Import fails with darts >= 0.35

[JWuzyk]

The _LikelihoodMixin class was removed from darts in version 0.35, this causes conformal tights to error on import when the latest version of darts is used since this class is used for the DartsForecaster.

See 0.35 release notes here
" Removed darts.models.forecasting.regression_model._LikelihoodMixin. Use dedicated Likelihood models instead."

[lsorber]

Thanks for reporting this! That's won't be the easiest change to work around if we want to continue to support older versions of darts.

[doxalogic]

Either the u8darts version doesn't match due to RegressionModel or the numpy version doesn't match if we pin to u8darts==0.28.0

Can there be a working requirements.txt ?

ValueError                                Traceback (most recent call last)
[/tmp/ipython-input-1911543797.py](https://localhost:8080/#) in <cell line: 0>()
----> 1 from conformal_tights import ConformalCoherentQuantileRegressor
      2 from sklearn.datasets import fetch_openml
      3 from sklearn.model_selection import train_test_split
      4 from xgboost import XGBRegressor
      5 

9 frames
/usr/local/lib/python3.12/dist-packages/pmdarima/utils/_array.pyx in init pmdarima.utils._array()

ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject