Skip to content

init

features ¤

Features

Feature extraction and summarization utilities for shaped timeseries.

  • NumericStatistics: Compute descriptive statistics for numeric columns.
  • column_mean: Mean of a column.
  • column_median: Median of a column.
  • column_std: Standard deviation of a column.
  • column_variance: Variance of a column.
  • column_min: Minimum value.
  • column_max: Maximum value.
  • column_sum: Sum of values.
  • column_kurtosis: Kurtosis of values.
  • column_skewness: Skewness of values.
  • column_quantile: Quantile of a column.
  • column_iqr: Interquartile range.
  • column_range: Range (max - min).
  • column_mad: Mean absolute deviation.
  • coefficient_of_variation: Standard deviation divided by mean (guarded).
  • standard_error_mean: Standard error of the mean.
  • describe: Pandas describe wrapper.
  • summary_as_dict: Comprehensive numeric summary as dict.

  • summary_as_dataframe: Comprehensive numeric summary as DataFrame.

  • StringStatistics: String-based statistics for categorical/text columns.

  • count_unique: Number of unique strings.
  • most_frequent: Most frequent string.
  • count_most_frequent: Count of the most frequent string.
  • count_null: Number of nulls.
  • average_string_length: Average length of non-null strings.
  • longest_string: Longest string.
  • shortest_string: Shortest string.
  • string_length_summary: Summary of lengths.
  • most_common_n_strings: Top-N most frequent strings.
  • contains_substring_count: Count of strings containing a substring.
  • starts_with_count: Count of strings starting with a prefix.
  • ends_with_count: Count of strings ending with a suffix.
  • uppercase_percentage: Percentage of uppercase strings.
  • lowercase_percentage: Percentage of lowercase strings.
  • contains_digit_count: Count of strings containing digits.
  • summary_as_dict: Comprehensive string summary as dict.

  • summary_as_dataframe: Comprehensive string summary as DataFrame.

  • BooleanStatistics: Boolean column statistics.

  • count_true: Count of True values.
  • count_false: Count of False values.
  • count_null: Count of nulls.
  • count_not_null: Count of non-nulls.
  • true_percentage: Percentage True.
  • false_percentage: Percentage False.
  • mode: Most common boolean value.
  • is_balanced: Whether distribution is 50/50.
  • summary_as_dict: Summary as dict.

  • summary_as_dataframe: Summary as DataFrame.

  • TimestampStatistics: Timestamp distributions and ranges.

  • count_null: Count of null timestamps.
  • count_not_null: Count of non-null timestamps.
  • earliest_timestamp: Earliest timestamp.
  • latest_timestamp: Latest timestamp.
  • timestamp_range: Time range (latest - earliest).
  • most_frequent_timestamp: Most frequent timestamp.
  • count_most_frequent_timestamp: Count of the modal timestamp.
  • year_distribution: Distribution by year.
  • month_distribution: Distribution by month.
  • weekday_distribution: Distribution by weekday.
  • hour_distribution: Distribution by hour.
  • most_frequent_day: Most frequent weekday.
  • most_frequent_hour: Most frequent hour.
  • average_time_gap: Average gap between consecutive timestamps.
  • median_timestamp: Median timestamp.
  • standard_deviation_timestamps: Standard deviation of consecutive differences.
  • timestamp_quartiles: 25th/50th/75th percentiles.

  • days_with_most_activity: Top-N active days.

  • TimeGroupedStatistics: Time-windowed aggregations for numeric series.

  • calculate_statistic: Single aggregation per window (mean/sum/min/max/diff/range).
  • calculate_statistics: Multiple aggregations merged.

  • calculate_custom_func: Apply a custom aggregation per window.

  • CycleExtractor: Build cycles from state/step/value changes.

  • process_persistent_cycle: True stretches define cycles.
  • process_trigger_cycle: True-to-False transition defines a cycle end.
  • process_separate_start_end_cycle: Separate starts and ends signals.
  • process_step_sequence: Start/end steps in integer values.
  • process_state_change_cycle: Sequential rows define boundaries.

  • process_value_change_cycle: Any value change defines a boundary.

  • CycleDataProcessor: Split/merge/group by cycle windows.

  • split_by_cycle: Split values by cycle ranges.
  • merge_dataframes_by_cycle: Annotate values with cycle UUIDs.
  • group_by_cycle_uuid: Group values by cycle key.

  • split_dataframes_by_group: Further split by column groupings.

  • CrossSignalAnalytics: Cross-signal analytics for multi-signal timeseries.

  • granger_causality: Test if one signal Granger-causes another.
  • transfer_entropy: Estimate information transfer between signals.
  • pairwise_transfer_entropy: Transfer entropy for all directed pairs.
  • synchronization_index: Phase or amplitude synchronization.
  • pairwise_synchronization: Synchronization for all pairs.
  • lead_lag: Detect lead-lag relationships via cross-correlation.

  • lead_lag_matrix: Lead-lag for all pairs.

  • PatternRecognition: Pattern discovery for univariate timeseries.

  • discover_motifs: Find top-k recurring subsequence patterns.
  • discover_discords: Find top-k anomalous subsequences.
  • similarity_search: Find subsequences most similar to a query (DTW).
  • template_match: Find all occurrences of a reference template.

  • compute_distance_profile: Distance from query to every subsequence.

  • SegmentExtractor: Extract time ranges from categorical signals (order/part number).

  • extract_time_ranges: Detect value transitions and extract segment boundaries.

  • SegmentProcessor: Apply time ranges to process data and compute metric profiles.

  • apply_ranges: Filter data by time ranges, annotate with segment info.

  • compute_metric_profiles: Compute statistical metrics per UUID per segment.

  • ProfileComparison: Distance, clustering, similarity, anomaly on metric profiles.

  • compute_distance_matrix: Pairwise distance matrix between groups.
  • cluster: Hierarchical clustering by metric similarity.
  • find_similar: Top-K most similar items to a target.
  • detect_anomalous: Flag items with unusual metric profiles.
  • detect_changes: Track metric shifts across consecutive segments.

  • find_similar_pairs: Find similar (UUID, segment) pairs across all data.

  • FeatureMatrixExporter: Convert long-format timeseries to wide ML-ready feature matrices.

  • to_feature_matrix: Pivot by uuid × value_col × agg into {uuid}{col} columns. Supports optional group_col (cycle, batch, segment) as row index.