• 1
  • 140
  • 141
  • 142(current)
  • 143
  • 144
  • 197
[RELEASE] Texture Cache Maintenance utility
Again thank you very much for the response.

(2016-12-12, 01:25)Milhouse Wrote: Can you provide the before and after poster artwork urls for one of your movies that loaded a thumbnail?

Although if a thumbnail is replacing a poster, that's down to the scraper you are using - the script is only telling kodi to remove the movie, then perform a video rescan (library update). Whatever is being scraped in is down to Kodi and the scraper you are using, the script has no involvement in what metadata is created during a library update.

I don't think there's an url, because it's all done manually and offline. Basically my process is that I make the folder for the movie, add a poster.jpg file, and then launch Kodi. Then, Kodi adds it with a scrapper called the 'Stupid Movie Scraper', which merely adds videos under their filename, and scraps nothing (later the .nfo is created through the process explained above, in which they're exported next to the poster.jpg file). This scraper works fine for posters, and it only ever uses the thumbnail as one when there is no poster.jpg file... any ideas on what's happening?

(2016-12-12, 01:25)Milhouse Wrote: 1. No. Movies (and TV show episodes) fail QA for a number of configurable reasons (defaults being: missing artwork, missing plot, missing mpaa). Movies/Episodes will be removed/rescanned whenever they fail QA
Wait, I don't know what QA is, but what I gather from this is that this whole update thing can be triggered if something inside the .nfo is missing, like <mpaa></mpaa>, right? This is such a stretch, but I'm gonna ask anyway: can it do the opposite, as in, detect that an <update></update> tag is present, and then update those? This would only really be useful if there was a way to automate the removal of this tag after updating though.

(2016-12-12, 01:25)Milhouse Wrote: 2. I'll need to see the before/after artwork urls (texturecache.py jd movies <somemoviename>). If you had a remote poster, and a local thumbnail, then it's possible the scraper will use the local thumbnail and replace the poster during remove/rescan. However without knowing what you had before, and what you have now, I can't really say what or why the scraper has replaced the poster with a thumbnail.

Oh, that's what you mean with url, ok. Well I ran the command with a movie that got the poster replaced (which really is all of them minus the ones that are foldered DVD/Blu-ray... like video_ts and such)

Code:
C:\Program Files (x86)\Kodi (Video)\portable_data\userdata>texturecache.py jd mo
vies Teorema
[
  {
    "art": {
      "thumb": "image://video@M:\\Media\\Video\\Single\\Teorema\\Teorema.mkv/"
    },
    "file": "M:\\Media\\Video\\Single\\Teorema\\Teorema.mkv",
    "label": "Teorema",
    "movieid": 593,
    "title": "Teorema"
  }
]

I tried with the DVD/Blu-ray ones, but they don't seem to be recognized (so probably they never updated to begin with... bug?)

There's no phyisical local thumbnail by the way. All the files each folder has is the video, the .nfo, and the poster.jpg. Not sure what a remote poster is, but I don't think it applies (sounds like it's for people who have Kodi or the videos separated on a different hardware/network, while I run both videos and Kodi from the same computer).
Reply
(2016-12-12, 06:57)firewater Wrote: Wait, I don't know what QA is, but what I gather from this is that this whole update thing can be triggered if something inside the .nfo is missing, like <mpaa></mpaa>, right? This is such a stretch, but I'm gonna ask anyway: can it do the opposite, as in, detect that an <update></update> tag is present, and then update those? This would only really be useful if there was a way to automate the removal of this tag after updating though.

A refresh will be triggered if a configured field (such as mpaa, or plot) is not populated in the media library database.

The script has no idea if the NFO contains an mpaa or plot tag, or even if an NFO exists - it will simply refresh the movie if the movie (as it exists in the media library) fails QA (Quality Assurance). By refreshing the movie, the hope is that any NFO file that may exist will include the missing fields, or the relevant information will be scraped from the internet.

