handwriting_features.features package

Subpackages

• handwriting_features.features.configuration package
  • Submodules
  • handwriting_features.features.configuration.mapping module
    • HandwritingFeaturesMapping
      • HandwritingFeaturesMapping.map()
  • handwriting_features.features.configuration.settings module
    • HandwritingFeaturesSettings
      • HandwritingFeaturesSettings.axes
      • HandwritingFeaturesSettings.get_feature_argument_default()
      • HandwritingFeaturesSettings.get_feature_argument_options()
      • HandwritingFeaturesSettings.get_feature_argument_type()
      • HandwritingFeaturesSettings.settings
      • HandwritingFeaturesSettings.statistics
  • Module contents
• handwriting_features.features.exceptions package
  • Submodules
  • handwriting_features.features.exceptions.mapping module
    • FeatureNameNotInMappingError
  • handwriting_features.features.exceptions.validation module
    • FeatureArgumentInvalidTypeError
    • FeatureArgumentMissingError
    • FeatureArgumentUnsupportedValueError
    • FeatureNameInvalidTypeError
    • FeatureNameMissingError
    • FeatureNameUnsupportedError
    • StatisticsForSingleValuedFeatureError
  • Module contents
• handwriting_features.features.implementation package
  • Subpackages
    • handwriting_features.features.implementation.conventional package
      • Subpackages
      • Submodules
      • handwriting_features.features.implementation.conventional.composite module
      • handwriting_features.features.implementation.conventional.dynamic module
      • handwriting_features.features.implementation.conventional.kinematic module
      • handwriting_features.features.implementation.conventional.spatial module
      • handwriting_features.features.implementation.conventional.temporal module
      • Module contents
  • Module contents
• handwriting_features.features.validation package
  • Module contents
    • HandwritingFeaturesValidation
      • HandwritingFeaturesValidation.features_settings
      • HandwritingFeaturesValidation.validate()

Submodules

handwriting_features.features.base module

class handwriting_features.features.base.HandwritingFeaturesBase(sample_wrapper, **config)

Bases: object

Base class for handwriting features

classmethod after_computation(feature, statistics=None)

Applies the after-computation hook.

Parameters

• feature (numpy.ndarray) – computed feature

• statistics (Any[str, list, tuple], optional) – statistics to be computed, defaults to None

Returns

features

Return type

numpy.ndarray

classmethod before_computation(method, kwargs=None)

Applies the before-computation hook.

Parameters

• method (callable) – feature computation method to be applied

• kwargs (dict) – feature computation-specific kwargs

Returns

kwargs

Return type

dict

classmethod compute(sample_wrapper, method, statistics=None, **kwargs)

Applies the computation (including before-after computational hooks).

Parameters

• sample_wrapper (HandwritingSampleWrapper) – sample wrapper object

• method (callable) – feature computation method to be applied

• statistics (iterable, optional) – statistics to be computed, defaults to None

• kwargs (**kwargs) – feature computation-specific kwargs

Returns

computed features

Return type

numpy.ndarray

classmethod from_json(path, labels=None, **config)

Initializes HandwritingFeatures object from a JSON file.

Parameters

• path (str) – path to a JSON file

• labels (list, optional) – labels for the data values

• config (**kwargs) – common configuration

Returns

HandwritingFeatures object

Return type

HandwritingFeatures

classmethod from_list(values, labels=None, **config)

Initializes HandwritingFeatures object from a list.

Parameters

• values (list) – data values

• labels (list, optional) – labels for the data values

• config (**kwargs) – common configuration

Returns

HandwritingFeatures object

Return type

HandwritingFeatures

classmethod from_numpy_array(values, labels=None, **config)

Initializes HandwritingFeatures object from a numpy array.

Parameters

• values (numpy.ndarray) – data values

• labels (list, optional) – labels for the data values

• config (**kwargs) – common configuration

Returns

HandwritingFeatures object

Return type

HandwritingFeatures

classmethod from_pandas_dataframe(values, labels=None, **config)

Initializes HandwritingFeatures object from a pandas DataFrame.

Parameters

• values (pandas.DataFrame) – data values

• labels (list, optional) – labels for the data values

• config (**kwargs) – common configuration

Returns

HandwritingFeatures object

Return type

HandwritingFeatures

classmethod from_sample(sample, **config)

Initializes HandwritingFeatures object from a sample. :param sample: handwriting sample object :type sample: HandwritingSample :param config: common configuration :type config: **kwargs :return: HandwritingFeatures object :rtype: HandwritingFeatures

classmethod from_svc(path, labels=None, **config)

Initializes HandwritingFeatures object from an SVC file.

