MTGJSON as source & automated building

[tooomm]

Dropping a thought here:
MTGJSON has token cards, and with the last release (5.2.1) there are also information on related cards.

The data should be from Scryfall as well. We could build the database automatically rather than trying to keep up manually, similar to our spoiler files.

For the future, once Oracle is not used anymore and there are ready-to-use, MTGJSON sourced card databases for Cockatrice, tokens can maybe be part of it as well.

[SlightlyCircuitous]

Sounds like it could be very useful. However, how would this work with token images that aren't on Scryfall? would they need special logic to pull the images from other sources?

[tooomm]

Don't mostly all new tokens have art?

There could be an additional file to list other cases as secondary input source.
When I think about the earlier spoiler approach in this repo, there were also several files that the script consumed for corrections and adding missing cards for example.

[SlightlyCircuitous]

So, we would have a manually curated file of tokens that don't have scryfall/gatherer art alongside the automatically generated file from MTGJSON and then stitch those two together for the tokens file for Cockatrice?

[SlightlyCircuitous]

also, yes, most new tokens have art, although for whatever reason Jumpstart '22 tokens didn't have art at all

[tooomm]

So, we would have a manually curated file of tokens that don't have scryfall/gatherer art alongside the automatically generated file from MTGJSON and then stitch those two together for the tokens file for Cockatrice?

Not two files, but two sources (Scryfall database + new special cases file) that are both consumed by a script to build one file for Cockatrice to use (tokens.xml).

At the end the repo would be more similar to our spoiler repo wich runs a script on a schedule and automates the update process (partially).