Adding an <update> tag to the NFO will have no effect as the script doesn't look at the contents of the NFO.

(2016-12-12, 06:57)firewater Wrote: Oh, that's what you mean with url, ok. Well I ran the command with a movie that got the poster replaced (which really is all of them minus the ones that are foldered DVD/Blu-ray... like video_ts and such)

Code:
C:\Program Files (x86)\Kodi (Video)\portable_data\userdata>texturecache.py jd mo
vies Teorema
[
  {
    "art": {
      "thumb": "image://video@M:\\Media\\Video\\Single\\Teorema\\Teorema.mkv/"
    },
    "file": "M:\\Media\\Video\\Single\\Teorema\\Teorema.mkv",
    "label": "Teorema",
    "movieid": 593,
    "title": "Teorema"
  }
]

I tried with the DVD/Blu-ray ones, but they don't seem to be recognized (so probably they never updated to begin with... bug?)

There's no phyisical local thumbnail by the way. All the files each folder has is the video, the .nfo, and the poster.jpg. Not sure what a remote poster is, but I don't think it applies (sounds like it's for people who have Kodi or the videos separated on a different hardware/network, while I run both videos and Kodi from the same computer).

A "remote poster" is a poster image (artwork) scraped from fanart tv, imdb or similar web site, assuming the scraper supports such sites.

The scraper you are using is extracting a video screen grab from the video, presumably because it can't find a suitable poster file (remote, or local).

I suggest you contact the scraper maintainer for support on why the scraper is not finding and using the local poster.jpg artwork.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Reply
(2016-12-12, 21:29)Milhouse Wrote: A "remote poster" is a poster image (artwork) scraped from fanart tv, imdb or similar web site, assuming the scraper supports such sites.

The scraper you are using is extracting a video screen grab from the video, presumably because it can't find a suitable poster file (remote, or local).

I suggest you contact the scraper maintainer for support on why the scraper is not finding and using the local poster.jpg artwork.

Oh, okay. And no, it doesn't support anything like that. It literally just adds a video under its name (I think it uses the folder name for that), and that's as far as it can do.

I think the behaviour for 'make a thumbnail if no poster exists' is a Kodi function. Don't think scrapers have any say on that part... like, if someone wanted to code a scraper that didn't do anything if no poster is found, I don't think that's possible because Kodi as a program does the thumbnail extraction as a separate thing.

And I'm pretty sure it's something about this script, or maybe some setting of mine, because again the scraper always detects the posters in other scenarios, and because I just tried removing my library, and then readding it with the Kodi default .nfo scrapper (not the same as the 'Stupid Movie Scraper'), and then running the texturecache.py script, and the same thing occurred. So is there any command that updates the artwork for these updated .nfos? I could run it along with the other one and it'd be fixed I think.
Reply
(2016-12-12, 22:16)firewater Wrote: I think the behaviour for 'make a thumbnail if no poster exists' is a Kodi function. Don't think scrapers have any say on that part... like, if someone wanted to code a scraper that didn't do anything if no poster is found, I don't think that's possible because Kodi as a program does the thumbnail extraction as a separate thing.

Yes, it's Settings > Media settings > Videos > Extract thumbnails from video files.

Kodi is only extracting the thumbnails because your scraper isn't setting any other suitable artwork (ie. a poster).

(2016-12-12, 22:16)firewater Wrote: And I'm pretty sure it's something about this script, or maybe some setting of mine, because again the scraper always detects the posters in other scenarios, and because I just tried removing my library, and then readding it with the Kodi default .nfo scrapper (not the same as the 'Stupid Movie Scraper'), and then running the texturecache.py script, and the same thing occurred.

The script has nothing to do with what metadata is actually scraped. The script will delete the movie, then instruct Kodi to perform a library update which should hopefully re-scrape the movie that has just been deleted. This should produce the same results as the "Refresh" button on the movie Info page. How and what Kodi scrapes is entirely down to the scraper you are using for the source being scraped, and how you have configured Kodi (eg. extract thumbnails).

