Help me improve filename recognition

Hi guys,

I need your help once again. This time it’s needed to improve recognition of filenames in Mizuu. To do so, I’ll need a dataset to work with – in other words, I need a bunch of different filenames and folder structures.

To get this, I’ve set up a Google Documents form that’ll allow you to enter a video filename. There’s a field for both movies and one for TV shows. Below the fields, you’ll find a short description of what the expected input is and an example of that.

Please add as many different filenames as you can – my goal is to create something that works as expected on the entire dataset. You can only submit one field at a time, but there’s a link to fill out the form again once you’re done.

ADD YOUR FILENAMES HERE

I’m mainly interested in TV show filenames, as that’s the area that needs the most improvements, but movies are of course appreciated as well.

Thanks a lot!

13 Comments

  1. I’d love to help but entering each file one by one is just SO inefficient. If you want, I can whip up the names in my database in a .txt format or something similar.

    My base is more or less organised though (everything follows the same idea: name (year) – sXXeXX – episode_name, or: name (year) for movies, all organised into folders named like: name (year) and further categorized into folders for each season.

    I could also send you the files of various movies/series I have downloaded but not organised (just whatever name they came with).

    If you’re interested, reply here or send me an email or something and please specify the format you’d want to have it in :).

    -NMB

    Reply
    • I just changed it to allow for longer inputs. If you separate by commas or semicolon, it’ll be great!

      Reply
      • Awesome, way better :).
        Want just the organised part or both? On one hand, if you manage to correctly identify the un-organised files too it’d be awesome for Mizuu, but it might mess with your algorithms too much, since I really can’t say anything about what the file names look like (just normal stuff you see on internet downloads).

        -NMB

        Reply
  2. Something has run a fowl in the latest updates. Movie and TV show identification is all screwed up. I have always had my movies in folders structured as movie name and year in brackets ie. Tron (2010) both folder and file name, Mizuu cannot identify any of them becus of the year. I need to go in and manually ID every one. TV shows merges episodes from one show into another with a bunch of unidentified episodes. In the beta releases and prior there was not that many issues with video scraping and ID, now if you are trying to rebuild your library it is unusable.

    Moreover when you start the process to search for shows it takes forever to start, it just hangs there for minutes.

    I suggest you look at XBMC and archos on how they ID shows, they never have these issues even with movie/tv database script changes.

    This is how my movies and TVshows are named which the 2 other programs above have no issues since forever.

    For Movies:
    13 Assassins (2010)

    For Tvshows:
    Folder (show name) >>> Sub Folder (season 1, 2, 3 etc)
    episode name:
    Ex: Cowboy Bebop – s01e02 – Stray Dog Strut

    Reply
    • Hey! Are you using the free version?

      Reply
      • I had the problem with the tv show too. So I delated the free version and paid for the aktuell app. But the problem is sill here. Not for all shows. But for “Damages” in my case (Damages-s03e01.avi). From 26 episodes “Mizuu” shows just three in the database. Rest is unknown filename.

        Reply
        • The issue is due to the single hyphen. Replace it with a space or similar and it’ll work. This is something I’ll try to improve, so please add that filename to the Google form. Thanks!

          Reply
  3. Hi,
    i just find issue with TMDB,
    My filename is L_HOMME_QUI_RIT and TMDB was not able to find somthing but if i put L’homme qui rit film is found.

    you d probably need to had rules that replace L_…. by L’

    regards

    Blairfancy

    Reply
    • I’ll keep it in mind. Please add a sample though, so I get it in the data set with the other filenames that other people have added.

      Reply
    • Hi again!

      Just wanted to let you know that I’ve managed to get it working. It also works on titles that start with “C_est”. The new filename recognition is soooo much smarter and better.

      Reply
    • Hi Bobby!

      Yep, specials must include “00” as the season number.

      Reply

Add a comment