Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

All agents except IMDB TV, are failing to match when TV folder has (year) in it #1192

Closed
blade316 opened this issue Jan 17, 2021 · 8 comments · Fixed by #1195
Closed

All agents except IMDB TV, are failing to match when TV folder has (year) in it #1192

blade316 opened this issue Jan 17, 2021 · 8 comments · Fixed by #1195
Labels
bug: unexpected behavior Not a bug but unexpected and therefore an issue component: TV shows TV shows, series, episodes
Milestone

Comments

@blade316
Copy link

Scrapers that don't work:

  • [*] TVMaze
  • [*] TMDb
  • [*] TheTvDb

Details which are wrong / not loaded:
All of my TV show folders have the year in the name e.g. Westworld (2016), 24 (2001), The Blacklist (2013) etc
I have just started using MediaElch and found that almost all my TV shows fail to match when the year is in the search.
All metadata agents except IMDB TV fail to match unless I remove the year from the search text.

e.g. Westworld (2016) won't match, but Westworld will.

I could just use IMDB TV, however it doesn't scrape a lot of the other images and fanart, so this means that if I want MediaElch to go out and automatically scrape all my TV shows AND get the extra fanart and images, I have to use a scraper other than IMDBTV, which means most of the shows will fail to match.

MediaElch Version:

  • [*] 2.8.5-dev (nightly)
  • [*] 2.8.4 stable

Operating System:

  • [*] Windows
  • [*] macOS
  • [*] Linux (distribution: Docker - UnRAID)

Additional context:

Here is some screens

IMDB TV matches

image

TBDb TV fails with year

image

image

TMDb TV matches without year

image

image

The TVDb and TVMaze also fail on most TV shows when the year is in the folder name.

I am hoping this could be fixed, so the automated searching works properly on scrapers that also allow the extra artwork at the same time.

@ticao2
Copy link

ticao2 commented Jan 17, 2021

Hi @blade316
I hadn't used the TV Scraper in a long time.
With these new changes I decided to update my Series.
I noticed this behavior different from Scraper for Movies.
I didn't even consider it very important. But it seems to you it is. :-)
I already commented with @bugwelle . I believe it will be resolved.
Watch the conversation at #1176

@blade316
Copy link
Author

Hi @blade316
I hadn't used the TV Scraper in a long time.
With these new changes I decided to update my Series.
I noticed this behavior different from Scraper for Movies.
I didn't even consider it very important. But it seems to you it is. :-)
I already commented with @bugwelle . I believe it will be resolved.
Watch the conversation at #1176

Thanks so much for the quick reply @ticao2 .... I will keep an eye on that then, fingers crossed! :)

@ticao2
Copy link

ticao2 commented Jan 17, 2021

@blade316 , in fact, I believe that clarification is necessary.
In Movies, although we can request with the year in the title, I believe that the search is in fact done without the year.
I realize this because in several cases in the search result we also received titles with another year of release.
But, just for our comfort, we are watching the year of the film we have and so we can more safely choose the data we receive.

@blade316
Copy link
Author

@blade316 , in fact, I believe that clarification is necessary.
In Movies, although we can request with the year in the title, I believe that the search is in fact done without the year.
I realize this because in several cases in the search result we also received titles with another year of release.
But, just for our comfort, we are watching the year of the film we have and so we can more safely choose the data we receive.

OK no worries, well let me know if you need any further information from me, or want me to test or clarify anything 👍

@bugwelle
Copy link
Collaborator

In Movies, although we can request with the year in the title, I believe that the search is in fact done without the year.

That is partially true. The TMDb scraper checks for a certain pattern and extracts the year if it was provided in the form of Title (2021), Title 2021 or Title - (2021).
IMDb on the other hand does not extract the year. IMHO MediaElch should simply remove the year from the search query when the dialog opens. Users can add it again for e.g. TMDb. This should also be done for the multi-movie-scraper.

I'll adapt the TV scraper dialog so that the year is removed. Thank you very much for investigating into this. 😄

@ticao2
Copy link

ticao2 commented Jan 19, 2021

In the case of TMDb, for Movies, in the UI I see the year of the Movie but in API Request the year is not used.
Is this method that will be implemented for TV Show? That is, TV Show equal to Movie?
I know that IMDB accepts the year for the survey. The others I don't know if they can handle the year.
Now, just to complicate your work a little: :-)
I believe that 1 year ago TMDb incorporated the possibility of using the year as a parameter in the API Request.
It's the &primary_release_year= parameter.
The API Request would look like this:
https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&language=pt-BR&query=Ad+Astra&primary_release_year=2019
I don't know if it is necessary to make this change for TV Show and/or Movies.
For me personally I see no need.
Of course, there are cases where the Movie Title is just a common word.
In fact, there are 2 worse cases, which I know because I suffered from them.
M (1931) - https://www.themoviedb.org/movie/832-m?language=en-US
total_pages 111 - total_results 2219
Z (1969) - https://www.themoviedb.org/movie/2721-z?language=en-US
total_pages 90 - total_results 1796
So, in these cases it is even better to locate the film and use the TMDb ID in the search.
That's what I did at the time.