Can you provide metadata details of a movie (ie. "jd movies <moviename>") before running "texturecache.py qax", and the metadata after qax, and also the directory contents for the movie? A debug-log from kodi showing the movie scan would also be helpful.

(2016-12-12, 22:16)firewater Wrote: So is there any command that updates the artwork for these updated .nfos? I could run it along with the other one and it'd be fixed I think.

If your artwork is correctly named, and is supported by the scraper, then it should be found and set on each movie automatically. Note that most Kodi scrapers only support limited artwork types (eg. fanart, poster) - if you need anything more exotic than that you'll need an add-on such as Artwork Downloader, or mklocal.py script.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Reply
Thanks for the clarifications.

(2016-12-12, 22:43)Milhouse Wrote: Can you provide metadata details of a movie (ie. "jd movies <moviename>") before running "texturecache.py qax", and the metadata after qax, and also the directory contents for the movie? A debug-log from kodi showing the movie scan would also be helpful.

Right, so I rescanned with the Stupid Movie Scraper, and then

Code:
C:\Program Files (x86)\Kodi (Video)\portable_data\userdata>texturecache.py jd mo
vies Amour

[
  {
    "art": {
      "poster": "image://M:\\Media\\Video\\Single\\Amour\\poster.jpg/"
    },
    "file": "M:\\Media\\Video\\Single\\Amour\\Amour.mkv",
    "label": "Amour",
    "movieid": 872,
    "title": "Amour"
  }
]

This folder is "M:\Media\Video\Single\Amour". The contents are "Amour.mkv", "Amour.nfo", and "poster.jpg". The poster is being displayed.

Then

Code:
texturecache.py qax movies @qaperiod=-1 @qa.nfo.refresh=30

This outputs a list that begins with lines like:

Code:
t; modified nfo; missing mpaa), WARN (missing fanart, local not found; missing p
oster, local not found)
Movie    [La Historia Oficial                               ]: FAIL (missing plo
t; modified nfo; missing mpaa), WARN (missing fanart, local not found)
Movie    [La Pianiste                                       ]: FAIL (missing plo
t; missing poster, local is available; modified nfo; missing mpaa), WARN (missin
g fanart, local not found)

"La Historia Oficial" is a foldered movie (the ones that do have the poster displaying after the script). Note that it makes no mention of the poster.
"La Pianiste" doesn't have a poster after this refresh (it has it before, it's a "poster.jpg", as always).

And then ends with

Code:
Updating library: New movieid     977 [A Collection]
Rescanning directory: M:\Media\Video\Single\Accattone...
Updating library: New movieid     978 [Accattone]
Rescanning directory: M:\Media\Video\Single\Aguirre, der Zorn Gottes...
Updating library: New movieid     979 [Aguirre, der Zorn Gottes]
Rescanning directory: M:\Media\Video\Single\Amargos...
Updating library: New movieid     980 [Amargos]
Rescanning directory: M:\Media\Video\Single\Amour...
Updating library: New movieid     981 [Amour]

Now, everything has the thumbnail as the poster, except for the foldered movies.

Code:
C:\Program Files (x86)\Kodi (Video)\portable_data\userdata>texturecache.py jd mo
vies Amour
[
  {
    "art": {
      "thumb": "image://video@M:\\Media\\Video\\Single\\Amour\\Amour.mkv/"
    },
    "file": "M:\\Media\\Video\\Single\\Amour\\Amour.mkv",
    "label": "Amour",
    "movieid": 1148,
    "title": "Amour"
  }
]

Here's kodi.log. Note that I had to run the script twice so I could copy the first type of line (they disappear after completion because of so many Updating lines in the end)

This is the entire content of the Stupid Movie Scraper, by the way:

