Cleaner movie pages in DMM

Movie pages on DMM now clean themselves up. A lot of what gets scraped under a movie isn't that movie at all. Sequels, the TV series with the same name, soundtracks, games, other films from the same year and random junk all end up on the wrong page.

Here's some of what got taken off on day one. Zootopia 2 had the 2016 Zootopia on it. Harry Potter and the Sorcerer's Stone had Half-Blood Prince, Shallow Hal and the PS2 game. Pirates of the Caribbean: The Curse of the Black Pearl had Dead Men Tell No Tales, Sinbad: Legend of the Seven Seas, Runaway Jury and Lego Pirates for the Wii. Captain America: Civil War had Kung Fu Panda 3 and The First Avenger. Avengers: Age of Ultron had Italian football matches. Barbarian had Conan the Barbarian and Waiting for the Barbarians. The Gentlemen had episodes of the TV series and The Dark Knight had The Dark Knight Rises. That's almost 300,000 releases off 935 movie pages so far.

When you open a movie, DMM checks every release it hasn't looked at before against every title IMDb knows the movie by, in every language, and its year. Simple rules catch the easy ones like episode numbers and the wrong year. For everything else a small and very fast AI model looks at each filename and says what kind of file it is and whether the title is really this movie's. It runs in the background after the page loads, so the next visit is already clean.

Nothing gets deleted. Every release taken off a page is kept with the movie it came from, so a wrong call can be put back. Packs and collections that include the movie stay, and so do translated titles. On 300 filenames I checked by hand it agreed with me 98% of the time, and I've been fixing every miss I found since it went live. Popular movies are already cleaner and the rest get cleaned as people open them.

If you see a real release missing from a movie page or junk that slipped through, tell me! I will soon apply this to TV shows and anime https://debridmediamanager.com