Skip to main content
Topic: foo_discogs (Read 856502 times) previous topic - next topic
0 Members and 2 Guests are viewing this topic.

Re: foo_discogs

Reply #2775
If you have not edited many Discogs entries in the past, your changes will probably be held until somebody approves them, so you may not be able to do it immediately.

Discogs edits now happen immediately and anyone with a Discogs account can edit a release. If some other user disagrees with the edit, they can choose to comment, vote or revert the change. Any edits have to conform to Discogs RSG
Quis custodiet ipsos custodes?  ;~)

Re: foo_discogs

Reply #2776
If you have not edited many Discogs entries in the past, your changes will probably be held until somebody approves them, so you may not be able to do it immediately.

Discogs edits now happen immediately and anyone with a Discogs account can edit a release. If some other user disagrees with the edit, they can choose to comment, vote or revert the change. Any edits have to conform to Discogs RSG

I've updated the page on discogs and it is working fine now.

I'll know what to do in future.

Thanks for all the help.

Re: foo_discogs

Reply #2777
Discogs edits are immediate but the Discogs API can cache older info so sometimes you have to wait a few hours for the new info to come. foo_discogs also caches info so you'd need to clear the cached releases in the component settings (or restart foobar2000).

Re: foo_discogs

Reply #2778
Hi zoomorph,

First, thank you for continuing the development of that Foobar plugin, it probably is the best option to get the most metadata from Discogs nowadays.

I do vinyl restoration regularly and I was thinking about writing matrix / runout infos in audio files, I'm guessing it's not implemented in the plugin at the moment? I tested the "barcode" field for CDs and this works great though.

Seems like the field for vinyl is called "Matrix / Runout" in the Discogs API.
Correct, it's currently not available in foo_discogs. There can be multiple Matrix / Runout variants for a single release. Not sure how this should be handled.

Late reply on my end.

If there are matrix variants for a single release, I seem to recall each should be separated in its own release page. I don't think I've seen one release page with multiple matrix variants, but I could be wrong.

Re: foo_discogs

Reply #2779
Late reply on my end.

If there are matrix variants for a single release, I seem to recall each should be separated in its own release page. I don't think I've seen one release page with multiple matrix variants, but I could be wrong.
See: https://support.discogs.com/hc/en-us/articles/360005054893-Database-Guidelines-5-Barcodes-Identifiers

Quote
One edition of a release may have many different matrix numbers on individual copies - especially for major label releases. These are considered manufacturing variations for the purposes of cataloguing on Discogs, and not unique releases, so one Discogs 'release' may contain multiple variations in matrix numbers etc. For smaller releases, or in special cases, the matrix variation may indicate a unique release / edition, so please ask in the Database Forum Topic if in doubt.

Re: foo_discogs

Reply #2780
Oh, very late reply, sorry!

@kode54

I'm on Win7 64Bit. (Both the VMs and my working computer).
Don't trust my User Agents as I use a lot of different browsers with a lot of different addons which -some of the time- will fake the User Agent/the OS.

If there ever is something special you or anybody else, that needs to be tested, you always can ask and I'll try to brute test it...

@fuffi It may also depend on which OS you're running foobar2000 on. I can't guess what you're running from your user agent, because that's kind of prying, and the two user agents I have seen from you don't agree completely.

Windows virtual machine? Which Windows? It may be a case of legacy Windows versions not being tested properly, if you're using the long out of support XP. It almost makes me wonder why Peter and I continue to build with the XP supporting toolchains, when at least I don't actually bother to do any testing to see if anything actually blows up there, and I don't know what Peter does for old OS testing, other than maybe doing some targeted testing when someone points out a bug.

@zoomorph

I know, that I'm out of the ordinary, when using your fine component. Most of the time (90% of 100%) if works very well and it is so damn useful, I couldnt live w/o it!
The other 10%, well, I'm trying to reproduce it as best and precice as I can, so others might reproduce it as well and we might track the cause down and maybe fix/improve the component. If not, well, 90% are being top when using a software nowadays! :-)

About the "not always returning country":
I pers. think, its not a problem with the API, but with your caching routine.
You can clearly see the country is shown at first, and after some clicking (no reloads from the API, just fooling around with some buttons in the component) the country vanishes.
Seems to me, that the concatenation of the used strings to be shown in the list of releases has a bug, when using cached information.

I also miss the useful info like CD, 2xCD, 3xCD,... in the list of releases. They appeared some months/years ago and suddenly vanished completely. (I'm not sure, if its a new boundary of the API, but it was very useful to filter all non-LP releases)


Thank you for testing. Strange....I don't know if something changed on the discogs side, because their API give and gave the correct result.
I just wasn't shown in the releases tab of the fine component.

