Hopefully this will be fixed soon because TheTVDB and TheMovieDB is making Dragon Ball and Dragon Ball Z unwatchable.
2018-03-05 23:05:18,339 (124a0) : ERROR (networking:219) - Error opening URL āhttp://127.0.0.1:32400/:/plugins/net.fribbtastic.coding.plex.myanimelistā
2018-03-05 23:05:18,339 (124a0) : WARNING (logkit:19) - We werenāt able to get information from the server about the agents.
check firewall.
Alright, time for an update (no not April fools 
)
Iām currently working on an update to the Myanimelist Agent. I moved to another API that provides the Data for the Agent, the new version will use a standalone installation of the Atarashii API which is way better than my mess of an API that barely holds itself together.
Since this API is purely Myanimelist.net related it wonāt supply any additional information like background images, banners or more covers. The previous version of the Agent did provide this information because the āoldā API provided it because it made the requests to TheTVDB.com.
But while I have completed the work on the Agent, for now, I still want to add the TheTVDB information as well. For that, I am currently iterating over every. single. id. on. Myanimelist. to get the ID and the available titles. I automated this process a bit and am currently at 765/38000-ish. If that is complete I can write another program which will iterate over all ids and map the titles from Myanimelist to TheTVD or TheMovieDB. This mapping will then be used by the agent to pull additional data from TheTVDB or TheMovieDB to add to the Metadata.
However, that is a very tedious and long process initially because to not overload the requests on Myanimelist I can only make a request every 15 seconds. I hope the initial work is only that high and I can run the script to update the title and mapping list every week or month so keep it up to date.
I created a GitHub repository here that will store those lists for easier distribution and more importantly, everyone can clone the repository and make changes to it if something is mapped wrong.
The old API will still be online for now, so the old agent version will keep working (let me know if it doesnāt). But as soon as I have this all sorted out I will release the new version and disable the old API.
@Daraku said:
Alright, time for an update (no not April fools)
Iām currently working on an update to the Myanimelist Agent. I moved to another API that provides the Data for the Agent, the new version will use a standalone installation of the Atarashii API which is way better than my mess of an API that barely holds itself together.
Since this API is purely Myanimelist.net related it wonāt supply any additional information like background images, banners or more covers. The previous version of the Agent did provide this information because the āoldā API provided it because it made the requests to TheTVDB.com.
But while I have completed the work on the Agent, for now, I still want to add the TheTVDB information as well. For that, I am currently iterating over every. single. id. on. Myanimelist. to get the ID and the available titles. I automated this process a bit and am currently at 765/38000-ish. If that is complete I can write another program which will iterate over all ids and map the titles from Myanimelist to TheTVD or TheMovieDB. This mapping will then be used by the agent to pull additional data from TheTVDB or TheMovieDB to add to the Metadata.
However, that is a very tedious and long process initially because to not overload the requests on Myanimelist I can only make a request every 15 seconds. I hope the initial work is only that high and I can run the script to update the title and mapping list every week or month so keep it up to date.
I created a GitHub repository here that will store those lists for easier distribution and more importantly, everyone can clone the repository and make changes to it if something is mapped wrong.
The old API will still be online for now, so the old agent version will keep working (let me know if it doesnāt). But as soon as I have this all sorted out I will release the new version and disable the old API.
Thanks for the continued dedication. Daraku you rock!
@Hiltjabrand said:
Thanks for the continued dedication. Daraku you rock!
Thank you very much but this is something I also create for myself so the reason to do this isnāt really that selfless. Though I appreciate the thanks and I hope it can help others too.
Hi there! I just found this and I love the idea! Iāve been trying to get my anime to display and play correctly for ages.
I have created a new library using this agent as well as using BABS as my scanner. Some of my series are displaying perfectly, but some are completely wrong. Attack on Titan is showing up as āAttack No.1ā from 1969 and KaIji: Ultimate Survivor is showing up as āAmakageā. Kaiji: Against All Rules is appearing, but not gathering any info (no cover, no episode names). I was using the Japanese titles for both the Kaiji series, but that didnāt seem to work either; unless I messed up the titles. Akagi is showing up as āUG? Ultimate Girlsā.
These are just examples - out of my 30-40 anime series, itās picking up about half of them PERFECTLY and then the other half is really messed up; either the wrong metadata/artwork or, for Kaiji: Against All Rules, picking up nothing at all.
Is there a way to name series folders with a MAL number or something that would link the series to the correct MAL page? Like, for example, 2167 for Clannad? If I could just put that somewhere, my problem would be sorted!
