Check that SAX and PAA are fitted (#69)

docs/more/versions/v1.1.rst

@@ -16,6 +16,10 @@ Wildboar 1.1 requires Python 3.8+, numpy 1.17.3+, scipy 1.3.2+ and scikit-learn
 Version 1.1.1
 =============
 
+Major changes
+-------------
+Correctly depend on the oldest supported Numpy release.
+
 Changelog
 ---------
 
@@ -28,6 +32,14 @@ Changelog
 
  - |Fix| Correctly return ``max_motif`` motifs from :func:`annotate.motifs`.
 
+ .. grid-item-card:: 
+
+ :mod:`wildboar.transform`
+ ^^^ 
+
+ - |Fix| Check that :class:`transform.SAX` and :class:`transform.PAA` is fitted
+ before ``transform``.
+
 
 Version 1.1.0
 =============

src/wildboar/transform/_sax.py

@@ -7,7 +7,7 @@
 import numpy as np
 from scipy.stats import norm, uniform
 from sklearn.base import TransformerMixin
-from sklearn.utils.validation import check_scalar
+from sklearn.utils.validation import check_is_fitted, check_scalar
 
 from wildboar.utils.validation import check_array, check_option
 
@@ -147,6 +147,7 @@ def fit(self, x, y=None):
  return self
 
  def transform(self, x):
+ check_is_fitted(self)
  x = self._validate_data(x, reset=False, dtype=float)
  x_paa = self.paa_.transform(x)
  thresholds = self.binning_.get_thresholds(x, estimate=self.estimate)
@@ -156,6 +157,7 @@ def transform(self, x):
  return x_out
 
  def inverse_transform(self, x):
+ check_is_fitted(self)
  if self.estimate:
  raise ValueError("Unable to inverse_transform with estimate=True")
 
@@ -208,10 +210,12 @@ def fit(self, x, y=None):
  return self
 
  def transform(self, x):
+ check_is_fitted(self)
  x = self._validate_data(x, dtype=float, reset=False)
  return self.interval_transform_.transform(x)
 
  def inverse_transform(self, x):
+ check_is_fitted(self)
  x = check_array(x, dtype=float)
 
  x_inverse = np.empty((x.shape[0], self.n_timesteps_in_), dtype=float)