Clarifying raw_mimetype?

[ickc]

trafficstars

Hi,

When experimenting in a Jupyterlab instance, if I change the "Raw NBConvert Format", it results in for example

  {
   "cell_type": "raw",
   "metadata": {
    "raw_mimetype": "text/restructuredtext",
    "tags": []
   },
   "source": []
  }

but from the doc here:

{
  "cell_type" : "raw",
  "metadata" : {
    # the mime-type of the target nbconvert format.
    # nbconvert to formats other than this will exclude this cell.
    "format" : "mime/type"
  },
  "source" : "[some nbformat output text]"
}

The difference is the former uses raw_mimetype as the key of the mime-type, and later uses format.

Could you clarify the usage of these 2? Thanks!

[bollwyvl]

Looks like a bug in jupyterlab, as in this repo, raw_mimetype does not appear anywhere except this issue. :shrug:

[mgeier]

If this is a bug in JupyterLab, it's also a bug in the Classic Notebook and in nbconvert, and has been there for many years, probably since the beginning.

nbconvert has several instances of raw_mimetype: https://github.com/jupyter/nbconvert/search?q=raw_mimetype

[bollwyvl]

Ok, so less bug and more undefined behavior according to nbformat as it appears in the free-for-all metadata section, under an un-reserved key, but is implemented in nbconvert, lab and notebook, while format is a reserved key but is not implemented by any of those downstreams.

[ickc]

Should the solution be releasing a major version of nbconvert that change the spec to raw_mimetype to match the "real world behavior in the wild"?

This will help when someone is implementing according to spec but unaware of the actual usage out there.

Thanks everyone.