Code:
<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<scraper name="Stupid Movie Database" content="movies" thumb="icon.png" framework="1.1" language="en">
    <CreateSearchUrl dest="3" clearbuffers="no">
        <RegExp dest="3" output="&lt;url&gt;http://search.yahoo.com/search?p=$$7&lt;/url&gt;" input="$$7">
            <RegExp dest="7" output="\1" input="$$1">
                <expression noclean="1" trim="1">(.+)</expression>
            </RegExp>
            <expression noclean="1"></expression>
        </RegExp>
    </CreateSearchUrl>
    <GetSearchResults dest="8" clearbuffers="no">
        <RegExp dest="8" output="&lt;results sorted=&quot;yes&quot;&gt;\1&lt;/results&gt;" input="$$5">
            <RegExp dest="5" output="&lt;entity&gt;&lt;title&gt;\1&lt;/title&gt;&lt;url&gt;http://search.yahoo.com/search?p=$$7&lt;/url&gt;&lt;/entity&gt;" input="$$1">
                <expression noclean="1" trim="1">&lt;title&gt;([^"]*)\-\sYahoo Search</expression>
            </RegExp>
            <expression noclean="1"></expression>
        </RegExp>
    </GetSearchResults>
    <GetDetails dest="3" clearbuffers="no">
        <RegExp dest="3" output="&lt;details&gt;\1&lt;/details&gt;" input="$$5">
            <RegExp dest="5+" output="&lt;title&gt;\1&lt;/title&gt;" input="$$1">
                <expression noclean="1" trim="1">&lt;title&gt;(.+)\s\-\sYahoo Search Results&lt;/title&gt;</expression>
            </RegExp>
            <expression noclean="1" trim="1"></expression>
        </RegExp>
    </GetDetails>
</scraper>

I have no idea why it references Yahoo Search, since it all runs offline without a need for an Internet connection. I know that the scraper is modified from a version that the creator used for music videos, so presumably he had this Yahoo Search function and didn't bother to remove it.

I can do it again with Kodi's default .nfo scraper in case you want to rule the scraper out.
Reply
(2016-12-13, 19:38)firewater Wrote: "La Historia Oficial" is a foldered movie (the ones that do have the poster displaying after the script). Note that it makes no mention of the poster.
"La Pianiste" doesn't have a poster after this refresh (it has it before, it's a "poster.jpg", as always).

"La Pianiste" doesn't have a poster assigned before the refresh - that's why it failed QA with "missing poster". However the script has identified a suitable local file alongside the movie that could be used if the scraper were able to find it (and apparently it does not).

When you talk about "foldered" movies, do you mean one movie per folder? In which case naming the poster as poster.jpg is correct (and apparently this is working, with no before/after issues).

And your Armour movie, since this isn't "foldered", does this mean it is one of many movies in a single folder? In which case you should be naming the poster Armour-poster.jpg, otherwise all movies in the same folder will have an identical poster (they'll all use the same poster.jpg).

It's not immediately clear why the re-scrape of a many-in-a-folder movie should be different to the original scrape, swapping the originally scraped poster for a thumbnail. It would be interesting to see if the Info page > Refresh button also replaces poster.jpg with a thumbnail for the movie Armour.

However this doesn't look right:
Code:
14:17:41 T:2520  NOTICE: VideoInfoScanner: Starting scan ..
14:17:41 T:2520   ERROR: CScraperParser::Parse: Could not find scraper function NfoUrl

I don't know which url it's trying to parse, but if it's the artwork url it might explain why it is dropping your posters. If the latest Kodi 17 beta 6 still has this error then I'd suggest reporting this to the scraper maintainer, or open a ticket on trac.kodi.tv.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Reply
(2016-12-13, 20:27)Milhouse Wrote: "La Pianiste" doesn't have a poster assigned before the refresh - that's why it failed QA with "missing poster". However the script has identified a suitable local file alongside the movie that could be used if the scraper were able to find it (and apparently it does not).

When you talk about "foldered" movies, do you mean one movie per folder? In which case naming the poster as poster.jpg is correct (and apparently this is working, with no before/after issues).

