[TP.One]

Also interested In Discoigraphy (but personaly i browse much more ofthen from the AlbumInfo Dialog to the ArtistInfo.., rather than vice versa via the discograpy to the AlbumInfo.)

How ever, for me a nice Info, which im apreciate.

Feels sometimes the Discograpy is sometimes incomplete, did recently had a Look at the discograpy of "Bob Dylan", mainly to see if there are still Releases.
Estuary did List 39 Albums. At Wikipedia there where 37 Studio Albums, 11 Live Albums and several Bootlegs and Compilation Listed.

May a bit a extreme Case, but may shows the limitations. (And guess also a different Topic, so OT)

At least in Estuary there where listed more albums for this specivic Artist, than in another Skin, where the discograpy seems to came from Extended Info (plugin://script.extendedinfo?info=discography&&artistname=$INFO[ListItem.Artist]) which lists "horrible" 25 items -> may a Script limitation.

--
so, writen much without saying anything

EDIT: More OT: Artist Bio of Bob Dylan was also incomplete (at least in German), but guess this is caused by the source the scraper used.

[User 421137]

Feels sometimes the Discograpy is sometimes incomplete, did recently had a Look at the discograpy of "Bob Dylan", mainly to see if there are still Releases.
Estuary did List 39 Albums. At Wikipedia there where 37 Studio Albums, 11 Live Albums and several Bootlegs and Compilation Listed.
...
EDIT: More OT: Artist Bio of Bob Dylan was also incomplete (at least in German), but guess this is caused by the source the scraper used.

Even if it is basically OT, just a few words:

The albums for the Discography is (also) based on the scraper and the used source. For example MusicBrainz has 37 "Release Groups" (i.e. Albums) in the category "Album". This are typically the studio albums. Other albums might be e.g. "Album + Live" or "Album + Soundtrack" or also "Unspecified type" and others.
AFAIK only the official Release Groups with category "Album" (and only with this single category) are used by the scrapers for the Discography when MusicBrainz is used (but I'm not sure in detail and I don't know if all scrapers work the same way).
The category in MusicBrainz is not always like I'd like to have it, is not always free of errors and sometimes it is not easily to say if there should be just one or two Release Groups.


Artist bio, especially in German, is a vexatious topic. At some data sources they are bad, outdated or just missing. Not every scraper supports all data sources (e.g. no Wikipedia). Scrapers using Wikipedia might possibly bring the best bio, but IMHO it is often also too much/too long. I came up to avoid bio in Kodi. It just blow up my DB, was often bad/outdated/not in my preferred language and keep it up to date is also not so easy. In most cases I don't really need the bio in Kodi or it's better for me to have nothing than something not good enough for me.

But again, it is OT ...

[DaveBlake]

@"NeroRome" @TP.One I have give this a new thread (split from https://forum.kodi.tv/showthread.php?tid...pid2823486) since the previous request for better matching of discography has been implemented for v19. OK so there is a bug in it for some MySQL/MariaDB versions that causes discography not to show, but a fix will be in v19.2

Here you both seem to be asking for artist info scraper changes/improvements over what discography contents or other artist info is picked up. That is something outside Kodi core, but as an addon change is not tied to the slower Kodi development cycle. Anyone (with the skills in Python) can write a new scraper, and they are welcome to do so.

[User 421137]

I have give this a new thread

Okay. Fine!
 

... since the previous request for better matching of discography has been implemented for v19.

Well, I saw there was a change/improvement but if I see it correct the improvement was to use MB Release Group ID. This is basically fine and a good improvement but my originally intent was not about this.
My initial wish was about e.g. non-case sensitive matching and matching of typewriter and typographic punctuation – especially without MB-IDs. If I see it correct there was no direct improvement therefore.
 

Here you both seem to be asking for artist info scraper changes/improvements over what discography contents or other artist info is picked up.

Well, I personally didn't ask for anything new (and I also don't see this from @TP.One's side by now).
Of course, scrapers can always be better/more flexible/etc. but I don't think this topic will end up in some changes/improvements for them. I guess it is too unspecific ... but maybe it will be the base for a more detailed topic.

