How does uTorrent parse RSS feeds

[luk32]

Hi there,

As the first thing I'd like to thank anyone who helped to make this project out, it's really great... and I'm also kind of astounded 'cause of the last changes.

Thus I have a question considering RSS, it's as stated in the topic. How does uTorrent parse RSS feeds?

I am interested how does it split the feed into sections as Title; Episode ...

So if, you're in a rush, or don't want to bother with irrevelant thing just don't read the "in detail..." part

In detail, it is because it doesn't work in my case or at least the www that I am using.

That would be very nice to automate torrents downloading on this page.

They are providing links to the anime episodes and the names aren't typical to the most TV releases, like

title.S[number]E[number].codec.etc.etc

In my opinion it would be nice, to develop a system where user could input the parsing formula, because sometimes it'd be very sophisticated, like when the torrent's name or link, or any info where you could get the idea what you would download by clicking it, is separated with spaces and there could be number or non-alpha characters like

".//hack sing" or "Canvas 2 ~Niji Iro no Sketch~ 13 [Lunar]"

If the user could develop his own parsing formula, I think the RSS system would be even more flexible.

Some of people with little determination could work-around really much thanks to this.

.. or maybe there's a different solution to my problem.. Not to complicated, at the moment I have some on my mind, but it's based on reading the rss, parsing it to the uTorrent-friendly syntax, and then placing it where uTorrent can reach it.

This one needs at least a task-schedule program and the script to parse the data and maybe some way to get the feed. I think all the needed things are provided or easy to get on windows or linux systems. Though, still, bringin' such a piece on legs would be a lot of work, beside the script-writing of course, that's additional

As I am thinking now, being able to successfully parse almost every kind of data would need a really powerful and complex parsing system too...

Okey, I stop smashing the reader with my thoughts now. If someone's would be willing to conversate over the manner I'll try to reply.

Thanks for reading this stuff, and sorry for my english skills .

Oh, and also, I didn't find anything in the faq or via search.. "sorrey" if I am touching something again.

Peace.

[splintax]

A lot of people want regex implemented in the RSS reader, assumedly this is one of the reasons. Unfortunately this is quite complex to implement and I don't think ludde's interested, but this is just hearsay..

As for a workaround to your solution, what comes up in the Name column for your anime episodes? You can use the wildcards *, ? and also | for OR, so I imagine you could get something worked out? *hack* might work, or something..

[luk32]

Hello,

sorry, that I didn't reply eariler, in fact i tried, but you now, preview, then back-button, can cause a lot of psychical damage to the poster ... simply I've lost the whole text.

I was wondering how does uTorrent behaves with handling automatic rss dowloads, e.g. when there's a torrent that is already on your list, or when you downloaded data and deleted the torrent, when it's still in the feed's queue. Would it download it again?

oh, i also provided one <item> entry from the feed

- <item>
  <title>Bomberman Jetters 48 [Ani-Kraze]</title> 
  <link>http://tracker.danomac.org/torrents/%28Ani-Kraze%29Bomberman_Jetterz_Ep48_%5BB1B0E31D%5D.avi.torrent</link> 
  <description><b>Bomberman Jetters 48</b><br />By <a href="http://www.animesuki.com/group.php/18.html">Ani-Kraze</a><br />Filesize: 171.0 Mb.<br />Show all torrents for <a href="http://www.animesuki.com/series.php/54.html">Bomberman Jetters</a></description> 
  <pubDate>Sat, 14 Jan 2006 06:58:14 +0100</pubDate> 
  <guid isPermaLink="false">http://www.animesuki.com/details.php/15244.html</guid> 
</item>

so as you see, i very probably could write such filter that would work for this one, but it gets complicated, when there's number in the title....

you know, whe i think more over this, i'm getting to the conclusion, that the problem would be only to parse the entry and to divide it. For now I could treat the title of the anime with all of it's special characters as the constant string that need to appear to be accepted by the filter e.g.:

*so/\/\e.a:\nime//title 3* and there coulde some info before, such as fansubbing group and after maybe the episode number...

am i thinking right ?

... or i guess i'll check it later and give my feedback.

thanks for replying

[rafi]

There is a short guide here:

http://forum.utorrent.com/viewtopic.php?pid=36682#p36682

As a general rule - exact torrents - that are already in the history list (have been DLed) will NOT be DLed again. But the problem is - there are may similar torrents distributed by different groups (RSS feeds) and many formats (qualities) .

Is order to avoid double DLing - as far as I can tell, the following two features are being added as we speak...:

1. Being able to DL only ONE torrent per filter per scan (the first one detected )

2. Being able to pause scanning feed(s) per filter for a timeout period (12 hours to ~1 week) so to make sure that only "next day/week's" episode will be detected.

It is not full proof, but may help to focus/DL only what we need.

If you want to play with it a bit - the beta version is located here:

http://utorrent.com/download/beta/utorrent-1.4-build-403.exe

[luk32]

Thanks for your points rafi and for replying too.

Of course as a part of making sure I am not doubling someone's issues, I've found geezer's guide. It didn't resolve my problem, though.

Maybe to make things clearer, i am working of trying to work with

http://www.animesuki.com/rss.php?link=enclosure

if one checked the link it's easy to see that right after the release's name is the subbing group, so i could easyly include it to the filtering string.

As for the features you mentioned, they won't work in certain cases, that sometimes occur e.g. when there are several episodes released at one time.

Obvioulsy you can't let to add every matching torrent, because different groups could release the same title, and you can't add just one in case when one group would release serveral eps in the same time.

Perhaps, it can be work-arounded as I am able, to apply group filter, I can let to download every torrent [just once per torrent of course but it's already provided] for one particular title of one particular group.

Also, my opinion about this kind of resolving problem, i mean, inventing thigs to leap over some certain cases can cause that such code would be even more sophisticated and less intuitive [sorry if misspelled] in use, than one complex solution.

Anyway, I don't know that, I am not experienced in developing such big projects [i know it's one app, but as for me it does really much stuff]. Don't know if it's better to make it bigger with many work-arounds than keep slim with complex solutions... I'm just thinking loud .

Thx for reading.

...

some time later

...

Okey, i admit that with the tools provided within build 403 i can handle, at least some of my cases, i think it's almost everything if not every thing, the "NOT" filter is nice addtion too.

Now by manipulating with a few filters and those options given, i can handle whole title (e.g. from several groups) preety nice.

Though i am thankful for implementing those, i still think that more complex filtering formula would be like the cherry on the top of that rss handling cake.

Thanks again, to rafi for directing me to this beta b403 and to the one responsible for developing the thing and implenting it.

Luk32.

[rafi]

There still are a few bug in this beta, so watch it...

http://forum.utorrent.com/viewtopic.php?pid=41211#p41211

[Firon]

http://www.utorrent.com/download/beta/utorrent-1.4.1-beta-build-405.exe