Check jupyter frontend widget availability before setting env to lab

[Koncopd]

#214

Unfortunately it doesn't work. About half of times it reports failure to access JupyterFrontEnd, despite it working later when it is actually needed and called.

On the other hand, trying to take into account the async nature and calling app.ready() (see this) coroutine and blocking the code execution before getting the result leads to infinite waiting times.

[github-actions]

🚀 Deployed on https://62ef9b4df3c82a74baad3fae--nbproject-3aa23rammb.netlify.app

[codecov]

Codecov Report

Merging #216 (be401de) into main (27f35e0) will decrease coverage by 0.44%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##             main     #216      +/-   ##
==========================================
- Coverage   88.20%   87.76%   -0.45%     
==========================================
  Files          20       20              
  Lines         882      891       +9     
==========================================
+ Hits          778      782       +4     
- Misses        104      109       +5     

Impacted Files	Coverage Δ
nbproject/dev/_jupyter_lab_commands.py	37.50% <25.00%> (-4.17%)	⬇️
nbproject/_header.py	87.83% <66.66%> (-1.06%)	⬇️
nbproject/_meta.py	94.54% <66.66%> (-1.61%)	⬇️

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[Koncopd]

By doing async i mean:

@run_sync # util to turn async into sync from nbclient
async def _check_frontend():
    global app
    if app is None:
        app = JupyterFrontEnd()

    await app.ready()

    return app.version != ""

and then using it in the header results in infinite waiting time sometimes. Obviously not applicable for us here.

[Koncopd]

This PR is just for the discussion on Monday.

[falexwolf]

Understood! Thank you for trying it!