I really want to get this working, my anime library being all messed up has been bugging me so much for the past few weeks!
Any help is much appreciated! Thanks all <3
A few things on this topic:
Plex scans the Video files and parses the title of the episode and series from the name of the file which is done by the Scanner. The Scanner (BABS in your case) also only adds the found and parsed video files to your library, the Agent will take the Title of the library entry and looks up the specific information for that title.
That means you need to have a good naming convention of your files that is producing matches on the location you get your information from.
So if you have files that are producing wrong metadata can have multiple issues. Either your files are named wrong on your file system, the title is parsed wrong because of the scanner or the name is not producing any matches through the Agent.
Another note on this: The current Behaviour is kind of crappy. The reason for this is that I needed to create my own backend, my own API, to get the necessary information. The official API is dead (but as I being told gets completely reworked and internally tested at the moment). My API only parses the official website, the agent requests the information with the name of your library entry from the API and the API will search on the official website for this title. Iām completely at the mercy of what the MAL Website search produces as results, my API just provides a better and more compact version of those results.
Which means that you should take the name you have in your library and do a search on the official MAL Website. If you find it then it is a bug or another issue of the API, if not then it is an issue with your naming convention.
You could also do a āfix matchā and use a better title and try to find it that way. However, this is a lot of work depending on how large your library is and how many mismatches you got.
In general, how it works right now, the best result is to use the main title used on the MAL website (the title above the poster when you browse a site) if that is not found you can also use synonyms or the English title. I generally would avoid using the Japanese title though because there could be many issues with the encoding of the Japanese characters.
But even that is not protecting you currently for errors or wrong searches on the official MAL Website
a few Examples:
MAL only allows at least 3 characters to search for so the Anime āKā will not find anything so you have to use a synonym āK-Projectā (which will then be āKā again after the matching was completed), or Kon! (the First season) can be matched as Kon!! (second season) because MAL actually delivered Kon!! first before Kon!.
Yes, I know it is strange but since I am at the mercy of how the Web search behaves there is nothing I can do.
However, with the new version of the Agent, this will change because I use a new API from the guys at Atarashii on my own server. though Iām already pretty far (only need to implement theMovieDB and theTVDB connection) it still will take some time to get it all together.
Not only that but I also need to fix the issues in the automation process of creating those lists to make the mapping between a MAL entry and a TheTVDB or TheMovieDB entry and I noticed only just yesterday that I had a problem in the data retrieval from those additional websites which means that I need to take another round over every found title and match them again with those websites to get a higher number of mappings.
And lastly to your āMAL numberā suggestion. While this could be possible it would also not really work since there are Animes out there that only consists of numbers as the title like 001 or 123 (Iām not kidding).
So I would need to find a way to distinguish between it being the MAL ID or the title. Pretty impossible if you ask me. But then again, if you are willing to edit the Anime to enter the ID as the title then why not just use the āFix matchā functionality Plex already provides?!
TL;DR: Mismatching might not be always the fault of the Agent but could be the poor naming conventions on your filesystem or the search results of the MAL Website. New API is already Up and running. The agent can already pull information from the API. The Mapping to TheMovieDB and TheTVDB is worked on but takes time to complete (4 hours for 1000 entries) and actually needs to be redone because I just found an error that leads to not found entries on TheMovieDB or TheTVDB. Using the MAL ID to find specific entries, probably not but you have the āFix matchā natively in Plex already.
Just a quick update on the state of the Agent.
I updated the Mapping generator and let it rerun (still only at 10k entries and have 28k left). A lot more mappings will now be created or āfoundā since I request data for every available title and use the most occurring ID of those results.
I spend the weekend with getting back into the Agent and I can say that I just completed the TV-Show section. The Agent will now pull the Mapping file from Github lookup the ID of the matched Anime in your Library and get the TheTVDB ID (or also the TheMovieDB ID for Anime Movies) if it is available. If there is an ID the Agent will request the background, the banners and additional Posters from TheTVDB and add them to the metadata.
What is left is the TheMovieDB data request and of course, finishing up the mapping generation. And if anyone can help me with JQ to merge and update JSON files that would be awesome.
Maybe you can do without JQ?
I did implement dome Dict functions in https://github.com/ZeroQI/Hama.bundle/blob/Beta-Agent/Contents/Code/common.py:
- def Dict(var, *arg, **kwarg): Return the value of an (imbricated) dictionnary, return āā if doesnāt exist unless ādefault=new_valueā specified as end argument
- def SaveDict(value, var, *arg): Save value to a Dictionary field (can be nested) and if value is a list, it extend the list instead
- def UpdateDict(var, var2): var.update(var2); return var
- def DisplayDict(items={}, fields=[]): to display a dict and selected fields all aligned
i used that to avoid key errors and assign only if the value is not empty and save to nested dict in one line, also allow to add to a list if the value is a list, which reduced dramatically the test structures needed in my codeā¦
JQ wasnāt really meant for the Agent but rather to merge all the JSON files generated for the mapping.
