Rename loss => elementwise_loss and full_objective => loss_function

Export SymbolicRegression.jl module

Add install function back with deprecation warning

Ensure julia is loaded as soon as possible

Remove Feynman problems and dataset

Remove generated version.py file

Remove Python 3.12 warning

Start third attempt at porting to PythonCall

Warn the user on 3.12

Move all deprecated functions to deprecated.py

Add pre-commit hook for isort

Add pre-commit hooks for unused imports

Patch scikit-learn to allow complex numbers

Moved julia helper functions out of sr.py

Remove silencer function from init

Add __version__ to package

Update parts of test to use ScikitLearn interface

Create scikit-learn API

Add workaround for statically-linked libpython

Add missing top-level import

Refactor lazy loading of torch and jax

Add torch format output; dont import jax/torch by default

Create torch export function

Move JAX export to separate file

Add JAX export functionality

Import FeynmanProblems into pysr namespace

Fix printed score; demonstrate best() function

Allow early quit + can pretty-print best equation

Create PyPI package under pysr