.. _auto_dataset_process_curve_columns:

*DATASET PROCESS CURVE COLUMNS*
===============================

Processes one or more curve columns in a dataset through a configurable multi-step pipeline: monotonic enforcement, pre-digitization, smoothing (average, LOESS, or CFC filter), outlier removal, clipping, cross-row X-sync, and final digitization. Use this worker to standardize and clean raw curve data in bulk before downstream analysis or model training.

When to use
-----------

Tagged: ``cfc_filter``, ``clip``, ``curve``, ``dataset``, ``digitize``, ``loess``, ``monotonic``, ``outlier``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Dataset
     - dataset
     - dataset
     - —
     - ✓
     - Input dataset (tabular) containing one or more curve-typed columns to be processed; must be non-empty.
   * - Curve Columns
     - curve_columns
     - select
     - —
     - 
     - Names of the curve columns to process; if left empty the worker returns the dataset unchanged.
   * - Output Column Postfix
     - output_postfix
     - text
     - —
     - 
     - String appended to processed column names (e.g. '_processed' yields 'OUT_C1_processed'); leave blank to overwrite the original columns in-place.
   * - Monotonic Type
     - monotonic_type
     - select
     - x
     - 
     - Axis on which to enforce monotonicity before all other steps: 'x' (default), 'y', 'xy', or 'none' to skip; applied uniformly to all selected columns.
   * - Pre-Digitize Points
     - pre_digitize_points
     - text
     - 0
     - 
     - [per-column] Integer number of evenly-spaced points for an optional pre-digitization pass; use 0 (default) to skip; comma-separated values apply different counts per column (e.g. '500,0,1000').
   * - Smooth Method
     - smooth_method
     - select
     - none
     - 
     - Smoothing algorithm to apply after pre-digitization: 'none' (default, skip), 'smooth' (forward-backward average), 'regression_smooth' (LOESS), or 'filter' (CFC filter).
   * - Smooth Points
     - smooth_points
     - scalar
     - 4
     - 
     - Number of averaging points used by the 'Average Smooth' method (default 4); ignored when another smooth method is selected.
   * - Smooth Percentage
     - smooth_percentage
     - scalar
     - 30
     - 
     - Fraction of total points (as a percentage, default 30) used as the bandwidth for the LOESS 'Regression Smooth' method; ignored otherwise.
   * - Filter Frequency
     - filter_freq
     - scalar
     - 60
     - 
     - Channel Frequency Class (CFC) cut-off frequency in Hz (default 60) for the 'CFC Filter' smooth method; ignored otherwise.
   * - Remove Outlier Type
     - remove_outlier_type
     - select
     - none
     - 
     - Strategy for removing outlier points after smoothing; set to 'none' (default) to skip; consult platform documentation for available type identifiers.
   * - Clip X Min
     - clip_xmin
     - text
     - -1e20
     - 
     - [per-column] Minimum x-value for clipping; points with x below this threshold are removed (default -1e20, effectively no lower clip); comma-separated for per-column control.
   * - Clip X Max
     - clip_xmax
     - text
     - 1e20
     - 
     - [per-column] Maximum x-value for clipping; points with x above this threshold are removed (default 1e20, effectively no upper clip); comma-separated for per-column control.
   * - Clip Y Min
     - clip_ymin
     - text
     - -1e20
     - 
     - [per-column] Minimum y-value for clipping; points with y below this threshold are removed (default -1e20, effectively no lower clip); comma-separated for per-column control.
   * - Clip Y Max
     - clip_ymax
     - text
     - 1e20
     - 
     - [per-column] Maximum y-value for clipping; points with y above this threshold are removed (default 1e20, effectively no upper clip); comma-separated for per-column control.
   * - Sync X Start and End
     - sync_x_start_end
     - text
     - no
     - 
     - [per-column] Whether to synchronize the x-axis start and/or end across all rows of a column before final digitization; 'no' (default) skips this step; comma-separated for per-column control.
   * - Digitize Points
     - digitize_points
     - text
     - 100
     - 
     - [per-column] Number of evenly-spaced points for the final digitization pass (default 100); use 0 to skip; comma-separated values apply different counts per column.

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - Processed Dataset
     - dataset
     - dataset
     - Dataset identical in structure to the input but with the selected curve columns replaced (or augmented, if an output postfix was specified) by their fully processed counterparts.

Disciplines
-----------

- data.curve.transform
- data.dataset.transform
- data.signal_processing

Runnable example
----------------

A runnable example is registered for this worker. Open the example workflow on the d3VIEW canvas: `/api/workflow/example?id=dataset_process_curve_columns <https://www.d3view.com/api/workflow/example?id=dataset_process_curve_columns>`_

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>platform</code> schema. Worker id: <code>dataset_process_curve_columns</code>. Schema hash: <code>c287abbde5a6</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>