Several updates to the liquefaction module

[LanaTodorovic93]

This pull request:

• Adds several regional liquefaction models to the secondary perils module (@LanaTodorovic93 and @raoanirudh)
  • ZhuEtAl2017LiquefactionCoastal and ZhuEtAl2017LiquefactionGeneral
  • RashidianBaise2020Liquefaction
  • AllstadtEtAl2022Liquefaction
  • AkhlagiEtAl2021LiquefactionA and AkhlagiEtAl2021LiquefactionB (experimental)
  • Bozzoni2021LiquefactionEurope
  • TodorovicSilva2022NonParametric
• Adds new site parameters required by some of these new models to site.py
• Adds LiqOccur as a valid IMT, referring to the occurrence or non-occurrence of liquefaction at a site
• Adds basic documentation for the secondary perils module (originally written by @cossatot and extended by @LanaTodorovic93)
• Adds unit tests for the newly added models (@LanaTodorovic93)
• Adds a QA test case for the TodorovicSilva2022NonParametric model (@LanaTodorovic93)
• Includes some minor refactoring of the existing liquefaction module (@raoanirudh)

To do before merging:

• Add scikit-learn as a dependency, which is required by TodorovicSilva2022NonParametric (@vot4anto)
• Add onnxruntime as a dependency. Remove scikit-learn, as pickle is not required anymore. (@vot4anto)
• Complete documentation of the newly added models (@LanaTodorovic93)
• Add rock-slope failure analysis (extended by @LanaTodorovic93)
• Add corresponding tests (@LanaTodorovic93)
• Tutorials (@LanaTodorovic93)
• Add a scenario liquefaction demo (@LanaTodorovic93)
• Add a stochastic event-based liquefaction demo (@LanaTodorovic93)
• Check if keeping test_sep_suite_1.py makes sense, given that currently, all tests in this suite are skipped and now that all site parameters are expected to be input through a CSV file rather than as rasters (@cossatot). Removed by @LanaTodorovic93
• Update the changelog once the PR is approved.
• (Optional, can go into a future PR) Check if it makes sense to refactor sep/classes.py and sep/liquefaction/liquefaction.py to adopt a structure similar to the hazardlib/gsim module, with a base.py providing the abstract base class and each liquefaction model going into its own file (@micheles)

[micheles]

Congratulations for the hard work, I see nothing stopping the merge of the PR now :-)

About the question Check if it makes sense to refactor sep/classes.py and sep/liquefaction/liquefaction.py to adopt a structure similar to the hazardlib/gsim module, with a base.py providing the abstract base class and each liquefaction model going into its own file I would keep the code as it is. I don't see a strong reason to go in the same direction as hazardlib, certainly not for the moment.