And your Armour movie, since this isn't "foldered", does this mean it is one of many movies in a single folder? In which case you should be naming the poster Armour-poster.jpg, otherwise all movies in the same folder will have an identical poster (they'll all use the same poster.jpg).

It's not immediately clear why the re-scrape of a many-in-a-folder movie should be different to the original scrape, swapping the originally scraped poster for a thumbnail. It would be interesting to see if the Info page > Refresh button also replaces poster.jpg with a thumbnail for the movie Armour.

However this doesn't look right:
Code:
14:17:41 T:2520  NOTICE: VideoInfoScanner: Starting scan ..
14:17:41 T:2520   ERROR: CScraperParser::Parse: Could not find scraper function NfoUrl

I don't know which url it's trying to parse, but if it's the artwork url it might explain why it is dropping your posters. If the latest Kodi 17 beta 6 still has this error then I'd suggest reporting this to the scraper maintainer, or open a ticket on trac.kodi.tv.

Sorry, I wasn't clear about that so your analysis doesn't apply for the most part. What I meant with foldered is what I mentioned previously about these movies being foldered Blu-ray/DVDs, so instead of being a single .mkv or .avi file, they're contained in folders like VIDEO_TS, or STREAM + MovieObject.

In what sense does La Pianiste not have a poster assigned? Because it's there as poster.jpg, and it's being displayed before the script runs, so what would be the process for Kodi (or the script) to recognize that this means it does in fact have a poster assigned? I guess it doesn't matter much since it says it's being found, but still.

Info > Refresh replaces the thumbnail image with the correct poster.jpg

The scraper has no support. He didn't even make an individual thread for it... someone asked for a scraper doing this and this other guy made a post about his "Stupid Movie Scraper". He also hasn't connected in a while nor responded when I contacted him on Github once, so my only real chance at fixing this is here. And then again I don't think it's the scraper since the official Kodi "Local information only" scraper has the same problem, and it doesn't have the parse line... this is the log after repeating the process with that one.

At last, I've tried the process with the TMDB scraper, and to my surprise, the posters are kept after refresh. It's too bad I have no use for this scraper for a variety of other reasons, so this is not an option for me. Conclusion? I don't know, but there is something in common about Kodi's default .nfo scraper, and the Stupid Movie Scraper that makes individual movies fail, but "foldered" ones work. If you could find what it is, great, if not... should I make a separate thread here on the forums, or submit it as a bug ticket? How do I even propose this problem... ".nfo scrapers have trouble with bulk refreshing movies?" because I feel like the devs of Kodi are gonna tell me that it's not their problem since the refresh is happening with an external script and that Kodi has no feature for bulk refreshing...
Reply
(2016-12-13, 22:20)firewater Wrote: In what sense does La Pianiste not have a poster assigned? Because it's there as poster.jpg, and it's being displayed before the script runs, so what would be the process for Kodi (or the script) to recognize that this means it does in fact have a poster assigned? I guess it doesn't matter much since it says it's being found, but still.

At the time you ran the qax command, "La Pianiste" did not have poster artwork assigned to it in the media library. I can't explain how you saw a poster, are you using a third-party skin or extension that is doing this for you?

(2016-12-13, 22:20)firewater Wrote: Info > Refresh replaces the thumbnail image with the correct poster.jpg

The scraper has no support. He didn't even make an individual thread for it... someone asked for a scraper doing this and this other guy made a post about his "Stupid Movie Scraper". He also hasn't connected in a while nor responded when I contacted him on Github once, so my only real chance at fixing this is here. And then again I don't think it's the scraper since the official Kodi "Local information only" scraper has the same problem, and it doesn't have the parse line... this is the log after repeating the process with that one.

At last, I've tried the process with the TMDB scraper, and to my surprise, the posters are kept after refresh. It's too bad I have no use for this scraper for a variety of other reasons, so this is not an option for me. Conclusion? I don't know, but there is something in common about Kodi's default .nfo scraper, and the Stupid Movie Scraper that makes individual movies fail, but "foldered" ones work. If you could find what it is, great, if not... should I make a separate thread here on the forums, or submit it as a bug ticket? How do I even propose this problem... ".nfo scrapers have trouble with bulk refreshing movies?" because I feel like the devs of Kodi are gonna tell me that it's not their problem since the refresh is happening with an external script and that Kodi has no feature for bulk refreshing...

