Pyneb

Add a test to avoid crash using new version of numpy.genfromtxt that read strings as string and not bytes anymore.

Add new test facilities
Update flow
Update calling sequence for AI4Neb to be optional (need to add pn.config.import_AI4Neb() to use ANN)

* Migrate the setup to pyproject.toml. Minimalist version of setup.py is kept.
* Remove the version.py file, the version is now in pyproject.toml
* Move the import ai4neb into a specific config method: pn.config.import_AI4Neb() before using the Machine Learning tools. This avoid annoying message about Tensorflow to appear and speedup the import of pyneb when not using ML facilities.
* Change the name of the package to pyneb instead of PyNeb (seems to be mandatory soon).

Add Calzetti extinction laws

Correct bug in utils/manage_atomic_data/extract_flt : bytes and str confusion.

Change the way the low and High density ratios are obtained. Old way was using the statistical weights ratio, but this only works for S+-type of density diagnostics. Not for IR line ratios.
Add a Notebook in the doc directory to explore Critical densities.