v0.1.3

Separated the C++ wheels from the python library. Added everything related to the analysis of the input data into the python library.

The full python pipeline is now in place:

pip install mag-cusps

import numpy as np
import mag_cusps as cusps

# define J, Rho, X, Y, Z, earth_pos

J_processed: np.ndarray = cusps.preprocess( J, X, Y, Z, new_shape )
Rho_processed: np.ndarray = cusps.preprocess( Rho, X, Y, Z, new_shape )
J_norm_processed: np.ndarray = np.linalg.norm( J_processed, axis=3 )

BS = cusps.get_bowshock( Rho_processed, ... )
MP = cusps.get_interest_points( J_norm_processed, Rho_processed, ... )

MP_res = cusps.fit_to_analytical( MP, ... )
if MP_res is None:
    raise Exception()
MP_params, MP_cost = MP_res

analytics = cusps.analyse( ... )
model = cusps.load_pretrained_model( ... )
quality_score = model.predict(analytics)                               ### <- the quality scored associated with the input data
uncertainty = model.get_sample_uncertainty(analytics)       ### <- the uncertainty associated with that score