Scores Reported as ZERO (2 Viewers)

[Big_Chubbz]

Here's a snip from my log. Does this mean that I'm over my call limit and eventually everything will be OK, or is something still messed up? I'm using 4.9.60 at this point.

10-Feb-2016 19:58:33 Info [ MovieManagerPane]: Refreshing movie #1: 99 Homes
10-Feb-2016 19:58:33 Info [ ScraperNode]: ############################ IMDb+ Detail Node ############################
10-Feb-2016 19:58:33 Info [ ScraperNode]: IMDb+ Scraper: Settings overruled via Options XML file
10-Feb-2016 19:58:34 Warn [ ScraperNode]: Could not connect to http:. Invalid URI: The Authority/Host could not be parsed.
10-Feb-2016 19:58:35 Info [ ScraperNode]: ###########################################################################

 

[RoChess]

hmmm, that's an error that could mean more, but I would need to see what URI it was trying to connect to.

You would have to give me scraper-debug results that give me that extra bit of info, because I won't be able to test myself until this weekend.

[Edit]: Actually what I think happened is that the initial API call to get RottenTomatoes http://address from the IMDb tt-ID is rejected, which then gives me no 'address', and then the following call to read the page only ends up reading `http:` which then causes MovPic to freak.

Scraper-debug log file would give me a better idea then.

But yes, in my initial testing I found that the next day RT works fine again for those first few calls, but the problem is that you have to manually tell MovPic to refresh the movie so that IMDb+ can try to fetch it again from RT. Now most people do not add massive amounts of movies to their library, but a couple here-and-there.

So it is only the initial mass import that poses an issue, at least from my initial testing, but RT might have already tightened the rules on that.

 

[Big_Chubbz]

Thanks again, @RoChess.

Here's the more detailed log. I'm not sure whats going on, because I've been trying to manually refresh these movies with the "zero scores" for a couple of days now, and they just won't populate with the correct data. I'll try once just after midnight, then again before bed around 4am , and once more after I wake up around lunchtime, but MovingPictures just won't grab that data.

I'm trying to refresh just one movie, and I've tried completely deleting, going to 'manage hidden' and re-importing as well as simply clicking 'refresh movie info from internet' (as I did for the attached log), but neither method has been successful.

 

[RoChess]

I think you forgot to enable Moving-Pictures scraper-debug logging with the green little bug icon shown.

But I am going to try to reproduce it myself.

 

[RoChess]

Okay, found cause, and added error-handling code to IMDb+ in v4.9.61 to deal with it.

"99 Homes" @ IMDb = http://www.imdb.com/title/tt2891174/
"99 Homes" @ RT = http://www.rottentomatoes.com/m/99_homes/

RottenTomatoes IMDb tt-ID lookup that IMDb+ is forced to make to locate RT URL = http://api.rottentomatoes.com/api/public/v1.0/movie_alias.json?type=imdb&id=2891174

And that is when we see the problem occur, because RT then gives me an error back that it does not recognize that movie.

So somebody has to go into RT, and edit http://www.rottentomatoes.com/m/99_homes/ to add IMDb tt-ID reference.

IMDb+ has no other option but to fail with 0.0 score until then, but at least movingpictures.log will have a nice new message added now saying:

IMDb+ Scraper: RottenTomatoes is missing a reference for: ${movie.imdb_id} (consider contributing it)

 

[Big_Chubbz]

Thanks! I'm guessing it's the same story for the rest of the movies that aren't working. How can I contribute that data to RT?

 

[RoChess]

I see others have reported bad IMDb aliases on the forum: http://developer.rottentomatoes.com/forum/read/162242

But that is an old post, so not sure if that is still the acceptable method.

Their FAQ has this to say -- http://flixster.desk.com/customer/en/portal/articles/166030-your-movie-actor-information-is-wrong-

I think signing up at forum might be the best approach though, because it relates to the API side, which is developer stuff.

PS: You want to make sure the others are wrong as well though, so in a browser open: http://api.rottentomatoes.com/api/public/v1.0/movie_alias.json?type=imdb&id=2891174 with "2891174" being the IMDb tt-ID without the "tt".

 

[Big_Chubbz]

Thanks again for all your help. I double checked, and you are right. 13 out of 35 of the movies in my DB are unfortunately not aliased properly in RT. All 13 report:

{"error":"Could not find a movie with the specified id"}

They are major movies (e.g. Chi-Raq), so it is surprising. I appreciate you taking the time to completely solve this issue and figure out precisely what is happening.

FWIW, I found this when I went over to RT's API forums to see about fixing those 13 films:

FlixsterCS – 1 year ago

@jackripper220 - We do provide some IDs in our APIs, like IMDB IDs, to help with matching but ultimately our database of IMDB IDs is not reliable for matching. We recommend matching based on title, director, year, etc. in order to avoid these errors.
​

Something to consider if these types of errors become more prevalent. It's already affecting over a third of my database, and my collection auto pulls from

https://itunes.apple.com/us/rss/topvideorentals/limit=10/xml

So not a lot of obscure titles.

 

[RoChess]

IMDb has a perfect database. I can understand RottenTomatoes wanting to promote their system, but thousands of movies have an IMDb tt-ID reference, and a correct one as well. The problem with switching from the quick IMDb tt-ID lookup with RT is that I then have no guarantee that the movie is the correct one, because searching on title, director, year, etc leaves a lot of room for error. It is very easy to search on IMDb and get 100 movies with the same title, or dozens with same title in the same year. And what if IMDb uses 2015 as year, but RT uses 2016 then I would fail as well.

I have an extensive amount of code to do the search like that with IMDb search engine, and that is the only time a user can confirm/correct it as well, after that I would have to rely on fuzzy results, which is the whole point of the API to search by a fixed known identifier such as IMDb tt-ID.

PS: Do you have a list of the movies? If RT now decides to stop adding IMDb tt-ID on well known releases then it would indeed be a problem. Probably dealing with a lazy RT editor, which is why I do not understand they do not adapt to allow community contributions like IMDb does.

 

[Big_Chubbz]

I completely understand. Just from the little bit of interaction I've had with RT over the last week, I'm not too keen on the way they operate. I'm still waiting on approval of my API key, and if I read their application properly, they are going to charge me for a key if I'm approved. They killed their public RSS feeds a while back (the reason I switched to the itunes feeds), and now this issue.

I do love those audience ratings, though. I find them to be the best indicator of the straight entertainment value of a film. Too many times a critic's darling will be showered with praise, and when you watch it, it's a real stinker!

Anyway, I attached a redacted log file with the list of movies that don't have the correct entries in their database. Thanks again, and keep up the awesome work!