Help needed with development of new scraper for filmdelta.se (Swedish Movie Scraper)?
#46
Oops? please open up a new ticket with attached fix, sorry Smile
Always read the online manual (wiki), FAQ (wiki) and search the forum before posting.
Do not PM or e-mail Team-Kodi members directly asking for support. Read/follow the forum rules (wiki).
Please read the pages on troubleshooting (wiki) and bug reporting (wiki) before reporting issues.
Reply
#47
vdrfan Wrote:Oops? please open up a new ticket with attached fix, sorry Smile

np. I guess you didn't do it on purpose Wink

I'll open a ticket later today. Just got a power failure here and I don't know how long my battery is going to last Confused

/Daniel
Reply
#48
Bah. I've been struggling half the day trying to get this to work again. It just won't. Have things in the scraper api changed in svn?

It simply seems like buffer 2 as sent to CreateSearchUrl is empty. Isn't the year supposed to be there?

Is there any description anywhere of exactly what is in the buffers when the different functions are called?

Tearing my hair. Strange thing is that I'm sure that this worked before Confused

/Daniel
Reply
#49
haha. vdrfan will be by shortly and explain my laughter Wink
Reply
#50
Thanks Wink
Always read the online manual (wiki), FAQ (wiki) and search the forum before posting.
Do not PM or e-mail Team-Kodi members directly asking for support. Read/follow the forum rules (wiki).
Please read the pages on troubleshooting (wiki) and bug reporting (wiki) before reporting issues.
Reply
#51
http://trac.xbmc.org/changeset/21992
http://trac.xbmc.org/changeset/22013

bad day Smile
Reply
#52
spiff Wrote:http://trac.xbmc.org/changeset/21992
http://trac.xbmc.org/changeset/22013

bad day Smile

Hmmm... Meaning I chose a very bad day for doing scraper development, right? Stare

Well, ok. Exactly what will I get in my buffers in revision > 22012?

Anyway, while there's development going on in the scraper api, may I make a suggestion? I would like to have the year sent to GetSearchResults as well. Now I'm doing it myself by extracting it in CreateSearchUrl and telling it to not clear the buffers, which feels like an ugly workaround. Would be nicer if it was a part of the api that some buffer always contained the year when calling GetSearchResults. Or?

/Daniel
Reply
#53
Some sites (like themoviedb) won't even get you one result if you search with year :/
Always read the online manual (wiki), FAQ (wiki) and search the forum before posting.
Do not PM or e-mail Team-Kodi members directly asking for support. Read/follow the forum rules (wiki).
Please read the pages on troubleshooting (wiki) and bug reporting (wiki) before reporting issues.
Reply
#54
Nah, scraper development time is fine, cptspiff just wants to let us know that i had a "bad day" when "sabotaging" the filmdelta scraper :p
Always read the online manual (wiki), FAQ (wiki) and search the forum before posting.
Do not PM or e-mail Team-Kodi members directly asking for support. Read/follow the forum rules (wiki).
Please read the pages on troubleshooting (wiki) and bug reporting (wiki) before reporting issues.
Reply
#55
Ok, what about the following:

For the search in createsearchurl, I use (.*?)%28 from buffer 1 to get rid of the year. I also put %28(.*?)%29 into buffer 9 for later use in getsearchresults. And in getsearchresults I use the contents of buffer 9 to sort out all movies with the wrong year. Does that seem like a good idea that's going to work?

/Daniel
Reply
#56
Sounds like a plan. Searching with contents of buffer 1 and the year won't work?
Always read the online manual (wiki), FAQ (wiki) and search the forum before posting.
Do not PM or e-mail Team-Kodi members directly asking for support. Read/follow the forum rules (wiki).
Please read the pages on troubleshooting (wiki) and bug reporting (wiki) before reporting issues.
Reply
#57
vdrfan Wrote:Sounds like a plan. Searching with contents of buffer 1 and the year won't work?

Well, it returns hits. Problem is it only searches in the title, so for example searching for "Robin Hood (1973)" returns a load of movies that has 1973 in the title but are otherwise unrelated. In other words, I get a cleaner hitlist if I remove the year before searching...

/Daniel
Reply
#58
Ok, now. This works good for me:

http://pastebin.com/m3f9308c1

Comments?

Btw, how do I easiest get my xml indented in a way that makes you happy? I wanna make people happy!

/Daniel

edit: Oh, forgot to tell. I made another change. I send more=1 to search.php. Makes it always return a list (instead of taking us direct to the movie if it fits perfectly). Seems like a good idea to always have a list, right?
Reply
#59
Daniel Malmgren Wrote:For the search in createsearchurl, I use (.*?)%28 from buffer 1 to get rid of the year. I also put %28(.*?)%29 into buffer 9 for later use in getsearchresults. And in getsearchresults I use the contents of buffer 9 to sort out all movies with the wrong year. Does that seem like a good idea that's going to work?

When thinking about this, there is one problem here. People whose movies does not follow the "title (year)" standard will have problem here, they will get no results at all. Is there any way of filtering on year in GetSearchResults if there is a year, and show everything if there is none?

/Daniel

edit: I guess conditionals are the answer here. Never used those, will start playing with them now :-)

edit 2: No. Not conditionals. Didn't seem right after a closer look. Does xbmc support if-then-else-stuff in the expressions?
Reply
#60
Daniel Malmgren Wrote:Btw, how do I easiest get my xml indented in a way that makes you happy? I wanna make people happy!

For those interested, I found the solution to this. Setting the environment variable XMLLINT_INDENT to $'\t' makes xmllint produce tabs instead of spaces for indentation.

/Daniel
Reply

Logout Mark Read Team Forum Stats Members Help
Help needed with development of new scraper for filmdelta.se (Swedish Movie Scraper)?0