Movie Lookup problem

[kim]

I don't know if you will call it a bug, but:

if filename "Paul Blart Mall Cops 2" (should be "Cop"), filebot don't/cant even list the movie as a possible option...

only this on list: Paul, Mall

if filename "Paul Blart Mall Cop 2" all is well

[rednoah]

So what's the file name? What's the movie it should be matching?

Have you read FAQ #2 on how to deal with especially badly named files?

[kim]

FAQ #2 is useless here as its the filename that has only 1 extra letter...

https://www.themoviedb.org/movie/256961 ... mall-cop-2
wrong filename "Paul Blart Mall Cops 2"
right filename "Paul Blart Mall Cop 2"

so why can filebot find this from this filename "Paul Blart Mall Cops 2"?:
Paul - https://www.themoviedb.org/movie/39513-paul
Mall - https://www.themoviedb.org/movie/276901-mall

BUT it cant find this:
Paul Blart Mall Cop - https://www.themoviedb.org/movie/14560- ... t-mall-cop
or
Paul Blart Mall Cop 2 - https://www.themoviedb.org/movie/256961 ... mall-cop-2

a search should be like this:
try complete name first "Paul Blart Mall Cops 2"
2nd Paul... if over 80% match, display
3th Blart... if over 80% match, display
4th Mall... if over 80% match, display
5th Cops... if over 80% match, display

if 100% match use
if less then 100% match display all matches down to 80%


btw: my sniffer found this query just after the real query (from a clean download):
http://api.themoviedb.org/3/search/movi ... KEY%20HERE

[rednoah]

1.
Looks like that one letter makes a big difference. "Paul Blart Mall Cops 2" contains "Paul" and "Mall" but "Paul Blart Mall Cop 2".

GOOD: https://www.themoviedb.org/search?query ... t+mall+cop
BAD: https://www.themoviedb.org/search?query ... +mall+cops

While doing a search for just "paul blart mall" might work in this very specific case, it doesn't generalise well for all the other spelling-related issues. Plus I have serious concerns for changing "1 query" to "N queries".

This specific issue can be easily be fixed by adding "Paul Blart Mall Cops 2" to the list of alternative titles.

2.
FileBot will use file.name and folder.name to guess queries. Apparently you put some of your files in a folder called "FileBot 4.6 portable".

[kim]

if "N queries" is a NO GO, how about:

1. if 0 results then search / look up the NON plural version of the word / ignore that word...
and do a new search on the other words if not 0 words

Spelling Plural Nouns
http://www.factmonster.com/ipka/A0886509.html

http://www.howtospell.co.uk/pluralrules.php
http://www.oxforddictionaries.com/words ... s-of-nouns


it looks like TMDb still finds a movie if e.g. removed the S from "Sneakers" to "Sneaker"

2. and it to may be a good idea to leave out the year if 0 results... and do a locale compare to results
closest to file/folder year, display/select 1st

http://api.themoviedb.org/3/search/movi ... n&api_key=
http://api.themoviedb.org/3/search/movi ... n&api_key=
http://api.themoviedb.org/3/search/movi ... n&api_key=

[rednoah]

1.
"Sneaker" is a substring of "Sneakers", but "Sneakers" isn't substring of "Sneaker". Technically problems aside, I don't intend to add such "language specific" features just because it happens to work in a few specific cases. Spelling mistakes (which are far more common), alternative wording (common for translated titles), UK/US spelling differences and lots of other language specific issues still won't work. You also have to keep in mind than whenever you make things more flexible to include a few extra true positives, you'll also get potential false positives along as well (plus for unattended automated setups it's extremely important that the false positive rate is as low as possible).

This change would affect all users all the time, but the movies that are positively affected by this change is just a handful. In which case we may as well fix it by adding alias titles to TheMovieDB specifically for these problematic cases, without potentially messing up everything else.

EDIT: added alternative title so this case is closed (until a new example of this issue pops up, if ever)

2.
What's the issue? If there's some file that doesn't work I'd appreciate the path.

You can guess a little bit what's going on by capturing http traffic, but there's ALOT more client-side processing going on that you can't see.