[MNT] Add codespell support (config, workflow to detect/not fix) and make it fix a "few" typos

.github/actions/numba_cache/action.yml

@@ -62,6 +62,6 @@ runs:
         path: ${{ github.workspace }}/.numba_cache
         # Try restore using today's date
         key: numba-${{ inputs.cache_name }}-${{ inputs.runner_os }}-${{ inputs.python_version }}-${{ env.CURRENT_DATE }}
-        # If cant restore with today's date try another cache (without date)
+        # If can't restore with today's date try another cache (without date)
         restore-keys: |
           numba-${{ inputs.cache_name }}-${{ inputs.runner_os }}-${{ inputs.python_version }}-

.github/utilities/codespell_ignore_words.txt

@@ -0,0 +1,4 @@
+fpr
+mape
+recuse
+strat

.github/workflows/pr_precommit.yml

@@ -60,3 +60,16 @@ jobs:
         with:
           commit_message: Automatic `pre-commit` fixes
           commit_user_name: aeon-actions-bot[bot]
+
+  codespell-annotations:
+    runs-on: ubuntu-24.04
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Annotate locations with typos
+        uses: codespell-project/codespell-problem-matcher@v1
+
+      - name: Codespell
+        uses: codespell-project/actions-codespell@v2

.pre-commit-config.yaml

@@ -78,3 +78,11 @@ repos:
     hooks:
       - id: check-manifest
         stages: [ manual ]
+
+  - repo: https://github.com/codespell-project/codespell
+    # Configuration for codespell is in pyproject.toml
+    rev: v2.4.1
+    hooks:
+      - id: codespell
+        additional_dependencies:
+          - tomli  # for python_version < '3.11'

aeon/anomaly_detection/series/distance_based/_kmeans.py

@@ -42,7 +42,7 @@ class KMeansAD(BaseSeriesAnomalyDetector):
 
     stride : int, default=1
         The stride of the sliding window. The stride determines how many time points
-        the windows are spaced appart. A stride of 1 means that the window is moved one
+        the windows are spaced apart. A stride of 1 means that the window is moved one
         time point forward compared to the previous window. The larger the stride, the
         fewer windows are created, which leads to noisier anomaly scores.
 

aeon/anomaly_detection/series/distance_based/_merlin.py

@@ -102,26 +102,26 @@ def _predict(self, X):
 
         r = 2 * np.sqrt(self.min_length)
         distances = np.full(len(lengths), -1.0)
-        indicies = np.full(len(lengths), -1)
+        indices = np.full(len(lengths), -1)
 
-        indicies[0], distances[0] = self._find_index(X, lengths[0], r, np.multiply, 0.5)
+        indices[0], distances[0] = self._find_index(X, lengths[0], r, np.multiply, 0.5)
 
         for i in range(1, min(5, len(lengths))):
             r = distances[i - 1] * 0.99
-            indicies[i], distances[i] = self._find_index(
+            indices[i], distances[i] = self._find_index(
                 X, lengths[i], r, np.multiply, 0.99
             )
 
         for i in range(min(5, len(lengths)), len(lengths)):
             m = mean(distances[i - 5 : i])
             s = std(distances[i - 5 : i])
             r = m - 2 * s
-            indicies[i], distances[i] = self._find_index(
+            indices[i], distances[i] = self._find_index(
                 X, lengths[i], r, np.subtract, s
             )
 
         anomalies = np.zeros(X.shape[0], dtype=bool)
-        for i in indicies:
+        for i in indices:
             if i > -1:
                 anomalies[i] = True
 

aeon/anomaly_detection/series/outlier_detection/_stray.py

@@ -20,7 +20,7 @@ class STRAY(BaseSeriesAnomalyDetector):
     ability to detect clusters of outliers in multidimensional data without
     requiring a model of the typical behavior of the system. However, it suffers
     from some limitations that affect its accuracy. STRAY is an extension of
-    HDoutliers that uses extreme value theory for the anomolous threshold
+    HDoutliers that uses extreme value theory for the anomalous threshold
     calculation, to deal with data streams that exhibit non-stationary behavior.
 
