Automatically publish HTML pages with embedded Croissant mark-up from Github datasets

[benjelloun]

Croissant datasets in our Github repository are represented as standalone JSON-LD files. In order to make them crawlable, they should be made accessible as web pages with the JSON-LD embedded inside a script tag.

Instead of creating those pages manually, we should automate the process by:

1. Creating a Python script that converts the JSON-LD standalone file into the desired HTML + JS representation
2. Creating a Github action that calls the script and commits a new or updated HTML file whenever a dataset description is changed in the repository.