IMDB+ collection issue (1 Viewer)

styrsven

Portal Pro
January 6, 2013
115
51
63
Home Country
Sweden Sweden
I have recently moved back from my films to moving pictures. When rescanning my colliction, I couldn't help but notice the movies belonging to the IMDB top 250 collection don't belong to other collections although they should. E.g. tt0112471 "Before sunrise", in swedish "Bara en natt", belongs to |» IMDb Top 250| according to the scraper, but according to C:\ProgramData\Team MediaPortal\MediaPortal\IMDb+\Rename dBase IMDb+ Scraper.xml it should also belong to
<rename id="tt0112471" title="Before Sunrise" sortby="Before Sunrise 1" collection="Before Sunrise" />

Anyone got a hint of where the error might be?
 

styrsven

Portal Pro
January 6, 2013
115
51
63
Home Country
Sweden Sweden
After some trial and error and memories slowly coming back, I swapped the order of the sortby and collection tags in the rename database for this collection, and suddenly the renaming started working.
 

styrsven

Portal Pro
January 6, 2013
115
51
63
Home Country
Sweden Sweden
More experiences of the renamer:
It seems it has some issues with collection names from different sources.
When I did the fix in the previous post, "Before sunrise" was added to "Befor sunrise" collection, but also disappeared from the IMDB top 250 collection.
Also, the Rocky collection has the collection extracted from the name, but Rocky I is not included in the collection because it also is specifically included in the Oscars collection. After adding Rocky in the collection field it is included in both.
The same goes for "Silence of the lambs".
 

RoChess

Extension Developer
  • Premium Supporter
  • March 10, 2006
    4,434
    1,897
    @styrsven, I noticed that bug myself as well on another movie, and was a logic mistake in the way Top250 works, including another bug on subtitle movies, and both are fixed now with new version.

    Did not get around to fixing it until today, so grab latest version and I would let the IMDb+ plugin do a full refresh on all the replacement dbase entries so that it can fix all the Top 250 and other collections issues, such as "Mission: Impossible IV - Ghost Protocol".

    PS: You just made me realize I overlooked the "Rocky I" mistake, which should be in Top250, but indeed fails on the Rocky collections, because it needs to be manually added to that when I overwrite the collection tag. Will fix that.
     

    styrsven

    Portal Pro
    January 6, 2013
    115
    51
    63
    Home Country
    Sweden Sweden
    Thanks RoChess,
    I also think there is an issue that the attributes in the translation file has to be in the exact same order as they are listed in the regular expression in the grabber script, or else some attribute(s) is/are missed. That means you have to peek in the grabber script to be sure to get the translation right. Maybe it can be addressed by splitting up in several regexes or using xpath?
     

    RoChess

    Extension Developer
  • Premium Supporter
  • March 10, 2006
    4,434
    1,897
    Well that is why I specify the exact syntax at the top. RegExp had a massive improvement in speed over xpath, because it allows me to immediately jump to the "one" entry, whereas xpath first resolves every node. There is logic behind the madness :)
     

    styrsven

    Portal Pro
    January 6, 2013
    115
    51
    63
    Home Country
    Sweden Sweden
    Ok..
    There are a number of samples in the translation file that doesn't follow the exact syntax, e.g.

    <rename id="tt0478970" sortby="Ant-Man I" collection="The Avengers|» Superheroes" />

    where 'sortby' preceeds 'collection' so the scraper finds 'sortby' but misses 'collection'.

    In some places (like the sample above) I suspect that 'sortby' should really be 'title', but in other places it is just an obvious error, like

    <rename id="tt0112471" title="Before Sunrise" sortby="Before Sunrise 1" collection="Before Sunrise" />
    <rename id="tt0381681" title="Before Sunset" sortby="Before Sunrise 2" collection="Before Sunrise" />
    <rename id="tt2209418" title="Before Midnight" sortby="Before Sunrise 3" collection="Before Sunrise" />
     

    RoChess

    Extension Developer
  • Premium Supporter
  • March 10, 2006
    4,434
    1,897
    The syntax with RegEx was designed to allow for options to be skipped, as long as the order is correct.

    But you are totally right, somewhere along the line I totally screwed things up, because the RegExp expects title/collection/foreign/sortby, and I have been editing the dbase as title/foreign/sortby/collection

    Can't believe nobody caught that before, especially me :)
     

    RoChess

    Extension Developer
  • Premium Supporter
  • March 10, 2006
    4,434
    1,897
    Okay, all fixed with new update on scraper-script, and master dBase, will make a post in main thread with more details.
     

    Users who are viewing this thread

    Top Bottom