This is the enhacement of Discogs Credits script to support multiple sources. It currently adds Qobuz and Tidal and I will add more if appropriate Source is found (we can API it or scrape it).
Try it on this release which contains both Tidal and Qobuz sources.
Note that not all sources are equal, some are more rich and capable then others (see docs).
I was inspired to do this as Discogs can be poor source on digital releases and MET import has shortcomings: it adds non-existing artists which you must resolve all at the import time. Credit Hoarder works the same way as Discogs Importer so it will present review table and you can safelly ignore unresolved things or come back later to it any time. Plus, creating missing artists in review table is way faster than in relationship editor.
This script is incredible, I have been wanting something like this since, well I guess since I joined MB earlier this year. No more manually adding each and every credit relationship, this does it for me! I’ll have to do more testing with it, but so far it worked perfectly on the one release I tried it on.
Edit: Are there plans to add more credit types? This release, for example, has stuff like Sound Engineer and Mixing Engineer.
Script scrapes HTML page on Qobuz and there it can’t find the link. Artist link probably exists with full commercial API access but its not there in HTML. If you open the credits on Qobuz page, you will see they are given linkless. The purpose of the link is to better identify artists on the MB side - by matching not only by name abut also by link which makes one much more confident that the right one is selected. The importer works without it as well.
If somebody knows a way to get links, that doesn’t include paying monthly for API access, let me know.
For now, Qobuz doesn’t use links, and always shows that message. TIdal/Discogs use it.
Yeah, anything missing should be reported and I will add it. I will see about that release
Mastering Engineer — artist→recording mastering is deprecated in MB; mastering belongs at the release level, not per-track, so there’s no clean target here.
Sound Editor, Studio Personnel — no clear MB equivalent.
One issue I’ve noticed, adding the “credited as” field even when the credit name is identical seems redundant, and leads to issues like this:
Two of those relationships were added manually (without the Credited As field), and two were added automatically (with Credited As).
Being a bit of a sadist I threw it at this release on Discogs as the first test:
After 81.2seconds of “preflight”, Impressive results. And well presented.
I like how it lets you decide how to handle things that don’t match. That initial stage where it shows you the matches worked really well. And then putting everything into place on the page so I can then give a final check, adjust, tweak. Excellent.
And I didn’t even bother reading the instructions, so that shows it is intuitive.
That release was a mess of missing works, unlinked artists, a dozen artists that didn’t exist in MB, clashing names and vagueness… so much easier to sort out on the one page like that.
Even the little niceness of closing the extra tab that you open when adding a new artist. Neat. Normally after adding a release like that I have 50+ tabs left open.
Big time saver. Thank you for the hours you will save of my future edits.
Thank you, @IvanDobsky, a lot of time was devoted to this one, I am glad it shows.
Yes, it seems you did. Work is either created when missing (always) or when needed (required by some credit and doesn’t exist currently among relationships) or never, depending on option in toolbar. Script never searches MB database for existing works or match them. So if you want to search MB and cherry pick, you will have to do it before import manually. Script will then use existing works to add credits to them if required.
Perhaps this can be part of future version but I don’t have much experience in matching works so this is intentional. I would like to have it though myself, that is the reason why Apollo already matches artists and recordings. Feel free to fill up issue on how it could work.
(I tried to drop in the log but the forum got upset at the size)
You can copy the log without JSON in review phase or during preflight, that will produce the lowest amount of text. Since log is big, I might add export to Github gists in the future.
I use this script to “guess works” - Guess related works in batch
But it is pretty hit and miss due to limitations of MB search finding one word songs. Example - attempts to find “Time” as performed by “Pink Floyd” comes up with “Time Time Time” due to weirdness of how the search works. Probably best to stay away from that step or “Money” matching to Abba’s “Money Money Money” always makes me laugh. I end up doing a search for “Money Waters” to get the right one (i.e. Song + Writer) which could help accuracy if you go that way…
That script is currently about 80% accurate.
Now I know to sort out Works first, that is a relatively easy step. As I said, I didn’t actually read any instructions before diving in.
And thank you for that time and effort. You will save people hours days.
I think I copied it after pre-flight? the plain log? 1981 lines…
Yeah, I played initially with it a bit and came to the same conclusion. It looks too error-prone and so far you can use Credit Hoarder on auto-pilot with 5% brain involvement as auto matching works superbly.
Your mid-step that lets us check the results is good. It would get overloaded if you then slotted Works in there as well. Staying away from Works and keeping focused on Credits seems a good decision.
If you are going to do something with Works, then make it a separate script (which includes take the “make works” step out of the current Credits script)
or “Money” matching to Abba’s “Money Money Money” always makes me laugh. I end up doing a search for “Money Waters” to get the right one (i.e. Song + Writer) which could help accuracy if you go that way…