What can be tested for oldest_article

[AlanRoberts]

Looking at the RSS I'm working up a recipe for, I don't see any date/time datum for the item:

  </item>
  <item rdf:about="http://www.csmonitor.com/World/Europe/2017/0714/During-Trump-s-Paris-visit-Macron-positions-himself-as-intermediary-between-Europe-and-president">
    <title>During Trump's Paris visit, ...</title>
    <link>http://www.csmonitor.com/World/Europe/2017/0714/During-Trump-s-Paris-visit-Macron-positions-himself-as-intermediary-between-Europe-and-president</link>
    <description>&lt;p&gt;French President Emmanuel Macron invited President Trump and ...</description>
  </item>

There is a time element within the article body that seems useful:

<time
	id="date-published"
	class="eza-publish_date outer_block all_caps"
	datetime="2017-07-14T10:40:02-04:00"
	title="Published: July 14, 2017 10:40 am EDT">July 14, 2017
</time>

Can I customize to make use of this element, or is oldest_article only applicable if the RDF contains a date?

Thanks,
Alan

[kovidgoyal]

It needs a date, but you can always customize how the rss is converted into articles by re-implementing parse_feeds() in your recipe class.

[AlanRoberts]

Okay, I guess that means learning a lot more Python than I've had to deal with so far. Are there any existing recipes that you recommend I look at for good examples of reworking parse_feeds()?

While I'm learning, I could try an alternate approach, and suggest to the RSS provider that they add a date/time to their feed. What would calibre need to see, a dc:date element, or something different?

Thanks,
Alan

[kovidgoyal]

Just look at the source code of the BasicNewsRecipe class. And the RSS standard specifies how to add timestamps to feed entries.