EDIT
For TV Show the new parameter is &first_air_date_year=
https://api.themoviedb.org/3/search/tv?api_key=THE_KEY&language=pt-BR&query=House&first_air_date_year=2004
In both cases, if you do not have the year, the API Request can be sent with the parameter without the value, blank.

@bugwelle
Copy link
Collaborator

In the case of TMDb, for Movies, in the UI I see the year of the Movie but in API Request the year is not used.
Is this method that will be implemented for TV Show? That is, TV Show equal to Movie?

TMDb for movies already uses the year as far as I have tested. For TV shows I'm unsure whether we should implement it? How accurate is TMDb for TV shows? I will have to test that first. :)

I know that IMDB accepts the year for the survey. The others I don't know if they can handle the year.

Good to know. It may be useful for the multi-movie scraper.

Now, just to complicate your work a little: :-)

😄 👍

I believe that 1 year ago TMDb incorporated the possibility of using the year as a parameter in the API Request.
It's the &primary_release_year= parameter.
The API Request would look like this:
https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&language=pt-BR&query=Ad+Astra&primary_release_year=2019

We use year for movies. I don't really see the difference between the two. Do you know more?
https://developers.themoviedb.org/3/search/search-movies

I don't know if it is necessary to make this change for TV Show and/or Movies.
For me personally I see no need.

Agree. But I'm interested in what primary_release_year means. :-)

Of course, there are cases where the Movie Title is just a common word.
In fact, there are 2 worse cases, which I know because I suffered from them.
M (1931) - https://www.themoviedb.org/movie/832-m?language=en-US
total_pages 111 - total_results 2219
Z (1969) - https://www.themoviedb.org/movie/2721-z?language=en-US
total_pages 90 - total_results 1796
So, in these cases it is even better to locate the film and use the TMDb ID in the search.
That's what I did at the time.

Oh, that's a very good example! I tested it using Z with year and primary_release_year but the first result is the same... :/

@ticao2
Copy link

ticao2 commented Jan 19, 2021

by Google Translate.
Certainly with some mistake.

TMDb for movies already uses the year as far as I have tested.

I always thought I didn't use it.
Maybe because ME is using only &year= and not &primary_release_year=.
The difference you can see below in the last item.

How accurate is TMDb for TV shows? I will have to test that first. :)

I don't have much personal experience with TV Show. My collection is small.
Year problems for films exist between IMDB and TMDb.
This is because the criteria for determining the Year are not the same.
IMDB uses the first screening, including festivals.
TMDb uses the first screening in movie theaters, open to the public.
Therefore, in some cases, the film participated in the Cannes festival in 1969 but had its first public screening in 1970.
Then there will be this difference between IMDB (1969) and TMDb (1970).

But in the case of TV Show there is no such problem. This is because the criteria of the two are the same.
And the big reference is TheTvDB.
I don't know of any difference. But I am not a reliable source. Very small collection.

Now, just to complicate your work a little: :-)

😄 👍

We use year for movies. I don't really see the difference between the two. Do you know more?
Agree. But I'm interested in what primary_release_year means. :-)

The TMDb has 3 date parameters:

  • release_date
  • primary_release_date
  • primary_release_year

For TMDb Year is any regional, local release_date, premiere date, in any country in the world.
The same film has released_dates in as many countries as it was shown.
So, Year is the year of a release_date.
But primary_release_date is the world premiere date, the first country to show the film.
The year of this date is the year used as the opening year in the world and shown in brackets on the film page.
So primary_release_year is the year of a primary_release_date.

Oh, that's a very good example! I tested it using Z with year and primary_release_year but the first result is the same... :/

First result or first page of results?

M (1931)
https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&language=pt-BR&query=M
Sem o ano = total_pages 111 - total_results 2219
https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&language=pt-BR&query=M&primary_release_year=1931
Com o ano = total_pages 1 - total_results 3

Z (1969)
https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&language=pt-BR&query=Z
Sem o ano = total_pages 90 - total_results 1796
https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&language=pt-BR&query=Z&primary_release_year=1969
Com o ano = total_pages 1 - total_results 8

The difference is very big. But this case is exceptional.
It can be useful, for example for Cosmos (1980) and Cosmos (2014).

Or... :-)
All the King's Men (1949)
All the King's Men (2006)

The Front Page (1931)
The Front Page (1974)

Little Women (1918)
Little Women (1933)
Little Women (1949)
Little Women (1994)
Little Women (2018)
Little Women (2019)

Chocolat (2000)
Chocolat (2016)

Dumbo (1941)
Dumbo (2019)

This week I helped someone with a problem like that.
Movie title with only one common word.
https://www.themoviedb.org/talk/5ff17aad1fb94f003ef540f4

@bugwelle bugwelle added this to the v2.8.6 milestone Jan 21, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug: unexpected behavior Not a bug but unexpected and therefore an issue component: TV shows TV shows, series, episodes
Projects
None yet
3 participants