API: Smarter cache updates

[oss]

Hi guys,

today I work on smarter cache updates.

Scenario before:

1) xml-rpc SearchSubtitles with moviehash, moviesize and sublanguageid
2) some results found or no results found, cached for 24 hours be default
3) (Try)UploadSubtitles() with given moviehash, moviesize, and sublanguageid
4) xml-rpc SearchSubtitles with moviehash, moviesize and sublanguageid => come from cache

Fixed now:
1) xml-rpc SearchSubtitles with moviehash, moviesize and sublanguageid
2) some results found or no results found, cached for 24 hours be default
3) (Try)UploadSubtitles() with given moviehash, moviesize, and sublanguageid, cache is expired
4) xml-rpc SearchSubtitles with moviehash, moviesize and sublanguageid => come from database and cached for 24 hours.

[watdafox]

If I understand well, the cache is refreshed whenever anyone uploads a subtitle? That would be awesome, results indeed take a really long time to appear for newly released episodes

[oss]

watdafox, exactly.

Got already good feedback about this, but not here It took couple of hours to sort this out, it is not absolutely perfect, but it is 99% better then before.

You can try.

[watdafox]

either it's not working, or it's because I don't send sublanguageid, I left the detection to you.

[oss]

should work, even without sublanguageid.

send me please what are you sending (or some more info). It is not working with query/tag searching, but with others (moviehash/moviebytesize/imdb) it should work.

send some more info so I can duplicate the problem. thx

[watdafox]

Step 1) I search subs:

Code: Select all

Object {imdbid: "tt1405406", season: "7", episode: "2"}

Returns 0 subtitles.

Step 2) I upload a subtitle:

Code: Select all

Object {subpath: "D:\The Vampire Diaries - 07x02.srt", path: "D:\the.vampire.diaries.702.mkv", imdbid: "tt1405406"}

Note: subpath=gzip/b64 subtitle content // path=extracts moviehash & moviebytesize
Response: subtitle was added (and I can see it uploaded on my account & is available on website, instantly)

Step 3) I search subs again:

Code: Select all

Object {imdbid: "tt1405406", season: "7", episode: "2"}

Returns 0 subtitles.

******

that was a few hours ago, when the episode had aired not long ago. Right now (1am CET) the call still returns 0 subtitles.

[MarcoTC]

Strange.
Maybe you can submit request(TryUpload and Upload) and responses -xml to look at? Are you getting the URL back in the Upload response?
For me, it works instantly.
It used to take 24 hrs depending if I was the first doing the search. Now it's instant.
I must say that I'm passing all the fields. Including language.

[oss]

watdafox, yes with your example it is not working (yet). To be more technical, so you guys understand whats going on.

- all searches are parsed, and then I calculate md5 of input fields. So when something like "imdb => 123456" coming, I serialize it, then calculate md5, then I check nosql db for value, if it doesnt exist, I script searches, and store it.

So for correct uncaching, I need to actually uncache all possible searches, how one can find the subtitles. Now working 100% uncaching moviehashes, also imdbis (but not with seasons and episode).

I will try to fix this, but not sure when