     Parameters
@@ -39,7 +39,7 @@ class STRAY(BaseSeriesAnomalyDetector):
         Proportion of possible candidates for outliers. This defines the starting point
         for the bottom up searching algorithm.
     size_threshold : int, default=50
-        Sample size to calculate an emperical threshold.
+        Sample size to calculate an empirical threshold.
     outlier_tail : str {"min", "max"}, default="max"
         Direction of the outlier tail.
 

aeon/base/_estimators/compose/collection_ensemble.py

@@ -42,7 +42,7 @@ class BaseCollectionEnsemble(ComposableEstimatorMixin, BaseCollectionEstimator):
         Only used if weights is a float. The method used to generate a performance
         estimation from the training data set i.e. cross-validation.
         If None, predictions are made using that estimators fit_predict or
-            fit_predict_proba methods. These are somtimes overridden for efficient
+            fit_predict_proba methods. These are sometimes overridden for efficient
             performance evaluations, i.e. out-of-bag predictions.
         If int or sklearn object input, the parameter is passed directly to the cv
             parameter of the cross_val_predict function from sklearn.

aeon/benchmarking/results_loaders.py

@@ -201,7 +201,7 @@ def estimator_alias(name: str) -> str:
 def get_available_estimators(
     task: str = "classification", as_list: bool = False
 ) -> Union[pd.DataFrame, list]:
-    """Get a DataFrame of estimators avialable for a specific learning task.
+    """Get a DataFrame of estimators available for a specific learning task.
 
     Parameters
     ----------
@@ -251,7 +251,7 @@ def get_estimator_results(
 
     Parameters
     ----------
-    estimators : str ot list of str
+    estimators : str or list of str
         Estimator name or list of estimator names to search for. See
         get_available_estimators, aeon.benchmarking.results_loading.NAME_ALIASES or
         the directory at path for valid options.

aeon/benchmarking/stats.py

@@ -57,7 +57,7 @@ def nemenyi_test(ordered_avg_ranks, n_datasets, alpha):
     ordered_avg_ranks : np.array
         Average ranks of estimators.
     n_datasets : int
-        Mumber of datasets.
+        Number of datasets.
     alpha : float
         alpha level for Nemenyi test.
 

aeon/benchmarking/tests/test_resampling.py

@@ -107,7 +107,7 @@ def test_resample_data_indices():
     concat = np.concatenate((train_indices, test_indices), axis=0)
     assert len(np.unique(concat)) == len(concat)
 
-    # expected indicies after resampling
+    # expected indices after resampling
     np.testing.assert_array_equal(
         concat,
         [
@@ -277,7 +277,7 @@ def test_stratified_resample_data_indices():
     concat = np.concatenate((train_indices, test_indices), axis=0)
     assert len(np.unique(concat)) == len(concat)
 
-    # expected indicies after resampling
+    # expected indices after resampling
     np.testing.assert_array_equal(
         concat,
         [

aeon/classification/compose/_ensemble.py

@@ -33,7 +33,7 @@ class ClassifierEnsemble(BaseCollectionEnsemble, BaseClassifier):
         Only used if weights is a float. The method used to generate a performance
         estimation from the training data set i.e. cross-validation.
         If None, predictions are made using that estimators fit_predict or
-            fit_predict_proba methods. These are somtimes overridden for efficient
+            fit_predict_proba methods. These are sometimes overridden for efficient
             performance evaluations, i.e. out-of-bag predictions.
         If int or sklearn object input, the parameter is passed directly to the cv
             parameter of the cross_val_predict function from sklearn.

aeon/classification/deep_learning/_disjoint_cnn.py

@@ -25,42 +25,42 @@ class DisjointCNNClassifier(BaseDeepClassifier):
         Number of 1+1D Convolution layers.
     n_filters : int or list of int, default = 64
         Number of filters used in convolution layers. If
-        input is set to a list, the lenght should be the same
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     kernel_size : int or list of int, default = [8, 5, 5, 3]
         Size of convolution kernel. If
-        input is set to a list, the lenght should be the same
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     dilation_rate : int or list of int, default = 1
         The dilation rate for convolution. If
-        input is set to a list, the lenght should be the same
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     strides : int or list of int, default = 1
         The strides of the convolution filter. If
-        input is set to a list, the lenght should be the same
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     padding : str or list of str, default = "same"
         The type of padding used for convolution. If
-        input is set to a list, the lenght should be the same
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     activation : str or list of str, default = "elu"
         Activation used after the convolution. If
-        input is set to a list, the lenght should be the same
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     use_bias : bool or list of bool, default = True
-        Whether or not ot use bias in convolution. If
-        input is set to a list, the lenght should be the same
+        Whether or not to use bias in convolution. If
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     kernel_initializer: str or list of str, default = "he_uniform"
         The initialization method of convolution layers. If
-        input is set to a list, the lenght should be the same
+        input is set to a list, the length should be the same
         as `n_layers`, if input is int the a list of the same
         element is created of length `n_layers`.
     pool_size: int, default = 5

aeon/classification/deep_learning/_fcn.py

@@ -36,7 +36,7 @@ class FCNClassifier(BaseDeepClassifier):
     activation : str or list of str, default = "relu"
         Activation used after the convolution.
     use_bias : bool or list of bool, default = True
-        Whether or not ot use bias in convolution.
+        Whether or not to use bias in convolution.
     n_epochs : int, default = 2000
         The number of epochs to train the model.
     batch_size : int, default = 16

aeon/classification/deep_learning/_inception_time.py

@@ -354,7 +354,7 @@ def _predict_proba(self, X) -> np.ndarray:
     def load_model(self, model_path, classes):
         """Load pre-trained classifiers instead of fitting.
 
