Margin content in feed.xml

[ekstroem]

Howdy.

When feed.xml is created the content is provided by

   <description>{{ post.content | xml_escape }}</description>

However, the side margin information is included in the post.content which - for footnotes and references - can make the text rather unreadable. Is it possible to only get the content corresponding to the main column (i.e,. the content you can see for the smallest screen)?

[ghost]

Hmm, I think that makes sense. I'm happy to write a pull request if you and @clayh53 will review it.

[ekstroem]

I'd be happy to test that out

[ghost]

@ekstroem please test. You can cherry-pick this commit into your repo to test the code: https://github.com/clayh53/tufte-jekyll/commit/81ce4e356e42715faf7d737923349cc8d9e256f4 Thanks!

[ekstroem]

Testing it out right now but I'm having some trouble with nokogiri. Installation worked fine (local) but running jekyll build throws.

`require': cannot load such file -- nokogiri (LoadError)

Trying to figure out why my installation isn't working.

[ghost]

Please list the following:

• OS and OS version
• Ruby version
• Jekyll version

[ekstroem]

• Mac OSX 10.9.5
• ruby 2.2.2p95 (2015-04-13 revision 50295) [x86_64-darwin13]
• jekyll 2.5.3

[ghost]

Hmm, what command did you use to install the gem locally? It may be a path issue with your gems.

[ekstroem]

I ran

gem install nokogiri 

which gave no errors and when I check gem list --local then I get

...
mustache (0.99.8)
netrc (0.10.3)
nokogiri (1.6.7.2, 1.6.6.2)
pango (2.2.5)
parslet (1.5.0)
...

[ghost]

I don't think you should be able to run gem install without sudo on Mac. Sorry, I'm not sure what next to suggest.

[ghost]

Maybe does this help? http://stackoverflow.com/questions/19643153/error-to-install-nokogiri-on-osx-10-9-maverick

[ekstroem]

Possibly. I did try installation with sudo gem install nokogiri also but that resulted in the same error. However, I don't get installation errors. Only when running jekyll. I'll dig further.

[ekstroem]

Alright. Got nokogiri up and running by updating ruby and reinstalling the jekyll gem.

Here's a snippet from feed.xml where there is/was a sidemargin note.

frem til december 2015<label for="sn-dst-fødselsmåned" class="margin-toggle sidenote-number"></label>. Figuren herunder 

I think it looks great. Thanks a bunch. I'm not sure if it's worth keeping the whole <label ...> .... since that won't work in an RSS reader anyway but it doesn't do any harm I suppose

[ghost]

So just to verify, the output you shared is being generated by my changes? Or was that from before you tested my changes.

[ekstroem]

Generated by your changes. Could be merged directly (IMO)

[ghost]

Hmm, that's not how it should look. I want to completely remove the HTML elements in question. I'll have a look at this.

[ekstroem]

The sidenote reference ends up being in < label> ... < /label> so it disappears altogether from the actual viewable text.

[ghost]

Sure but there's still that mess of <label for="sn-dst-fødselsmåned" class="margin-toggle sidenote-number"></label> in the feed which shouldn't be in there.

[ghost]

Ah, I forgot a CSS selector, I need to search and remove labels that have margin-toggle sidenote-number

[ghost]

@ekstroem Ok, I just updated my branch and all elements of margin and side-note content should be fully removed from feed.xml - please test again when you can. thanks!

[ekstroem]

Just ran the updated version of strip_margin_content.rb and when looking at the same position in feed.xml as before it now reads

frem til december 2015. Figuren herunder viser 

All the unnecessary html fluff is gone. Bravo! I really appreciate you spending time on writing the ruby code for this.

[ghost]

Hurrah! Glad to do it. Thanks for testing it.

[ghost]

Btw, I modified an existing filter for this, so credit goes to @sumdog and this gist https://gist.github.com/sumdog/99bf642024cc30f281bc

[ghost]

Thinking further about this: removing marginnote and sidenote content completely isn't the best way to handle this. It would be better to convert all marginnote and sidenote content into footnotes for feed.xml. This way if someone is reading the site via a feed reader, they could still see the content at the bottom of each article.

Using nokogiri, you can replace a node in a document, as well as move nodes around. So to convert to footnotes perhaps something like this:

1. For each instance of <span class='marginnote'></span> and <span class='sidenote'></span> replace that node with a super-script number and footnote id, perhaps <span class='footnote' id='footnote-1'>1</span> or something and style this as super-script.
2. Move just the content of each node into a new footnote node located at the bottom of the page. Link this content back to the footnote id.
3. Cleanup the page by deleting all existing instances of marginnote and sidenote nodes. This commit does exactly that: https://github.com/clayh53/tufte-jekyll/commit/8f13e2a3a6a6f182677da019ed9bc8e2bdb8ae65

This leaves you with a page of footnote numberings in the text content, and associated footnote pairings at the end of the content. Write this as a Jekyll filter, and put it right after post.content in feed.xml to process each post's content before rendering it in the feed.

This isn't something I have time for. Maybe someone else can use these notes to write the necessary plugin.

Is it possible to only get the content corresponding to the main column (i.e,. the content you can see for the smallest screen)?

In reply to OP: If you just want to strip out all marginnotes and sidenotes from feed.xml, cherry-pick this commit: https://github.com/clayh53/tufte-jekyll/commit/8f13e2a3a6a6f182677da019ed9bc8e2bdb8ae65

[ekstroem]

I'd be fine for me to close this since I got the tweak I wished for, but I'm not sure if @clayh53 / @xHN35RQ have figured out what the best way to proceed would be.

[clayh53]

Another thing I want to investigate is whether we can tweak the jekyll rendering engine to do this for us when it creates the feed's xml file.

[ghost]

Right now the feed xml contains the generated post content, so it already contains any margin or side notes. You'd need to write a custom content renderer plugin to generate a feed specific post content where the margin/side notes are turned into footnotes. Essentially, for each margin/side note instead of including directly into the post, transform it into a footnote and include the note content in a new footnote section at the bottom of the post. Again, this seems like an awful lot of work just for the feed content.

Recently I found this plugin which uses gsub to filter out content to make it appropriate for inclusion in RSS feed. If you want to just filter note content entirely, using this approach is probably better than my solution which uses nokogiri, since gsub is part of ruby so you don't need an extra gem requirement:

https://github.com/mpc-hc/mpc-hc.org/blob/master/source/_plugins/fix_rss.rb