCollard Sure thing. In our regional DA meeting last week, Ting and Shun mentioned interest in being able to have the entire dataframe returned. I could be mistaken and maybe they meant returning the dataframe after being indexed.

[aerorahul]

See comment.

[aerorahul]

Why can't you add this method in Class GSIdiag since that is the class from which all other inherit from?

[CoryMartin-NOAA]

@kevindougherty-noaa yes I remember that, so I guess you're saying this is so one can have numpy arrays if they want, and a Pandas dataframe through the .return_df() method? I thought the issue you linked would basically imply that the read method would have an argument like read_data(return_df=True) and then:

if return_df:
    return self.df
else:
   return data

or whatever the above pseudo code should be

[kevindougherty-noaa]

@CoryMartin-NOAA you are right, that is what should be done. I will fix.

[CoryMartin-NOAA]

@kevindougherty-noaa is this indentation correct?

[CoryMartin-NOAA]

so for non data frame, it returns a dict of dicts, for u,v. For the data frame case, are both u and v inside the data frame?

[kevindougherty-noaa]

Yes, both u and v are in the dataframe. All the variables will be included.

[aerorahul]

Just a quick glance.
See if you notice these as well.

[kevindougherty-noaa]

Based on conversations with @CoryMartin-NOAA and @aerorahul, I am going to close this PR and develop diags.py to only return dataframes rather than specific variables. It should cut out a good amount of code.