[DaveBlake]

If there is still a wish for more changes to dicography it is better that it gets restated here on this thread (discussing v19 functionality and future) than lurk on the old thread when at least some of that discussed there has been implemented.

My initial wish was about e.g. non-case sensitive matching and matching of typewriter and typographic punctuation – especially without MB-IDs. If I see it correct there was no direct improvement therefore.

No I didn't implement any kind of fuzzy matching, just use of Musicbrainz release group ids. I see requests for features as a negotiation, I didn't mean to imply thought that I had completely statisfied your wishes! I moved from having no personal interest and no intention of doing anything to implementing an improvement that helped with duplicates and mismatches. Matching is not case sensitive, but deeper levels of matching is not something I think is worth my limited time.
 

Well, I personally didn't ask for anything new (and I also don't see this from @TP.One's side by now).
Of course, scrapers can always be better/more flexible/etc. but I don't think this topic will end up in some changes/improvements for them.

@"NeroRome" I read your post as saying that the discography scraped from Musicbrainz was not always of those albums that you would want included. Want better filtering of what category of releases included, or scraping from sources other than Musicbrainz then those are the responsibility of the addon.

@TP.One said that the disography for Bob Dylan seemed incomplete and limited, again that is a matter of what release data the scraper fetches.

Artist scraper addons could look elsewhere other than Musicbrainz, filter results differently, it could even do fancy matching to the albums in the library and adjust what it returns accordingly (adding mbid via a look-up so that Kodi can match-up what it is given too). There is huge potential for changes to discography via the addons without any need for changes to Kodi core. Given the greater number of people that can write Python compared to those that understand the core Kodi design, run an IDE and able to makes changes C++, I offer this as an approach in case anyone is interested in taking it up now or in the future.

[User 421137]

I don't know to whom you address your reply at all and therefore I give some aspects to clarify (incl. them addressed to me):

• There was no new whish/suggestion for Kodi itself from my side (and as I already said: I didn’t see it for @TP.One’s side)
  
• The posting from @TP.One and also my replay have been OT, I clearly marked it so (and it should be clear not to go on with OT in an excessive way) but I think it was worth to answer TP.One.
  
• I will not repeat whishes/suggestions again and again. It’s fine if they are implemented but if not, I (we) have to live without. That’s how it goes. Nobody drives you (or anybody else) to do it.
  
• Everybody has limited time. Neither you nor anybody else have to explain why not implementing something. Everybody is free to do it or not.
  
• For me there is a big difference if something is implemented completely or not, and if the main and initial suggestion (the non-case sensitive matching, without MB-IDs) is not implemented or not. If the main and initial suggestion was not implemented it’s improper to say it has be done.
  
• https://github.com/xbmc/xbmc/pull/18079 is a good improvement but I see less connection to (the initial suggestion) of https://forum.kodi.tv/showthread.php?tid=340947
  
• If “matching is not case sensitive” now it is be fine (and it is the main suggestion of https://forum.kodi.tv/showthread.php?tid=340947)! I didn’t test it now but I see no change in the code for it in https://github.com/xbmc/xbmc/pull/18079/files (but my knowledge in C++ is very modest!). For me https://forum.kodi.tv/showthread.php?tid...pid2823486 shows it clearly it was case sensitive in the past. If it is working case insensitive now, it’s fine and sorry for not seeing it (also not in a different issue)!
  
• No, I did not complain the scraping of the discography from MB but the data in MB itself! It is clear that the scraped result cannot be correct if data in MB is “wrong”.
  
• I guess it’s clear to everyone why Dylan’s discography is incomplete. It was in my reply to the posting of TP.One (and the major reason for me to reply him).
  

[DaveBlake]

@"NeroRome" you sound grumpy with me for some reason, I have no idea why,  I certainly have no beef with you. My reply was an open response and explanation to anyone that was interested.

I can confirm that matching is not case sensitive. There is no C++ code to see this is as the match is done using SQL, and the music database is defined to have case insensive "LIKE".