I’m running into a problem with my Audiobook Agent that I maintain. Somewhere in the recent server updates, changes have been made to how the agentkit passes off the manually entered search term for ‘fix match’ for a music album.
Previously, I was able to reference what was entered into the the Title field using the media.name variable. That seems to be deprecated now. According to the logs, it looks like it’s attempting to pass it through media.title, but when I try to use it it’s not being respected the way album.name was and I’m not getting anything.
I know that they announced a lot of changes coming to the music libraries back in like June/July. Anybody have an information on this? These changes have crippled half of my agent.
What I can find in the logs is that it looks like it’s passing it differently than it used to. I entered “TEST BOOK TITLE” in the manual matching box to see where I could make it pop up.
Log from 1.18.1.1966 2019-10-16 15:14:21,525 (7ff5e0ff9700) : INFO (agentkit:957) - Searching for matches for {'title': 'TEST BOOK TITLE', 'parentGUID': 'local://130094', 'year': '2019', 'id': '130095', 'parentID': '130094'}
Log from 1.13.9.5439 (it’s an old test/dev server I just have updated) 2019-10-16 15:17:51,332 (89f0) : INFO (agentkit:957) - Searching for matches for {'parentID': '36107', 'year': '2019', 'id': '36108', 'name': 'TEST BOOK TITLE', 'parentGUID': 'com.plexapp.agents.audiobooks://Craig-A-Falconer%20B00CTNURW2?lang=en'}
My first clue was that calling media.name suddenly started throwing errors. But I was only using it for manual matching, so adding media that auto-matched properly didn’t give me an indication that there was a problem. It wasn’t until I had some users report to me through github and here that I dug enough to see what was going wrong.
It looks to me like the variables being passed are slightly different. Don’t suppose there’s a current guide out there anywhere?
Thanks for taking the time to look into it. It’s much appreciated. I have been legitimately surprised by the number of people using my agent. Makes me eager to keep it working for everyone.
Thank you macr0dev for all the work on this, it really is hugely appreciated. Also thanks to elan for getting personally involved and tracking down the issue. For all the recent “Plex don’t care blah blah blah” this clearly isn’t the case.
I’m on the latest version of PMS and still have the same issue.
What irks me is that after adding a new book and I scanned the library all the cover art of my existing books disappeared.
I hope this get solved soon.
This is doubly annoying because I just found the Prologue app and was just about to migrate my audiobook library to Plex. So if Plex can just scrap audiobook files properly…
I understand everyone’s frustration, but these things happen. As soon as they’ve corrected whatever was inadvertently broken everything should start working again. But we have to wait on their release process for their changes to make it through testing.
This is probably the only 3rd party agent that exists for audio libraries, so I’m not overly surprised that the bug wasn’t caught ahead of time. We just have to be patient.
I hate to be the one to break it to you, but since you haven’t picked this up on your own I must.
The audiobook scraping is a NOT an innate part of Plex. It’s a metadata agent that I wrote and distribute for others to use. I do not work for Plex, I’m just a dude who wanted my audiobooks cataloged in Plex and then shared that code with others.
I’m a 3rd party. And this is my agent. Which makes this a 3rd party agent.
Oh I know. I’m just saying scrapping audiobooks using Plex defaults seem to be broken too.
FWIW I switched between Plex Music, Plex Music Scanner, Last.fm. Deleted my library and recreated it. Same issue persists. No album art for my audiobooks.
Actually, Plex Music now uses Musicbrainz as a source. Musicbrainz supports audiobooks. So Plex supports audiobook metadata as long as the source database (which is community-edited) has that data. @Bjoure, if Musicbrainz isn’t finding an audiobook, I suspect it’s because no one has added it there yet—not because the agent talking to it is broken.
Music in Plex starts with the best-in-class database of music content from MusicBrainz and then also adds in the best-in-class musical metadata from AllMusic.
First and foremost. That is an awesome book. I have book 8 but am working on Sector 64: Ambush and probably won’t get to it until next week.
I was aware that there were some entries in musicbrainz for some books. Unfortunately, I don’t consider it prolific enough to be viable. Especially given some of the examples I’ve been given to test with from users when they have problems. There’s also the big differences between people with properly chapter-ized books, vs single mp3 files, and the dreaded “cut along time stamps”. How does it handle those?
As a legit question - did your track titles come from the tags or from the entries you created in musicbrainz? How would it react if they weren’t the same?
Congratz on being the kind of person who contributes to user generated databases. I wish there were more of us out there! I’ve even considered making a db for audio books with an API for people to pull against. I just always considered getting all the content together to be more of a challenge than the actual coding piece.
Thanks for your patience and for using Plex!
