Add include_path_column arg to read_json

[bryanwweber]

As suggested in #8539, this adds the include_path_column argument to dd.read_json(). The paths can be munged using the path_converter argument, which is slightly different from the dd.read_csv() case, because pd.read_json() does not support the converters argument.

• Closes Add include_path_column to read_json #8539
• Tests added / passed
• Passes pre-commit run --all-files

[GPUtester]

Can one of the admins verify this patch?

[jrbourbeau]

add to allowlist

[jrbourbeau]

Thanks @bryanwweber! Looking forward to seeing this merged

On Windows, some of the tests added here are failing with

 E               AssertionError: Attributes of DataFrame.iloc[:, 2] (column name="path") are different
E               
E               Attribute "dtype" are different
E               [left]:  CategoricalDtype(categories=['C:/Users/RUNNER~1/AppData/Local/Temp/tmpupda7hgh.json'], ordered=False)
E               [right]: CategoricalDtype(categories=['C:\Users\RUNNER~1\AppData\Local\Temp\tmpupda7hgh.json'], ordered=False)

The other failures are unrelated to the changes in this PR (xref #8580, #7406)

[bryanwweber]

On Windows, some of the tests added here are failing with

 E               AssertionError: Attributes of DataFrame.iloc[:, 2] (column name="path") are different
E               
E               Attribute "dtype" are different
E               [left]:  CategoricalDtype(categories=['C:/Users/RUNNER~1/AppData/Local/Temp/tmpupda7hgh.json'], ordered=False)
E               [right]: CategoricalDtype(categories=['C:\Users\RUNNER~1\AppData\Local\Temp\tmpupda7hgh.json'], ordered=False)

Ugh, Windows paths. I'll look at how read_csv() deals with this issue, I don't recall seeing normpath anywhere.

[jrbourbeau]

Thanks @bryanwweber!

[jrbourbeau]

It looks like this needs the same path formatting you applied to test_read_json_with_path_column

@martindurant do you have any insight into why pandas.to_json and dask.dataframe.read_json are using different path separators on windows (i.e. pandas outputs files with \ and Dask reads them in with a /)?

[jrbourbeau]

Thanks @bryanwweber! This is in