Title detection

Hi,

Some feed contain multiple titles, perhaps a short and a long version. For example see http://feeds.bbci.co.uk/news/rss.xml which contains:

item
  title Assad says Britain bullying Syria 

while the article itself:

http://www.bbc.co.uk/news/world-middle-east-21646091

contains

Short title: Assad says Britain bullying Syria
Long title: Syria conflict: Assad accuses UK of bullying

Now the default site_config contains a title rule which seems to match the LONG title, but the resulting feed seems to only get the short title (from the original feed), and the long title is stripped out completely unless I comment out title title: line.

Is there any way to force full-text-rss to use the full title in this case rather than defaulting to the shorter title it’s getting from the RSS? Many thanks for any insight and thanks for a great tool!

Hi Mark, this will be possible in Full-Text RSS 3.1. At the moment the title specified in the site config file is only used in the generated feed when the input URL points to a web article. When dealing with a feed, we assume the item title in the feed is the preferred one. This will still be the case in Full-Text RSS 3.1, but we’ll have an option to allow you to override it.

Hi Keyvan, that’s great thanks - at least I now know I’m not going mad! You might like to consider updating the documentation for site configuration rules if 3.1 is going to be a few weeks away to make it clear that Title: fields don’t quite work as expected for RSS feeds.

It seems quite a few sites use an abbreviated title in their feeds and it’s nice to convert that to the “full text” version rather than losing the often slightly broader context.

Looking forward to 3.1, many thanks!!

Mark Knight

Hi Mark, sorry about that, the site config docs were quite misleading. I’ve updated the help page here http://help.fivefilters.org/customer/portal/articles/223153-site-patterns - will update again when 3.1 is ready - should be out by the end of the week.

Thanks for updating the documentation, much appreciated :wink:

Mark Knight