Upload JSON directly and split it into multiple files

[aaronfranke]

It would help with version control and viewing history (as this is GitHub...)

Splitting into multiple files is necessary in the long term, since the maximum file size is 100 MB.

[DanMan]

Agreed on the raw upload. One can give it a generic name then, too, without the date in it - easier to consume.

[bryanmr]

This is an annoyance for me on the Steam library categorization script I am writing. I'd love for there to be a consistent file name to grab.

[DanMan]

@bryanmr This is hosted here but the reports are up to a month behind because that's how often dumps are released. Shout out to @maxpoulin64.

[bryanmr]

@DanMan The raw JSON that is hosted there saves me from parsing the original files or having to put it up myself on Github, which now that I am thinking of it was the better option. :)

FYI, you can find the raw json file for ProtonDB on the page you linked, which is probably better than me having a script that hits the API a few thousand times in a few seconds whenever someone with a large library runs it.

[maxpoulin64]

@bryanmr Would an API that accepts many game IDs be helpful for your use case? That way you wouldn't need to download the whole JSON file and could do it in a single API request, so win win. If so open an issue on my GitHub page. Glad to see more people using it 🙂

I'll take the occasion to repeat my offer to build, host and maintain an official API so we don't need to work off monthly database dumps. I feel a month old is a lot and might not reflect the latest status of Proton, DXVK and Wine. All I need is sufficient access to set up a database slave on my server and take it on from there. I've jumped to Linux full time 12 years ago and would love to contribute to expanding the Linux gaming universe. This thread shows all the cool stuff people could build on top of the ProtonDB reports!

[bryanmr]

@maxpoulin64 My current command to parse the JSON is jq -r '.[] | [.appId, .rating] | join(",")' and it just spits out a CSV with the ratings, which I count up to get the mean and mode. It's not object oriented or responsive, it just produces a config file for Steam to parse.

I'd assume that something like Enhanced Steam would be one of the good use cases, where you integrate the information into the store pages for games on what their ratings are. That would be a fun project, on top of the other ones I am wanting to do.

[DanMan]

FYI, you can find the raw json file for ProtonDB on the page you linked

I'm aware, it's why I posted it.

integrate the information into the store pages for games on what their ratings are

Way ahead of you. ;-)