This includes models that combine some quadratic aspect & a cross term.
Remove random test

the selection for the model with minimal aic only takes into account models with with finite aic (there are often several models with infinite aic)

[contrasting] aic and bic test ok for my first model combining a linearity in 2 different coordinates.

[contrasting] fix polynomial_margin_model.py creation. extend the form_dict for spatio temporal models.

[contrasting] add quadratic model for gumbel/gev and location/scale. modify code for the name of the models

[contrasting] add assertion that the max_abs_change is different from np.inf otherwise it makes the computer freeze
IMPORTANT MODIFICATION, THAT MIGHT CHANGE PAPER 1: change unconstrained_estimator_gev_params to unconstrained_estimator_gev_params_last_year

[exceeding] add full diagnositic plot (qqplot & intensity plots) for gumbel stationary & add gumbel gof tests from r package

[exceeding] add full diagnositic plot (qqplot & intensity plots) for gumbel stationary & add gumbel gof tests from r package