Additional Lithogeochemical plots [FEATURE]

[Antoine-Cate]

Is your feature request related to a problem? Please describe.
In addition to the TAS, other commonly used diagrams could be added.

Describe the solution you'd like
Using a template, the following diagrams could be added:

• Box Plot alteration plot (Large et al, 2001) and similar alteration plots using indexes
• Modified Wynchester and Floyd diagram (Barrett, T. J., and MacLean, W. H., 1994)
• Pearce basalt discrimination diagrams
• Verma and Agrawal (2011) tectonic discrimination diagrams
• Ross and Bedard (2009) magmatic affinity diagram
• Hastie et al (2007) arc rock discrimination diagram
• plenty of other diagrams
  Using a template, anyone could contribute and add their diagram of interest.

Describe alternatives you've considered
I do not see other alternatives to add new diagrams.

Additional context
I would love to participate building this diagrams library.

Relevant References

[morganjwilliams]

These would be some good additions to pyrolite.plot.templates, which currently hosts the TAS and Pearce diagrams.

I'll have a think about a good way to do this and keep all the metadata (axes variables, values, limits and descriptions of conditions for which the diagrams were intended to be applied - some are more restrictive!).

My first approach would be to use a subclassed template class with a method to add it to an axes. Then write up a bit of a guide for constructing these templates and submitting them as part of the docs development guide.

@Antoine-Cate and others who come across this issue, please feel free add your own (design and specific template) suggestions here too!

[morganjwilliams]

The Pearce diagrams are constructed from specified geometric components which preserve some of this information (i.e. line labels etc), but there are likely a much better ways to achieve this. It'd also be great to make this a bit more readable and approachable for python beginners.

[morganjwilliams]

Another thing to keep in mind is that pyrolite will eventually move towards having multiple plotting backends (to leverage interactive plotting, likely largely through pandas-based interfaces to start with). This is part of the reason that the Pearce diagrams were constructed with the geometry decoupled from the plotting.