If you're concerned that there is a Kodi core issue, then you should be able to prove it by removing a movie, and performing a library update via the GUI - this is what the texturecache.py script is doing (using JSON) and similar to what the Refresh button is doing (but you say that works, so results may vary - did you answer yes or no to "Locally stored information found. Ignore and refresh from Internet"? - you should reply "No"). Then open a trac ticket. If you can provide the exact steps to reproduce (including if necessary any nfo contents) then that would help. Keep information to a minimum, don't refresh multiple movies (if that's not part of the problem) - focus on one movie.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Reply
(2016-12-14, 03:32)Milhouse Wrote: At the time you ran the qax command, "La Pianiste" did not have poster artwork assigned to it in the media library. I can't explain how you saw a poster, are you using a third-party skin or extension that is doing this for you?

If you're concerned that there is a Kodi core issue, then you should be able to prove it by removing a movie, and performing a library update via the GUI - this is what the texturecache.py script is doing (using JSON) and similar to what the Refresh button is doing (but you say that works, so results may vary - did you answer yes or no to "Locally stored information found. Ignore and refresh from Internet"? - you should reply "No"). Then open a trac ticket. If you can provide the exact steps to reproduce (including if necessary any nfo contents) then that would help. Keep information to a minimum, don't refresh multiple movies (if that's not part of the problem) - focus on one movie.

Weird. No, just the poster.jpg file. I do use a skin, but it works the way any other skin does (that is, it just displays the poster, it doesn't download anything).

I selected "no", yeah.

Ok so I'm gonna record a video of me doing this:

1. Display the contents of my movie folder in file explorer.
2. On Kodi, show that my library is just 3 movies ("foldered" + 2 of the normal ones), and they display posters just fine.
3. Remove one movie.
4. Show that I'm using the official .nfo scraper.
5. Scan the library for changes from Kodi.
6. The movie is readded correctly.
7. Run the script on cmd.
8. All 3 movies get updated with the wrong poster.

Hope it suffices.
Reply
Also enable texturecache.py @logfile, that way it's obvious what JSON requests are being issued to Kodi - should just be VideoLibrary.RemoveMovie followed by VideoLibrary.Scan.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Reply
What the absolute HECK... I'm so confused... as I recorded the video, the situation has completely changed, for seemingly no reason at all. Now information is being fetched after running the script... posters are being downloaded, plot is being scrapped, even the title of the movies are being adapted according to some scraper.. but I'm still using the local .nfo scraper HuhHuh I even tried with the Stupid scraper and still the same thing occured, things were scraped as if I had TMDB selected Huh ... so to make this clear: I erased my video library, used .nfo / stupid scraper (did it twice, once with each) to add it again, they get added with no plot (the .nfo doesn't have them), the poster.jpg inside their respective folders, and they have their "real" titles (À ma sœur!) - then I run the script, and the movies get their poster fetched from TMDB, get plot, and the title adapted in english (Fat Girl). Is the script fetching all this stuff? Why now? I don't understand anything aaaaaaaaaaah

Code:
Microsoft Windows [Version 6.2.9200]
(c) 2012 Microsoft Corporation. All rights reserved.

