Multiple features (update backend to 2.0)

[MilesCranmer]

Features:

• n-ary operators
• support for user-passed "guesses" of equations
• Mooncake and Enzyme support as autodiff backends
• feature node mutation

This also fundamentally changes some of the mutation operators to support the new n-arity operators. For example, a "rotation" of a binary tree has been generalised in such a way that its behavior changes on binary trees as well.

Similarly, the sampling patterns for other mutations has been changed, so as to improve the uniformity of mutations.

TODO:

• Do we have to use safe operators in the guesses? Or is that not needed? (Safe operators are used automatically via options.operators)
• Can we do complexity_of_operators for operators that have safe versions?
• Add examples to docs for n-arity operators
• Add some 3-arity operators to the docs and to the builtin sympy mappings (muladd, max, etc.)
• Get working with pytorch/jax export?

[codecov]

Codecov Report

❌ Patch coverage is 95.79832% with 5 lines in your changes missing coverage. Please review.
✅ Project coverage is 94.21%. Comparing base (c64ab94) to head (41fad00).
⚠️ Report is 50 commits behind head on master.

Files with missing lines	Patch %	Lines
pysr/julia_extensions.py	60.00%	2 Missing ⚠️
pysr/sr.py	98.16%	2 Missing ⚠️
pysr/_cli/main.py	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #999      +/-   ##
==========================================
+ Coverage   93.55%   94.21%   +0.66%     
==========================================
  Files          21       21              
  Lines        1566     1642      +76     
==========================================
+ Hits         1465     1547      +82     
+ Misses        101       95       -6     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.