Kodi Community Forum

Full Version: Problems with Scraper (TMDB) - Scraping wrong movie titles
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I completely reworked my MediaBoxes (with Kodi and CoreElec). However, several times I have already noticed that the scrapper (TMDB, German) scrapes the wrong film titles, although the folder and file names correspond exactly to the german spelling of TMDB!
This is particularly noticeable if two identical film covers suddenly appear next to each other under "Movies". Here are only a few examples where the scrapper is not working correctly:
 
  1. 2018/Apostle (2018)/Apostle (2018).mp4
    2018/Paulus, der Apostel Christi (2018)/Paulus, der Apostel Christi (2018).mp4

    Both are scrapped as... "Paulus, der Apostel Christi" (2018)


  2. 2018/Solo (2018)/Solo (2018).mp4
    2018/Solo - A Star Wars Story (2018)/Solo - A Star Wars Story (2018).mp4
     
    Both are scrapped as... "Solo - A Star Wars Story" (2018)


  3. 2016/Viking (2016)/Viking (2016).mp4
    2016/Viking Legacy (2016)/Viking Legacy (2016).mp4

    Both are scrapped as...  "Viking Legacy" (2016)


  4. 2017/Mountain (2017)/Mountain (2017).mp4
    2017/Zwischen zwei Leben (2017)/Zwischen zwei Leben (2017).mp4

    Both are scrapped as... "Zwischen zwei Leben" (2017)  <-- TMDB english original title: "The Mountain Between Us", which is not the movie "Mountain" (2017)


  5. 2014/Guardians of the Galaxy (2014)/Guardians of the Galaxy (2014).mp4
    2017/Guardians of the Galaxy Vol. 2 (2017)/Guardians of the Galaxy Vol. 2 (2017).mp4

    Both are scrapped as... "Guardians of the Galaxy Vol. 2" (2017)


  6. 2018/Shadow (2018)/Shadow (2018).mp4

    There are two movies from 2018 with the same title... The scrapper grabs directly the first one - without asking the user which is the correct / wanted one.


It seems to me that the scrapper will unfortunately - as soon as only the first word in the title matches one of many other titles on TMDB - incorrectly take this as a hit and transfer it to the Kodi film database.
Even with film titles, where there are two different films in the same year - but with the same title, the scrapper apparently always takes the first hit directly and incorrectly.

My wish would be that the scrapper really take the full file / folder movie name and check for an exact match instead of just the first word.
I would also very much like the scrapper to show the user a selection of what the original file name is in the event of two or more identical hits from TMDB - and which matching hits were found. So the user can choose the right film directly.
So that the user is not constantly asked during scrapping, the scrapper could theoretically first process everything that is unambiguous - and display all ambiguous films to the user at the end of the scrapping that has been completed so far.

PS:
In the meantime I accidentally discovered some other films in my library that were assigned the wrong film titles / covers. (Unless two identical film covers are displayed side by side; this is very difficult to determine and fix)



Finally, another mysterious problem with the scraper in the following film ...
  1. /2013/Dinosaurier 3D - Im Reich der Giganten (2013)/Dinosaurier 3D - Im Reich der Giganten (2013).mp4

    This is exactly the german title as in TMDB... The scrapper will not scrape this movie nor seems to write an error to the logfiles.
    The scrapper recognizes this movie with its german title only, when renaming the folder- and file to "2013/Walking with Dinosaurs (2013)/Walking with Dinosaurs (2013).mp4"
    Very, very strange - is there any explanation for this? Seems to be a bug to me too
Hi guys,

i found one more movie, where the folder and file names correspond exactly to the german spelling of TMDB! https://www.themoviedb.org/movie/430155
Again the scrapper does not regonize the movie.
  1. 2019/Coma (2019)/Coma (2019).mp4
This movie will only be added to the movie database, when renaming to the orginal RUSSIA movienames using RUSSICA CHARSETS - wtf?
So i renamed it like flollowing and the movie scrapper directly found and added this movie to the db.
  1.  2019/Кома (2019)/Кома (2019).mp4
Would be really nice, if this scrapper problems / bugs may be fixed soon. thanks in advance.
and one more movie like the problem #2 ...
  1. 2017/Attraction (2017)/Attraction (2017).mp4 

    This fails.... even its indentical to the german title - https://www.themoviedb.org/movie/401513/...mary_facts

    This is the russia movie name: Притяжение

What the hell is wrong with the scrapper or TMDB that it does not works as it should?
@va!n

We need a Debug Log which captures the problem. Thanks.

A scraper is never 100% accurate. But there are many ways to fix wrong entries. Using Refresh and either picking the correct movie from the offered list, or adding the ID as shown in the images here... https://kodi.wiki/view/Add-on:The_Movie_...hon#Search

But, until you provide a proper Debug Log, there is nothing for us to do.
@Karellen:
Thanks for your reply. The problem about the logs is following:
  1. Movies scrapped as wrong movie title   (I only coincidence noticed this now)
  2. Movies that are not scrapped (because the folder- / filename is apparently not recognized and is skipped) are not captured in the logfiles.

    First when i noticed that the comparsion between the number of movies inside a year folder on the hdd didn't matches to the number of movies of the same year in the kodi movie db, i noticed that there is something wrong.
    Since there were in these cases unfortunately no entries in the logs, it was quite difficult and a lot of work for me to come across this bug.
@va!n

You haven't told me anything of use. I don't need you to further describe what you are experiencing. I need the log that captures the problem.

You won't hear from me again until I see you post the proper log.