C:\Windows\system32>cd C:\Program Files (x86)\Kodi (Video)\portable_data\userdat
a

C:\Program Files (x86)\Kodi (Video)\portable_data\userdata>texturecache.py qax m
ovies @qaperiod=-1 @qa.nfo.refresh=30
Movie    [24 Hour Party People                              ]: FAIL (modified nf
o)
Movie    [31 de Abril-HD                                    ]: FAIL (missing mpa
a; missing plot; missing poster, local is available), WARN (missing nfo)
Movie    [71 Fragments of a Chronology of Chance            ]: FAIL (modified nf
o)
Movie    [199 Tips to Be Happy                              ]: FAIL (missing mpa
a; modified nfo)
Movie    [Accattone                                         ]: FAIL (modified nf
o)
Movie    [Aguirre: The Wrath of God                         ]: FAIL (modified nf
o)
Movie    [Amargosa                                          ]: FAIL (modified nf
o), WARN (missing fanart, local not found)
Movie    [Amour                                             ]: FAIL (modified nf
o)
Movie    [Antichrist                                        ]: FAIL (modified nf
o)
Movie    [Atras De La Ventana-SD                            ]: FAIL (missing mpa
a; missing plot; missing poster, local is available), WARN (missing fanart, loca
l not found; missing nfo)
Movie    [Before Sunrise                                    ]: FAIL (modified nf
o)
Movie    [Benny's Video                                     ]: FAIL (missing mpa
a; modified nfo)
Movie    [Bodysong                                          ]: FAIL (modified nf
o), WARN (missing fanart, local not found)
Movie    [Fat Girl                                          ]: FAIL (modified nf
o)
Movie    [Mouchette                                         ]: FAIL (modified nf
o)
Movie    [Teorema                                           ]: FAIL (modified nf
o)
Movie    [The Collection                                    ]: FAIL (modified nf
o)
Movie    [The Official Story                                ]: FAIL (modified nf
o)
Rescanning directory: M:\Media\Video\Movies\199 Recetas para ser feliz...
Updating library: New movieid      72 [199 Tips to Be Happy]
Rescanning directory: M:\Media\Video\Movies\24 Hour Party People...
Updating library: New movieid      73 [24 Hour Party People]
Rescanning directory: M:\Media\Video\Movies\31 de Abril...
Updating library: New movieid      74 [31 de Abril-HD]
Rescanning directory: M:\Media\Video\Movies\71 Fragmente einer Chronologie des Z
ufalls...
Updating library: New movieid      75 [71 Fragments of a Chronology of Chance]
Rescanning directory: M:\Media\Video\Movies\A Collection...
Updating library: New movieid      76 [The Collection]
Rescanning directory: M:\Media\Video\Movies\Accattone...
Updating library: New movieid      77 [Accattone]
Rescanning directory: M:\Media\Video\Movies\Aguirre, der Zorn Gottes...
Updating library: New movieid      78 [Aguirre: The Wrath of God]
Rescanning directory: M:\Media\Video\Movies\Amargos...
Updating library: New movieid      79 [Amargosa]
Rescanning directory: M:\Media\Video\Movies\Amour...
Updating library: New movieid      80 [Amour]
Rescanning directory: M:\Media\Video\Movies\Antichrist...
Updating library: New movieid      81 [Antichrist]
Rescanning directory: M:\Media\Video\Movies\Atras de la Ventana...
Updating library: New movieid      82 [Atras De La Ventana-SD]
Rescanning directory: M:\Media\Video\Movies\Before Sunrise...
Updating library: New movieid      83 [Before Sunrise]
Rescanning directory: M:\Media\Video\Movies\Benny's Video...
Updating library: New movieid      84 [Benny's Video]
Rescanning directory: M:\Media\Video\Movies\Bodysong...
Updating library: New movieid      85 [Bodysong]
Rescanning directory: M:\Media\Video\Movies\La Historia Oficial...
Updating library: New movieid      86 [The Official Story]
Rescanning directory: M:\Media\Video\Movies\Mouchette...
Updating library: New movieid      87 [Mouchette]
Rescanning directory: M:\Media\Video\Movies\Teorema...
Updating library: New movieid      88 [Teorema]
Rescanning directory: M:\Media\Video\Movies\À ma sœur!...
Updating library: New movieid      89 [Fat Girl]

