Hello all, how’s your hollyday? I hope you are not watching movies and not still sitting in front computer like me 🙂 Anyway, there are some news:
Support for movies not in IMDB
this feature was most missing and users asked for it from early times, so here it is. Now you can add your own movie/video to database and then upload subtitles to it – please don’t add movies which are already in IMDB. Technically – it is just workaround – it adds movie to database with IMDB greater than 10.000.000 – and that means it is internal movie. Should work good, if not, let me know. Some more details (rating, cast, directors…) is not supported, but should be done in future…Support for adding movies using XMLRPC coming soon 🙂
ALLPlayer automatic subtitles upload
Guys at ALLPlayer implement automatic upload of subtitles, in newest version it detects .nfo files and it has support for 2CD movies and some more cool features. Automatic subtitles upload works really nice – after 80% of watching movie are subtitles uploaded. I like to see this feature in more players out there – read first for developers.
There are more changes: voting is disabled for anonymous users, smart detecting similar subtitles, better response times (still should improve:), improved handling subtitles requests for admins…
Enjoy summertime!
Start the discussion at forum.opensubtitles.com
XMLRPC InsertMovie() done:
http://trac.opensubtitles.org/projects/opensubtitles/wiki/XMLRPC#InsertMovie
It is possible voting using XMLRPC: http://trac.opensubtitles.org/projects/opensubtitles/wiki/XMLRPC#SubtitlesVote
Also add comments should be nice, eh ? 🙂
This is VERY nice.
I’m not so convinced about the InsertMovie method (I can see a lot of margin for error, when incorrect IMDB searches mean more “non-imdb” movies added). But the SubtitlesVote is VERY sweet.
I just noticed it’s possible to add comments through XMLRPC but not to see the comments that exist (or how many there are).
Wouldn’t we want to have both of these on the XMLRPC? The comments are an important factor to decide whether or not to select one subtitle or another.
eduo: Ok, so you want to return in SearchSubtitles() number of comments and also new method GetComments(). OK, I can do that, no problem 🙂
You kill me 🙂
I kno you’re doing this just so I don’t insist in TV support. I just know it.
eduo: done:
http://trac.opensubtitles.org/projects/opensubtitles/wiki/XMLRPC#GetComments
http://trac.opensubtitles.org/projects/opensubtitles/wiki/XMLRPC#SearchSubtitles
I should make a new article about changes…
Sweet.
You’re right, tracking changes wouldn’t be a bad idea.
On the SearchSubtitles changes:
It’s nice to see that all possible searches (except for tv series, *hint*, *hint*) are covered. I would’ve added a parameter for incremental search.
Explanation: As you’ve currently made it the search is made depending on the available data. Hashes -> IMDBID -> Name.
The problem is that the searches are exclusive. Searching for one means not searching by others.
I would do it like this (adding a parameter for “flexible search”, for example):
1.-If present, use Hashes.
-> If hash is not succesfull but has IMDB match use IMDB and IMDB Name in next steps.
2.-If IMDB present and/or Hash failed use IMDB for search.
3.-If Name present and/or Hash failed use IMDB Name for search
This way you can always pass all the data you have. One, two or three types of query and obtain a single set of results and the result would say what query type was used:
Result1: Hash Matched
Results2, 3: User-provided IMDB Matched
Result 4: IMDB-From-Hash Matched
Result 5: User-Provided Name Matched.
Result 6: No match
The other option is returning all the results for all (Hashes, User-provided name, User-provided or Hash-provided-IMDB) and flag each result for relevance depending on the source.
This may be complicating things a lot, so feel free to ignore it. It would only save multiple searches (search once by hash, if IMDB is returned search by IMDB, if not search by name).
Those searches in SearchSubtitles are not exclusive. You can search by anything, you just can not combine anything in one array. I can pass result so you can see where it comes from, thats not a problem. Technical background: I am searching by hashes and moviesizes, that should be the fastest (if in input array are these data), then by imdbid, then by fulltext (sphinx), and everything is joined by union, so I can add some explanation where it comes from (“MatchedBy” => “moviehash|imdbid|fulltext”), no problem, thats 2 minute work.
I know TV series is a must. I am still improving the site a little bit and this will be the next I guess, but I need a more time for this, it is huge improvement (bigger then other).