Parallelize pyodide based docs

[philippjfr]

Allows parallelization of pyodide based docs by:

• Storing a state dictionary per document
• Writing resources out to disk (guarded by a lockfile) to ensure we can gather all required resources
• Making the subprocesses daemons

[hoxbro]

I'm curious: Is the processing still needed when using sphinx parallelization?

[philippjfr]

I'm curious: Is the processing still needed when using sphinx parallelization?

Yes, absolutely. Each pyodide execution environment expects a clean slate. In theory we could try to clean up state but that seems very prone to issues.