fix: Completely skip pages excluded by match_path

[prcr]

I was experimenting @rmorshea's great contribution on #43 to include only a small subset of relevant pages in the feeds for my site, hoping that it would be an elegant way to overcome the limitation in #23. However, I noticed the following:

• The plugin was still outputting warnings because it failed to obtain the correct Git dates from pages that I was excluding with match_path
• None of the pages that I was including with match_path would show up on the output

This was because the match_path filter was only being applied during the on_post_build event handler, and after the list of pages to include in the output had already been truncated by the configuration length.

To fix this issue I propose that we completely skip processing the pages that aren't included in match_path, with the additional advantages:

• On bigger sites, there is a potential performance improvement if you only need RSS feeds for a small subset of pages. The version that I'm proposing completely skips processing for pages that aren't to be included in the output as soon as possible, and also compiles the regex pattern only once to be reused on the pages that are included.
• This solution is more self-contained since the code used to handle match_path is less scattered, and could facilitate subsequent work on Add ability to include/exclude files using regex pattern #45.

[codecov-io]

Codecov Report

Merging #49 (6880ec8) into master (19f89e6) will decrease coverage by 0.20%.
The diff coverage is 40.00%.

@@            Coverage Diff             @@
##           master      #49      +/-   ##
==========================================
- Coverage   37.66%   37.45%   -0.21%     
==========================================
  Files           5        5              
  Lines         300      299       -1     
  Branches       57       57              
==========================================
- Hits          113      112       -1     
  Misses        182      182              
  Partials        5        5              

Impacted Files	Coverage Δ
mkdocs_rss_plugin/customtypes.py	100.00% <ø> (ø)
mkdocs_rss_plugin/plugin.py	33.33% <25.00%> (-0.39%)	⬇️
mkdocs_rss_plugin/util.py	25.15% <100.00%> (+<0.01%)	⬆️

[Guts]

Thanks @pauloribeiro-codacy for taking time to contribute. I'll try to make a review in the next days.

[prcr]

(I inadvertently closed the pull request when I mentioned it in codacy/docs#567. 😅 )

In the meantime, I'm testing this fix on the site that I'm maintaining:

• https://docs.codacy.com/feed_rss_created.xml and https://docs.codacy.com/feed_rss_updated.xml
• https://github.com/codacy/docs/blob/master/mkdocs.yml#L66-L68
• https://github.com/codacy/docs/blob/master/requirements.txt#L13

[Guts]

It's working for me.

@pauloribeiro-codacy with this behavior, is it possible to match multiple folders? for example with your website, include release-notes and articles.

[prcr]

@Guts yes, since match_path gives you all the power of regular expressions you can have more complex patterns to include multiple directories. For example, to include all pages under both release-notes and articles:

    - rss:
          match_path: "(release-notes|articles)/.*"

[Guts]

Ok thanks. I will add it for reference in the documentation.