-        When calling this function, all funcationalities can be used
+        When calling this function, all functionalities can be used
         such as predict, predict_proba, etc. with the loaded models.
 
         Parameters

aeon/classification/deep_learning/_lite_time.py

@@ -286,7 +286,7 @@ def _predict_proba(self, X) -> np.ndarray:
     def load_model(self, model_path, classes):
         """Load pre-trained classifiers instead of fitting.
 
-        When calling this function, all funcationalities can be used
+        When calling this function, all functionalities can be used
         such as predict, predict_proba, etc. with the loaded models.
 
         Parameters
@@ -448,7 +448,7 @@ class IndividualLITEClassifier(BaseDeepClassifier):
     References
     ----------
     ..[1] Ismail-Fawaz et al. LITE: Light Inception with boosTing
-    tEchniques for Time Series Classificaion, IEEE International
+    tEchniques for Time Series Classification, IEEE International
     Conference on Data Science and Advanced Analytics, 2023.
     ..[2] Ismail-Fawaz, Ali, et al. "Look Into the LITE
     in Deep Learning for Time Series Classification."

aeon/classification/dictionary_based/_boss.py

@@ -719,11 +719,11 @@ def boss_distance(X, Y, i, XX_all=None, XY_all=None):
     """Find the distance between two histograms.
 
     This returns the distance between first and second dictionaries, using a non-
-    symmetric distance measure. It is used to find the distance between historgrams
+    symmetric distance measure. It is used to find the distance between histograms
     of words.
 
     This distance function is designed for sparse matrix, represented as either a
-    dictionary or an arrray. It only measures the distance between counts present in
+    dictionary or an array. It only measures the distance between counts present in
     the first dictionary and the second. Hence dist(a,b) does not necessarily equal
     dist(b,a).
 

aeon/classification/dictionary_based/_cboss.py

@@ -28,7 +28,7 @@ class ContractableBOSS(BaseClassifier):
     described in [2]_.
 
     Overview: Input "n" series of length "m" and cBOSS randomly samples
-    `n_parameter_samples` parameter sets, evaluting each with LOOCV. It then
+    `n_parameter_samples` parameter sets, evaluating each with LOOCV. It then
     retains `max_ensemble_size` classifiers with the highest accuracy
     There are three primary parameters:
         - alpha: alphabet size

aeon/classification/dictionary_based/_redcomets.py

@@ -36,7 +36,7 @@ class REDCOMETS(BaseClassifier):
         RED CoMETS variant to use from {1, 2, 3, 4, 5, 6, 7, 8, 9} to use as per [1]_.
         Defaults to RED CoMETS-3. Variants 4-9 only support multivariate problems.
     perc_length : int or float, default=5
-        Percentage of time series length used to determinne number of lenses during
+        Percentage of time series length used to determine number of lenses during
         pair selection.
     n_trees : int, default=100
         Number of trees used by each random forest sub-classifier.
@@ -185,12 +185,12 @@ def _build_univariate_ensemble(self, X, y):
         Returns
         -------
         sfa_transforms :
-            List of ``SFAFast()`` instances with random word length and alpabet size
+            List of ``SFAFast()`` instances with random word length and alphabet size
         sfa_clfs :
             List of ``(RandomForestClassifier(), weight)`` tuples fitted on `SFAFast`
             transformed training data
         sax_transforms :
-            List of ``SAX()`` instances with random word length and alpabet size
+            List of ``SAX()`` instances with random word length and alphabet size
         sax_clfs :
             List of ``(RandomForestClassifier(), weight)`` tuples fitted on `SAX`
             transformed training data
@@ -321,13 +321,13 @@ def _build_dimension_ensemble(self, X, y):
         Returns
         -------
         sfa_transforms : list
-            List of lists of ``SFAFast()`` instances with random word length and alpabet
-            size
+            List of lists of ``SFAFast()`` instances with random word length and
+            alphabet size
         sfa_clfs : list
             List of lists of ``(RandomForestClassifier(), weight)`` tuples fitted on
             `SFAFast` transformed training data
         sax_transforms : list
-            List of lists of ``SAX()`` instances with random word length and alpabet
+            List of lists of ``SAX()`` instances with random word length and alphabet
             size
         sax_clfs : list
             List of lists ``(RandomForestClassifier(), weight)`` tuples fitted on `SAX`

