Create tooling to translate Jupyter notebooks with IBM infrastructure

[Eric-Arellano]

The IBM translations pipeline from #670 will not be able to handle jupyter notebooks because it can't understand that file type.

So we will need to instead have some custom tooling that extracts the translatable strings from the notebook, sends it to get translated, then applies the translations properly. The start and end product should be a valid Jupyter notebook where only the markdown code blocks are translated. This notebook needs to be executable, including because users may download the translated notebook.

We can choose any key-value format like a PO file or JSON file.

Look at https://github.com/WittmannF/jupyter-translate for inspiration. We can also probably use https://github.com/Qiskit/documentation/blob/main/scripts/js/lib/markdownReader.ts

#2368 is a follow up to to also translate code comments.

[Eric-Arellano]

This was moved to closed source, since the tooling will live there. Indeed, we will only globalize markdown cells for now and use code similar to markdownReader.ts.