Some Feeds are not showing all items.

Hi!

I’m using Full-Text RSS 3.4 (selfhosted) and I like it. But sometimes I’m stuck and I don’t know why. Perhaps you’ve an explenation.

So I’m trying to setup “my” newspaper. I want to add two feeds:
http://www.faz.net/rss/aktuell/politik/
http://www.faz.net/rss/aktuell/feuilleton/

As you see: It’s the same website. I want to limit to 25 items. When I try that with my first link, it works superb. When I try this with the second link, I’ve to limit to five (!) items. Otherwhise it will not work.

Any ideas?

Thanks!

Gabe

Hi Gabe, I think this has more to do with a character encoding issue with Full-Text RSS than the item limit. I can see that the second URL fails even if I set the limit to 3 due to character encoding issues. We’re looking into this and hope to have a fix for it very soon.

Hi Keyvan!

Oh, an issue. Nice to hear, that you’re on it.

Meanwhile may I ask another question about the same source. As you see, the Full-Text RSS contains images. My newsreaders (iOS-Apps like http://newsify.co/ or http://www.airsource.co.uk/theearlyedition/ )are showing the images on the FrontPage, but not in the article. On the FrontPage some images are cropped.

If I do another side, there are the images in the FrontPage and the article. This is much better, because I get the text for the Image too. Example: http://ftr.fivefilters.org/makefulltextfeed.php?url=http%3A%2F%2Fwww.n-tv.de%2Frss&max=3

Do you got a hint for me?

Thanks!

Gabe

Hi Gabe,

My guess is that images associated with articles which appear in your news reading app but not in the article view are images which are embedded as part of the feed’s metadata. My guess is that your feed reader looks for this first and if it can’t find it, it will look for it in the article. Full-Text RSS preserves such metadata when processing the original feed, but might not find those images in the article content when it’s processing the web page for the article itself. That would explain why you’d see images in the list view and not the article view in your app.

If you give me an example of a feed where the images do not appear in the article view, but do in the list view of your feed reading app, I can see if my guess is correct or not. If it is, we might be able to help by improving the extraction for that site to include images.

Hey Keyvan!

Thanks for your getback.

Images only in the “Newspaper”-FrontPage:
http://www.faz.net/rss/aktuell/

Images in the “Newspaper”-FrontPage and in the article:
http://www.n-tv.de/rss

Now I viewed both feeds at the same time. I think, you’re right: www.faz.net Shows the featured image in the feed. Is there a way to get that into the article too?

But there are some articles with more than one article if you want to see one: http://www.faz.net/aktuell/wissen/weltraum/mission-rosetta/rosetta-zeigt-nahaufnahmen-von-komet-67p-tschuri-13433746.html

Thanks!

Gabe

Hi Gabe,

I tested the first feed and Full-Text RSS seems to be able to preserve images fine. Is yours producing different results (if you try in your copy of Full-Text RSS only)?

http://ftr.fivefilters.org/makefulltextfeed.php?url=www.faz.net%2Frss%2Faktuell%2F&max=3

We do not have a feature to insert images into the extracted content. So our only solution at the moment is to write custom extraction rules to try to preserve images when they’re not being included in the output. In this case, however, it seems what we have already is doing okay.

Hi Keyvan!

When you take a look into your browser, everything looks pretty perfect. But you’ve to check feed-readers (I checked five, my favorites are linked at my post from Feb 17, 2015 05:00PM CET).

Any news about the character encoding issue? :slight_smile:

Gabe

Hi Gabe,

It’s hard for us to check all feed readers (there are many!). But that URL I posted in my last email works perfectly fine in Feedly.com - you can try it without signing up if you don’t have an account, just enter

http://ftr.fivefilters.org/makefulltextfeed.php?url=www.faz.net%2Frss%2Faktuell%2F&max=3

In the search box you see after clicking ‘Explore’. I don’t know why the others don’t display the images like Feedly does.

As for the character encoding issue, we actually realised it was only affecting our hosted service, not Full-Text RSS 3.4. So if you’re still having issues with your copy, get back to us.

As for the problem you experienced. My guess is that setting the item limit to 25 is too much for your server, or the remote server starts to reject your requests after your first 5. That doesn’t really explain why it works for the first URL though, as they seem to be from the same server - but hard to tell, perhaps there are larger articles in the second feed that take up more memory than the first. If you enable debug output do you get any clues as to why it works for 5 items but not 25?

Our general advice regarding the item limit is to keep it low. For most sites you will not have 25 new items published between requests, so requesting 25 each time should not be necessary. The problem is that in processing one request to Full-Text RSS for 25 feed items, your server has to make 25 HTTP requests to fetch content, keep all that in memory and extract article contents from each response. This can exhaust your server’s resources or PHP limits. We generally don’t recommend the item limit being increased so much. If the feed is in fact updated so frequently that reducing the max items number means you miss items between requests, it’s still far better to decrease the item limit (to 5 or 10) but increase the frequency the feed is checked to ensure items are not missed between requests.

Hope that’s some help.