aeon/classification/dictionary_based/_tde.py

@@ -618,7 +618,7 @@ class IndividualTDE(BaseClassifier):
     bigrams : bool, default=False
         Whether to record word bigrams in the SFA transform.
     dim_threshold : float, default=0.85
-        Accuracy threshold as a propotion of the highest accuracy dimension for words
+        Accuracy threshold as a proportion of the highest accuracy dimension for words
         extracted from each dimensions. Only applicable for multivariate data.
     max_dims : int, default=20
         Maximum number of dimensions words are extracted from. Only applicable for

aeon/classification/early_classification/_probability_threshold.py

@@ -47,7 +47,7 @@ class ProbabilityThresholdEarlyClassifier(BaseEarlyClassifier):
         List of integer time series time stamps to build classifiers and allow
         predictions at. Early predictions must have a series length that matches a value
         in the _classification_points List. Duplicate values will be removed, and the
-        full series length will be appeneded if not present.
+        full series length will be appended if not present.
         If None, will use 20 thresholds linearly spaces from 0 to the series length.
     n_jobs : int, default=1
         The number of jobs to run in parallel for both `fit` and `predict`.
@@ -73,7 +73,7 @@ class ProbabilityThresholdEarlyClassifier(BaseEarlyClassifier):
     state_info : 2d np.ndarray (4 columns)
         Information stored about input instances after the decision-making process in
         update/predict methods. Used in update methods to make decisions based on
-        the resutls of previous method calls.
+        the results of previous method calls.
         Records in order: the time stamp index, the number of consecutive decisions
         made, the predicted class and the series length.
 

aeon/classification/early_classification/_teaser.py

@@ -56,7 +56,7 @@ class TEASER(BaseEarlyClassifier):
         List of integer time series time stamps to build classifiers and allow
         predictions at. Early predictions must have a series length that matches a value
         in the _classification_points List. Duplicate values will be removed, and the
-        full series length will be appeneded if not present.
+        full series length will be appended if not present.
         If None, will use 20 thresholds linearly spaces from 0 to the series length.
     n_jobs : int, default=1
         The number of jobs to run in parallel for both `fit` and `predict`.
@@ -82,7 +82,7 @@ class TEASER(BaseEarlyClassifier):
     state_info : 2d np.ndarray (4 columns)
         Information stored about input instances after the decision-making process in
         update/predict methods. Used in update methods to make decisions based on
-        the resutls of previous method calls.
+        the results of previous method calls.
         Records in order: the time stamp index, the number of consecutive decisions
         made, the predicted class and the series length.
 

aeon/classification/early_classification/base.py

@@ -66,7 +66,7 @@ def __init__(self):
 
         """
         An array containing the state info for each decision in X from update and
-        predict methods. Contains classifier dependant information for future decisions
+        predict methods. Contains classifier dependent information for future decisions
         on the data and information on when a cases decision has been made. Each row
         contains information for a case from the latest decision on its safety made in
         update/predict. Successive updates are likely to remove rows from the
@@ -308,7 +308,7 @@ def get_state_info(self):
         Returns
         -------
         An array containing the state info for each decision in X from update and
-        predict methods. Contains classifier dependant information for future decisions
+        predict methods. Contains classifier dependent information for future decisions
         on the data and information on when a cases decision has been made. Each row
         contains information for a case from the latest decision on its safety made in
         update/predict. Successive updates are likely to remove rows from the

aeon/classification/ordinal_classification/_ordinal_tde.py

@@ -592,7 +592,7 @@ class IndividualOrdinalTDE(BaseClassifier):
     bigrams : bool, default=False
         Whether to record word bigrams in the SFA transform.
     dim_threshold : float, default=0.85
-        Accuracy threshold as a propotion of the highest accuracy dimension for words
+        Accuracy threshold as a proportion of the highest accuracy dimension for words
         extracted from each dimensions. Only applicable for multivariate data.
     max_dims : int, default=20
         Maximum number of dimensions words are extracted from. Only applicable for

aeon/classification/shapelet_based/_sast.py

@@ -28,7 +28,7 @@ class SASTClassifier(BaseClassifier):
     ----------
     length_list : int[], default = None
         an array containing the lengths of the subsequences to be generated.
-        If None, will be infered during fit as np.arange(3, X.shape[1])
+        If None, will be inferred during fit as np.arange(3, X.shape[1])
     stride : int, default = 1
         the stride used when generating subsquences
     nb_inst_per_class : int default = 1