Uploading coverage results as part of CI

[RobPasMue]

🐞 Problem

When reviewing your CI jobs, I realized that you are computing the coverage of your library, but you are indeed not uploading it. This is easily seen in the pyfluent/Makefile file, in section "unittest"

As you may see in the image, you are running a coverage analysis, but this analysis is just kept as a result of the log of the CI job. It could be beneficial to upload it and show the resulting coverage as part of the README buttons (in a similar fashion to GitHub Actions or License - please see pyansys/pymapdl)

💡 Proposed solution

The coverage results should be generated in XML and HTML formats so that they can be uploaded to a coverage visualization web page or tool. For example, in pyansys/pymapdl, they are using codecov.io which provides enhanced visualization graphs for code coverage review as well as other metrics. You can access the results in this link to get a grasp of the potential of this tool.

Another option would be uploading the HTML files as artifacts of the GitHub Actions process (CI). These artifacts would be uploaded in each run and they would be accessible for the end user - this way, one could download the HTML reports generated by pytest and process the line-by-line coverage easily

Proposed for evaluation and review: @dnwillia-work @seanpearsonuk @mkundu1

Once again, I'd be happy to implement the modifications myself if you decide to go through with my proposal :)

[seanpearsonuk]

Thanks @RobPasMue please go ahead with this

[dnwillia-work]

Sounds fine with me, I had no idea we needed to do anything on this.

[RobPasMue]

Talking to @akaszynski, he informed me that the use of codecov.io is kind of restricted for "private" repositories. There is a limited amount of commits which can be performed... Thus, it is probably best to wait until PyFluent goes public to connect and upload metrics to codecov.io.

Nonetheless, I propose to generate the coverage reports in HTML format in the meantime, and upload them as artifacts. That way, at least one can download the coverage analysis and go through the line-by-line coverage. I'll get onto this on my branch feat/coverage-upload.

And once PyFluent goes public I'll open a new issue to enable the use of codecov.io.

Is it fine with you @seanpearsonuk @dnwillia-work?

[dnwillia-work]

Yeah, that sounds good to me.

[seanpearsonuk]

Agreed, thanks again

[RobPasMue]

Thanks for confirmation! Agreed changes are already available in PR #281

Looking forward to your feedback!