Parameters

• path (str) – path to an SVC file

• labels (list, optional) – labels for the data values

• config (**kwargs) – common configuration

Returns

HandwritingFeatures object

Return type

HandwritingFeatures

handwriting_features.features.interface module

class handwriting_features.features.interface.HandwritingFeatures(sample_wrapper, **config)

Bases: HandwritingFeaturesBase

Class implementing interface for computation of handwriting features

acceleration(axis='xy', in_air=False, statistics=())

Extracts the acceleration.

Parameters

• axis (str, optional) – axis to compute the acceleration from, defaults to “xy”

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

acceleration

Return type

numpy.ndarray or numpy.NaN

azimuth(in_air=False, statistics=())

Extracts the azimuth.

Parameters

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

azimuth

Return type

numpy.ndarray or numpy.NaN

jerk(axis='xy', in_air=False, statistics=())

Extracts the jerk.

Parameters

• axis (str, optional) – axis to compute the jerk from, defaults to “xy”

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

jerk

Return type

numpy.ndarray or numpy.NaN

number_of_changes_in_azimuth(fs, fc=None, n=None)

Extracts the number of changes in azimuth.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

number of changes in azimuth

Return type

int

number_of_changes_in_pressure(fs, fc=None, n=None)

Extracts the number of changes in pressure.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

number of changes in pressure

Return type

int

number_of_changes_in_tilt(fs, fc=None, n=None)

Extracts the number of changes in tilt.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

number of changes in tilt

Return type

int

number_of_changes_in_velocity_profile(fs, fc=None, n=None)

Extracts the number of changes in velocity profile.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

number of changes in velocity profile

Return type

int

number_of_changes_in_x_profile(fs, fc=None, n=None)

Extracts the number of changes in x profile.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

number of changes in x profile

Return type

int

number_of_changes_in_y_profile(fs, fc=None, n=None)

Extracts the number of changes in y profile.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

number of changes in y profile

Return type

int

number_of_inter_stroke_intersections()

Extracts the number of inter-stroke intersections.

Returns

number of inter-stroke intersections

Return type

int

number_of_interruptions()

Extracts the number of interruptions.

Returns

number of interruptions

Return type

float

number_of_interruptions_relative()

Extracts the number of interruptions relative to the duration.

Returns

number of interruptions relative to the duration

Return type

float

number_of_intra_stroke_intersections(statistics=())

Extracts the number of intra-stroke intersections.

Parameters

statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

number of intra-stroke intersections

Return type

numpy.ndarray or np.NaN

pressure(statistics=())

Extracts the pressure.

Parameters

statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

pressure

Return type

numpy.ndarray or numpy.NaN

ratio_of_stroke_durations(statistics=())

Extracts the ratio of stroke durations: on-surface / in-air.

Parameters

statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

ratio of stroke durations

Return type

numpy.ndarray or numpy.NaN

ratio_of_writing_durations()

Extracts the ratio of writing durations: on-surface / in-air.

Returns

ratio of writing durations

Return type

float

relative_number_of_changes_in_azimuth(fs, fc=None, n=None)

Extracts the relative number of changes in azimuth.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

relative number of changes in azimuth

Return type

int

relative_number_of_changes_in_pressure(fs, fc=None, n=None)

Extracts the relative number of changes in pressure.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

relative number of changes in pressure

Return type

int

relative_number_of_changes_in_tilt(fs, fc=None, n=None)

Extracts the relative number of changes in tilt.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

relative number of changes in tilt

Return type

int

relative_number_of_changes_in_velocity_profile(fs, fc=None, n=None)

Extracts the relative number of changes in velocity profile.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

relative number of changes in velocity profile

Return type

int

relative_number_of_changes_in_x_profile(fs, fc=None, n=None)

Extracts the relative number of changes in x profile.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

relative number of changes in x profile

Return type

int

relative_number_of_changes_in_y_profile(fs, fc=None, n=None)

Extracts the relative number of changes in y profile.

Parameters

• fs (float) – sampling frequency

• fc (float, optional) – cutoff frequency for the low-pass filter, defaults to None

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

Returns

relative number of changes in y profile

Return type

int

relative_number_of_inter_stroke_intersections()

Extracts the relative number of inter-stroke intersections.

Returns

relative number of inter-stroke intersections

Return type

float

relative_number_of_intra_stroke_intersections(statistics=())

Extracts the relative number of intra-stroke intersections.

