Introduce paragraph count summary

[gagath]

Let the users use only the n-first paragraphs of the article as article
summary.

The advantage of this approach is that we avoid the random word count
ellipsis that will cut content in pieces, while not having to copy the
first paragraph of the article into the article's summary metadata.

Pull Request Checklist

• Ensured tests pass and (if applicable) updated functional test output
• Conformed to code style guidelines by running appropriate linting tools
• Added tests for changed code
• Updated documentation for changed code

[boxydog]

The functionality (paragraphs) seems reasonable. Using negative and positive SUMMARY_MAX_LENGTH feels confusing. I'd introduce a different config var. But, my opinion is not binding.

[boxydog]

Less confusing would be SUMMARY_MAX_LENGTH (for words) and SUMMARY_MAX_PARAGRAPHS (for paragraphs).

If both are set, could 1) apply both, or 2) pick a winner (e.g., SUMMARY_MAX_PARAGRAPHS), or 3) raise an error, or 4) hidden option 4.

[paulocoutinhox]

+1

[avaris]

My time on the interwebs taught me that the possibility of finding a closing tag for <p> is like a coin toss. Even the HTML spec is permissive about it and "formally" allows omitting </p> in most cases. Long story short, I am trying to say that this is quite unreliable.

[gagath]

The HTML is mostly generated from a generator (reST, Markdown…) that hopefully emits closing tags, so in practice this should not happen very much.

[avaris]

in practice this should not happen very much.

That doesn't cut it though :). rst and md generators may play nice (for now) but content can come from any number for sources not limited to those, including direct html files, plugins for other formats etc. In software, edge cases happen more common than one would think initially :).

I'm not suggesting employing full DOM parsers like lxml or similar, although that would be more complete. But that can be postponed until the need arises. It should at least produce something reasonable in those cases: i.e. if </p> is not present (substr.find('</p>') == -1), take up to the next <p>:

tag_stop = substr.find('</p>')
if tag_stop == -1:  # no closing p
    tag_stop = substr.find('<p>', tag_start + len('<p>'))
    #  if tag_stop == -1:  # if there are no more <p>s take everything else?
    #      tag_stop = None
else:
    tag_stop += len('</p>')

[gagath]

• Rebased
• Added new setting SUMMARY_MAX_PARAGRAPHS
• Added a new test to verify what happens when both SUMMARY_MAX_PARAGRAPHS and SUMMARY_MAX_LENGTH are set
• Fixed formatting issues
• Added missing documentation

[gagath]

@avaris Thanks for your comments, fixed them.

[justinmayer]

Many thanks to @gagath for the enhancement as well as for your patience.

Thanks also to @boxydog and @avaris for reviewing!