feat: bootstrapper for pandas

xeofs/pandas/__init__.py

@@ -1,2 +1,3 @@
 from .eof import EOF
 from .rotator import Rotator
+from .bootstrapper import Bootstrapper

xeofs/pandas/bootstrapper.py

@@ -0,0 +1,73 @@
+from typing import Optional
+
+import numpy as np
+import pandas as pd
+
+from ..models._base_bootstrapper import _BaseBootstrapper
+from .eof import EOF
+from ..utils.tools import squeeze
+
+
+class Bootstrapper(_BaseBootstrapper):
+ '''Short summary.'''
+
+ def __init__(
+ self, n_boot : int,
+ alpha : float = 0.05,
+ test_type : Optional[str] = 'one-sided'
+ ):
+ '''Short summary.
+
+ Parameters
+ ----------
+ n_boot : int
+ Description of parameter `n_boot`.
+ alpha : float
+ Description of parameter `alpha` (the default is 0.05).
+ test_type : Optional[str]
+ Description of parameter `test_type` (the default is 'one-sided').
+
+ Returns
+ -------
+ type
+ Description of returned object.
+
+ '''
+ super().__init__(n_boot=n_boot, alpha=alpha, test_type=test_type)
+
+ def bootstrap(self, model : EOF):
+ super().bootstrap(model)
+
+ def get_params(self):
+ return super().get_params()
+
+ def n_significant_modes(self):
+ return super().n_significant_modes()
+
+ def explained_variance(self):
+ expvar, expvar_mask = super().explained_variance()
+ expvar = pd.DataFrame(
+ expvar.T,
+ columns=pd.Index(self._params['quantiles'], name='quantile'),
+ index=self._model._idx_mode,
+ )
+ expvar_mask = pd.DataFrame(
+ expvar_mask,
+ columns=['is_significant'],
+ index=self._model._idx_mode[:-1]
+ )
+
+ return expvar, expvar_mask
+
+ def eofs(self):
+ eofs, eofs_mask = super().eofs()
+ eofs = [squeeze(self._model._tf.back_transform_eofs(e)) for e in eofs]
+ eofs_mask = squeeze(self._model._tf.back_transform_eofs(eofs_mask))
+
+ return eofs, eofs_mask
+
+ def pcs(self):
+ pcs, pcs_mask = super().pcs()
+ pcs = [squeeze(self._model._tf.back_transform_pcs(q)) for q in pcs]
+ pcs_mask = squeeze(self._model._tf.back_transform_pcs(pcs_mask))
+ return pcs, pcs_mask