doomeer
commented
1 year ago Thanks for the heads-up, I was not aware that it was no longer maintained!
I had a very quick look at dat-schema. If I understand correctly, it describes the format of GGPK files (more precisely, the data files inside the GGPK archive). It could be used to generate a parser for dat files in GGPK files, in any language. That sounds very useful. I didn't look at whether dat-schema had the information for all the files that Kalandralang needs though.
(Fun fact: a few years ago I had written OCaml code to extract and parse data files directly from the GGPK file. The code was not very good quality unfortunately so it's not really usable again. It was a lot of efforts though.)
In order of difficulty for me, from the easiest to the hardest:
- someone starts maintaining something that looks like RePoE (same file names, same format), but it looks like this is also the least sustainable option if community efforts are now focusing on something else;
- someone generates JSON files (or YAML or XML or whatever as long as there is a parser for it in OCaml) that contain the same data, with a similar format to RePoE, just different field names and perhaps some minor structural changes; and someone hosts those JSON files somewhere; all I have to do is fix field names, so it's easy;
- same as the previous item, but with more structural changes (data in different places, need a transformation to get back what Kalandralang understands): depending on the changes and how intuitive the data format is, it can be easy or hard;
- I use dat-schema to generate OCaml code that generates something that Kalandralang understands (whether I, or someone, then hosts this data for Kalandralang to fetch, or whether Kalandralang just asks the user for the location of the GGPK file, is also a question).
I would obviously have enough time for the first solution (it's just a URL to change). I would also have time for the second for sure. I would probably have time for the third solution, depending on how different the file structures are. I would probably have a lot of fun doing the fourth or fifth option, but I don't foresee myself having enough time for it in the near future.
I hope this answers your question. If you think you can hosts JSON files in some format somewhere, I'll be happy to discuss the file format to see how hard it would be to migrate Kalandralang.
There is also the question of long-term sustainability; do you happen to know:
- why RePoE is no longer maintained?
- whether dat-format will likely be maintained?
- is there some community consensus about how to proceed?
https://github.com/brather1ng/RePoE is no longer maintained.
I looked into it and found a lot of promising tools to get the data myself, including crafting costs for harvest recipes (for future changes and cost of harvest crafts).
Lightweight tool to look around the data: https://snosme.github.io/poe-dat-viewer/
poe dat viewer also offers a library to extraxt data from the patch cdn of poe (the complete ggpk file is not needed with this method) following https://github.com/poe-tool-dev/dat-schema it is possible to export via json.
Question is if the switch is necessary.
I could provide the additional tooling for the data, and a up to date repository for kalandralang to pull the data from.
Further more this would change the json structure of the files. Probably could do a hacky translation to RePoE Format, but it could break in the future or kalandralang uses the new format of the future repository (which will basically be the dat-schema format)
For the language files (mod stats) I have to cook my own thing, could reimplement RePoE schema or an entirely different one, depending on your input.