Parameters

statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

relative number of intra-stroke intersections

Return type

numpy.ndarray or np.NaN

relative_total_number_of_intra_stroke_intersections()

Extracts the relative total number of intra-stroke intersections.

Returns

relative total number of intra-stroke intersections

Return type

float

stroke_duration(in_air=False, statistics=())

Extracts the stroke duration.

Parameters

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

stroke duration

Return type

numpy.ndarray or numpy.NaN

stroke_height(in_air=False, statistics=())

Extracts the stroke height.

Parameters

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

stroke height

Return type

numpy.ndarray or numpy.NaN

stroke_length(in_air=False, statistics=())

Extracts the stroke length.

Parameters

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

stroke length

Return type

numpy.ndarray or numpy.NaN

stroke_width(in_air=False, statistics=())

Extracts the stroke width.

Parameters

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

stroke width

Return type

numpy.ndarray or numpy.NaN

tilt(in_air=False, statistics=())

Extracts the tilt.

Parameters

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

tilt

Return type

numpy.ndarray or numpy.NaN

total_number_of_intra_stroke_intersections()

Extracts the total number of intra-stroke intersections.

Returns

total number of intra-stroke intersections

Return type

int

velocity(axis='xy', in_air=False, statistics=())

Extracts the velocity.

Parameters

• axis (str, optional) – axis to compute the velocity from, defaults to “xy”

• in_air (bool, optional) – in-air flag, defaults to False

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

velocity

Return type

numpy.ndarray or numpy.NaN

vertical_peaks_distance(fs, n=None, statistics=())

Extracts the vertical peaks distance.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical peaks distance

Return type

numpy.ndarray or np.NaN

vertical_peaks_duration(fs, n=None, statistics=())

Extracts the vertical peaks duration.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical peaks duration

Return type

numpy.ndarray or np.NaN

vertical_peaks_indices(fs, n=None, statistics=())

Extracts the vertical peaks indices.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical peaks indices

Return type

numpy.ndarray or np.NaN

vertical_peaks_values(fs, n=None, statistics=())

Extracts the vertical peaks values.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical peaks values

Return type

numpy.ndarray or np.NaN

vertical_peaks_velocity(fs, n=None, statistics=())

Extracts the vertical peaks velocity.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical peaks velocity

Return type

numpy.ndarray or np.NaN

vertical_valleys_distance(fs, n=None, statistics=())

Extracts the vertical valleys distance.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical valleys distance

Return type

numpy.ndarray or np.NaN

vertical_valleys_duration(fs, n=None, statistics=())

Extracts the vertical valleys duration.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical valleys duration

Return type

numpy.ndarray or np.NaN

vertical_valleys_indices(fs, n=None, statistics=())

Extracts the vertical valleys indices.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical valleys indices

Return type

numpy.ndarray or np.NaN

vertical_valleys_values(fs, n=None, statistics=())

Extracts the vertical valleys values.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical valleys values

Return type

numpy.ndarray or np.NaN

vertical_valleys_velocity(fs, n=None, statistics=())

Extracts the vertical valleys velocity.

Parameters

• fs (float) – sampling frequency

• n (int, optional) – number of samples of a Gaussian filter, defaults to None

• statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

vertical valleys velocity

Return type

numpy.ndarray or np.NaN

writing_duration(in_air=False)

Extracts the writing duration.

Parameters

in_air (bool, optional) – in-air flag, defaults to False

Returns

writing duration

Return type

float

writing_duration_overall()

Extracts the overall writing duration.

Returns

overall writing duration

Return type

float

writing_height(in_air=False)

Extracts the writing height.

Parameters

in_air (bool, optional) – in-air flag, defaults to False

Returns

writing height

Return type

float

writing_length(in_air=False)

Extracts the writing length.

Parameters

in_air (bool, optional) – in-air flag, defaults to False

Returns

writing length

Return type

float

writing_stops(statistics=())

Extracts the writing stops.

Parameters

statistics (Any[list, tuple], optional) – statistics to compute, defaults to ()

Returns

writing stops

Return type

numpy.ndarray or np.NaN

writing_tempo(in_air=False)

Extracts the writing tempo.

Parameters

in_air (bool, optional) – in-air flag, defaults to False

Returns

writing tempo

Return type

float

writing_width(in_air=False)

Extracts the writing width.

Parameters

in_air (bool, optional) – in-air flag, defaults to False

Returns

writing width

Return type

float

Module contents