C:\Program Files (x86)\Kodi (Video)\portable_data\userdata>

edit: I think I know that's happening... its a guess, at least. Remember how I did switch to TMDB for the test I did some posts above? I think Kodi has all these plot, localized titles, fanart, and posters information from TMDB scraper, still stored on its cache - and it's getting triggered and taking over when the script runs. Any way to completely delete Kodi's cache and see if it's that? I'm out of my league here, but another guess is that before I did this, and the cache didn't have TMDB info, the cache had nothing so it'd create the thumbnails... so to illustrate what I mean

before:

kodi's cache is empty > running the script appeals to the cache > all this non-info from the cache gets pasted into the movies > movies with poster now have no poster because the poster cache was empty, and kodi's behaviour is to make a thumbnail when there's no poster

now:

kodi's cache is filled with tmdb info > running the script appeals to the cache > all this info gets pasted into the movies > all movies have tmdb info
Reply
The script does not fetch *anything*! All the script does is automate the otherwise manual GUI process of refreshing individual movies by removing movies from the library using JSON, and then instructing Kodi to rescan the library for new videos (in particular the movies just deleted, so that they are re-added and thus in effect "refreshed"). The script absolutely does not set or change any metadata as part of this process - the only JSON requests being issued are Remove and Scan.

What Kodi does when it scans/rescans its library is entirely determined by whatever scraper is selected and how it is configured, and has absolutely nothing to do with the script. The script has no idea what scraper is being used by Kodi.

If Kodi is now pulling down missing plot information then it must be getting this from the internet via the currently selected scraper, which means you're not using the local NFO scraper. I don't know if the Stupid scraper is able to access internet connect (I don't know anything about this scraper) but from the "Stupid" name presumably not, so somehow TMDB would appear to still be selected (this is also not a script issue!)

Not sure if this is what you are doing but if you are initially loading/creating the library with the local NFO scraper, then changing scraper to Stupid, and THEN refreshing movies (now with Stupid scraper), then that might explain the different poster behaviour. If this is not what you are doing then I've really no idea...
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Reply
I could't figure it out after reading your message, and today I realied... the reason was that I had removed my entire library and only left 3 movies for the demonstration video, and then, when I readded all my movies and at that point I used TMDB just to try it out, so the .nfo tags for almost all movies got rewritten with a <fanart> and <poster> and <title> and so on linking to the TMDB library! It didn't even occur to me that this was a feature...

So now I'm back at the old poster-replacing behaviour, even on a new fresh install. But what I found interesting is that, as you can see, if the .nfo has the poster linked to the TMDB site, it gets loaded from there and doesn't show a thumbnail, so I wonder if I could somehow link the local poster from the .nfo.
Reply
(2016-12-15, 18:41)firewater Wrote: so I wonder if I could somehow link the local poster from the .nfo.

Generally speaking, scrapers should automatically use local artwork if it is appropriately named (and, presumably, if there is no relevant artwork url already present in the NFO). Removing remote artwork urls from the NFOs may restore your local artwork functionality.

Scraper issues are probably a little OT for this thread - I'd suggest starting a thread specific to your scraper of choice if you need specific scraper help. I vaguely understand the scraping process, but not really the specifics of NFOs or the order in which NFO fields are processed or whether NFO artwork takes precedent over local files etc.
Texture Cache Maintenance Utility: Preload your texture cache for optimal UI performance. Remotely manage media libraries. Purge unused artwork to free up space. Find missing media. Configurable QA check to highlight metadata issues. Aid in diagnosis of library and cache related problems.
Reply
With all the hassle your going through, why not use a offline scraping program like mediacompanion, scrape all movie data to be with your mother, pick your artwork etc, and then set Kodi scrapes to local info only to import your movie data.

Just my 2 cents ...
Media Companion Dev.
Media Companion - Kodi / XBMC - Media Companion
Reply
  • 1
  • 140
  • 141
  • 142(current)
  • 143
  • 144
  • 197

Logout Mark Read Team Forum Stats Members Help
[RELEASE] Texture Cache Maintenance utility17