I found this pretty useful to use, I can already append one JSON file to the other but I have to find out the values that you need to specify to update existing entries or append them.
My Idea was to give JQ per command line the full mapping and the file it just creates and updates or appends the entries to the JSON Array. That would make it so much easier to keep the list useful without the need to merge every single JSON file (not to mention that I could just let the generator run for a single ID and automatically update the entry or just add it if it doesnāt exist already.
Muchas gracias.
Ā”Animo con el proyecto!
This is awesome but I have a quick question here:
For example with One Piece I want to make a different season for each ARC which I then can rename with this trick https://forums.plex.tv/discussion/285038/custom-names-for-seasons posted by the user named ānavelsā.
But heres the thing when I name the episode One Piece S39E575 then the agent doesnt get the metadata but Plex creats a
new season.
When I name it One Piece E575 on the other hand it does find the metadata but Plex does not create a new season but instead groups everything into Season 1.
Is there a way for me to tell the agent to ignore the S39 and find the metadata I am looking for?
FYI I want it to show up in Plex according to this website http://onepiece.wikia.com/wiki/Episode_Guide and it currently works but as I said I dont get metadata with that naming.
For starters, the MyAnimeList Metadata Agent requires absolute episode naming! (That is also why I recommend using BABS)
The Agent should, on the other hand, pull general Metadata information regardless of your choice though. If you could provide your log file so that I can take a look at potential errors and fix them in the new version of the Agent, that would be great.
However, even if general Metadata Agent information like Name, Synopsis etc. should be pulled and added to your library you will absolutely lose episode information. The reason for this is that MyAnimeList does not support what you are looking for and neither my API nor the new Atarashii API (that I use for the new Agent) does this.
If you want to use seasonal Animes I think TheTVDB is a better choice for you. If not then you should either switch to Absolute episodes or take the loss of episode information.
Yes I do get the general Metadata but not the episode metadata but as you explained I am out of luck but thanks anyway!
Just a quick update: I think Iām pretty much done with the Agent.
The only problem that is left is to get the API back up and running since and fix the Internal Server Error that I get whenever I try to request information for an ID, search works though.
Maybe it is just a problem resulting from the MyAnimeList being down a couple of weeks ago and it fixes itself I donāt know yet.
But so far Iām happy with how the Agent turned out and after the API works again to test it against my own library. If that is going well and there is not much to āfixā then I will be able to release the new version.
Have a Nice Day
Iām having issues with getting the metadata for several of the new series. They are matching, just not getting metadata.
- Asobi Asobase
- Hanebado!
- Hataraku Saibou
- Hyakuren no Haou to Seiyaku no Valkyria
- Happy Sugar Life
- Lord of Vermilion Guren No Ou
- Harukana Receive
- Grand Blue
- Yama No Susume Omoide Present (This one is actually old)
Others match and get metadata fine.
- Island
- Ongaku Shoujo
I have a feeling it is same as before when this happened to me. Something happened on the MAL side that you might need to bump your API.
Oh, I am using your latest bundle (MALAgent v5.0.2)
I am having the same issue, Hataraku Saibou and Hyakuren no Haou to Seiyaku no Valkyria get no metadata (and a few others) but strangely enough, Grand Blue did manage to get the metadata just fine.
Actually the agent does retrive metadata sometimes but fails at updating it, on these cases, matching and unmatching seems to have worked for older anime. Here is a log for requesting metadata Chio-chan no Tsuugakuro net.fribbtastic.coding.plex.myanimelist.log (11.6 KB)
That is strange, I checked the Database and it was OK, I repaired it just in case.
Though I wonāt spend too much time trying to fix this because it has served its purpose and with the new Version on the horizon, there is not really much reason to put additional work into it.
I still have some things to do with the new version, the API works and I have tested it with my renamer and mapping programs over the last few months and could get the developer even fix the issues while the MAL website was down and deactivated its features. So I need to grab the last like 10K of the mapping IDs to be up-to-date with the mapping and let the new Agent Version run against my own Library and fix any bugs that turn up.
So far my tests were successful and quite promising I hope you can wait some more time. I should get some time to work on this in a few weeks.
Ok, Iāll sit back and wait a little bit. I appreciate the work your doing on this.
