Bug: common_base not found on Windows

[hectormz]

First of all, I really like the intent of this package! 👍

Is this a BUG REPORT or FEATURE REQUEST?:

• bug
• feature

Description of Bug or Feature

Environment

• interrogate version: 1.14
• Operating System(s): Windows 10
• Python version(s): 3.8

What happened

utils.get_common_base() does not find the common base for files on Windows.

This happens because path levels are split by \\ in Python on Windows. As a result, the common_base is an empty string (""). This has a downstream effect when generating the summary tables.

Specifically, in _get_filename(), since the common_base is an empty string, a filepath:

C:\path\to\project_root\src\file.py

becomes
:\path\to\project_root\src\file.py

instead of:
\src\file.py

Suggestion

Since this is a python 3.5+ project, I think that this could be fixed in a flexible way by handling filepaths with pathlib instead.
For example this in utils.py:

level_slices = zip(*[f.split("/") for f in files])

could be replaced with:

from pathlib import Path
level_slices = zip(*[Path(f).parts for f in files])

(if file names had not already been converted to Path objects)

Using pathlib here would help make this feature work across systems, but I think it could also replace just about all path manipulation interrogate uses os for.

If you'd like I could work contributing to any of these proposed changes.

[econchick]

Hey @hectormz ! Thanks so much for such a thorough bug report, it's very helpful. I'd welcome any contribution :) LMK if I can help at all.

[hectormz]

@econchick sure thing. I'll make a PR and see what you think!

Another thing that I found was that I had attrs version 19.1.0 installed, which causes an error when using interrogate. The error goes away if using attrs>=19.2.0.

I can include that version requirement in setup.py during this PR unless you'd like a separate one.

[econchick]

Ah good catch - yea feel free to through that in the same PR (separate commit if possible though). I appreciate it!

[econchick]

So instead of fixing all the paths, I tried to keep the bug fix very narrow. Unfortunately I don't have an easy way to test on a windows machine. Would you mind checking out #18 for me? In the mean time, I'm going to try and get github actions to work w windows, but it's not as easy as I thought...

[econchick]

@hectormz aha! I got windows testing to work on github! and found lots of test failures :D I am working through them :)

[hectormz]

So instead of fixing all the paths, I tried to keep the bug fix very narrow. Unfortunately I don't have an easy way to test on a windows machine. Would you mind checking out #18 for me? In the mean time, I'm going to try and get github actions to work w windows, but it's not as easy as I thought...

Yes, let me get to my computer and test this out for you.

[hectormz]

@hectormz aha! I got windows testing to work on github! and found lots of test failures :D I am working through them :)

@econchick awesome! When I was working through this I found it easier to test on a Linux VM then figure out the Windows test failures 😁. I assumed they might be due to the posix-style path patterns in the tests?

[econchick]

Addressed by #18 (thanks again @hectormz!)