Q Raw data of docs for JsonSchema

I want to collect CSS docs and save as JsonSchema like

{
  "border": {"propertyNames": {"enum": ["left","right"]}} 
}

Is somewhere more structured data than HTML values @ client/build/**/index.json ? To not produce weird exporter.

[peterbe]

@peterbe Oops, sorry. Can you say which one is correct?

Hehe. I was wrong in saying wrong repo. Sorry. If you want to get structured data out of MDN there are some options:

1. The Stumptown initiative which is about turning unstructured blobs of documentation (in HTML with macros) into Markdown with structured data. This project might be hard to understand because it’s early stages and a lot of flux.
2. The browser-compat-data is more about compatibility than anything.
3. What yari produces is just blobs of text to support the rendering of complete HTML pages. Those index.json are considered implementation details. Not an API. But you can still attempt to use it and do your own post-processing.

[peterbe]

There are sources in content/files/**/index.html files. On disk. And checked into git. The node content build ... command transforms that into index.html and index.json files. These can be found in client/build/**/index.json afterwards. On disk. These are all part of the .gitignore just like any build artifact like client/build/static/**