I guess its the same strange phenomen in the component like here: https://hydrogenaud.io/index.php?topic=50523.msg967983#msg967983 (you clearly see it in the animated gif)

The component swallows some information retrieved from the cache under some unclear circumstances :-)
If it worked the first time, then it should continue to display the same result while the info is in the foo_discogs cache (unless there's a bug in foo_discogs). Not working 5 hours later suggests to me that some or all of the info was probably kicked out of the foo_discogs cache by then and so it was reloaded from the Discogs API. The Discogs API could have returned something different at that point, or there could be a bug in foo_discogs (ie. maybe something goes wrong if part of the data is cached while another part is reloaded).

It's always hard to investigate a bug that can't be reliably reproduced. This seems to happen rarely so perhaps it's not worth putting much effort into. I played around a little with a small cache size in foo_discogs, but wasn't able to trigger it.

As for that other one about the country being displayed in the find release dialog, I suspect that one is probably related to the Discogs API search endpoint sometimes returning country and sometimes not (at least that's how it used to be). The release endpoint, which is loaded after you press "Next>>", always returns country. So when the results are repopulated when you go back and change the filter, the new information is reflected there.



 

Re: foo_discogs

Reply #2782
Not sure if dev wants individual releases that cause errors, but here's one:

https://www.discogs.com/Sonny-Rollins-The-Freelance-Years-The-Complete-Riverside-Contemporary-Recordings/release/5315647
That would be an:
Code: [Select]
(FATAL) Error: Error loading release 5315647: JSON Parser ExceptionError parsing release credits.

[ESCAPE to close]
error.

I guess, this is (on of the) the problematic part (taken from the credits part of the link you posted):
Code: [Select]
Thelonious Monk (tracks: 1-1 to 1-)
(The  "to 1-" which indicates, a  missing tracknumber)

Unfortunately, the fine component cannot do anything about it (otherwise it would lead to false tagging of your files), so you have to correct the release on the discogs site/db yourself or kindly ask in the discogs forum anyone to do it.

Next time, please show us the error message too (not only the link to the release), so it is easier for us to guess the problem :-D

Re: foo_discogs

Reply #2783
(f2k 1.5, but earlier version too and component Version: 2.20, Compiled: Aug  7 2019, but earlier versions too)

Today I'd like to show a screenshot of a greyed-out NEXT button. :-)

I had this from time to time (and when this is comming, its time to restart f2k, otherwise the finde component will cause a crash the next couple of invocations), but closing the Find Release window (via Cancel) and start the component again will help to reappear that NEXT button. So its not critical (the crash is, but its not reproducable. I guess, the more I have used the compoent = big cache and other stored things in RAM...)

The NEXT button is never greyed-out from the beginning (aka start of the fine component), I have to go to the Match Tracks windows and go back a few times and and check other releases too.

So its not really reproducable ...

Just wanted it to mention, maybe others can better reproduce or I'll find a way the next years :-)

Beside the (from time to time) crashes, I'm very thankful for such a usefull, solid and reliable helper here in foobar2000!!


Re: foo_discogs

Reply #2785
why track 2 & 4 would cause foo_discogs to spew out:
"(FATAL) Error: Error generating tag DISCOGS_ARTISTS_ALIASES [] for file file://D:\* *\02-Kraut feat. Bachan-Lentamente.flac
Interesting...

My error is not in ALBUM ALIASES, but in ALBUM ARTISTS:

Code: [Select]
(FATAL) Error: Error generating tag ALBUM ARTIST [] for file file://E:\test\test.mp3

[ESCAPE to close]

Here is my code:
Code: [Select]
$multi_if($multi_and($any($contains($first(%<<RELEASE_FORMATS_DESCRIPTIONS>>%),$array(Compilation,Mixed))),$multi_greater($length($unique($flatten(%<<RELEASE_TRACKS_ARTISTS_NAME>>%))),$div(%RELEASE_TOTAL_TRACKS%,2)),$multi_greater($length($unique($flatten(%<<RELEASE_TRACKS_ARTISTS_REAL_NAME>>%))),$div(%RELEASE_TOTAL_TRACKS%,2))),Various Artists,$joinnames($multi_replace(%<RELEASE_ARTISTS_NAME>%,'Various','Various Artists'),%<RELEASE_ARTISTS_JOIN>%))

Lets compare it with your code, maybe we find the culprit...
but i guess, its the empty VARIOUS .... https://www.discogs.com/artist/194-Various

Re: foo_discogs

Reply #2786
Hai.
Does anyone see anything on here:
https://www.discogs.com/Various-Confluencias-I-Selva/release/11123050
why track 2 & 4 would cause foo_discogs to spew out:
"(FATAL) Error: Error generating tag DISCOGS_ARTISTS_ALIASES [] for file file://D:\* *\02-Kraut feat. Bachan-Lentamente.flac"
If you click through on Bachan (2) and Yeahman!, both artists give 404 pages.
I've seen this happen now and again lately - I'm not sure how it would happen, because surely when they're listed on a release they would get an artist page of some sort, even if they have no releases other than appearing on this compilation?

Could be worth posting in the Discogs forums about them?

 
SimplePortal 1.0.0 RC1 © 2008-2019