pelagos_py.steps.input_output.correct_values#

class pelagos_py.steps.input_output.correct_values.CorrectValues(name, parameters=None, diagnostics=False, context=None)[source]#

Bases: pelagos_py.steps.base_step.BaseStep

Apply an affine correction corrected = slope * value + intercept to a variable.

The correction is conditional when expected_range is given: the median of the valid (non-NaN) data is compared against [min, max], and the correction is applied only when the median falls outside that range (i.e. the data still looks uncorrected). When expected_range is omitted the correction is always applied.

This makes a config robust to upstream fixes: e.g. a CNDC unit conversion (slope: 10) targeting expected_range: [20, 45] will scale S/m data into mS/cm, but quietly skip files that already arrive in mS/cm.

Parameters:

  • target_variable (str) – Name of the variable to correct (e.g. "CNDC").

  • slope (float, optional) – Multiplicative factor (default 1.0). For a x10 unit conversion, set 10.

  • intercept (float, optional) – Additive offset applied after scaling (default 0.0). Use for alignment.

  • expected_range (list, optional) – [min, max] for the corrected variable. The correction is applied only when the data’s median falls outside this range. If omitted, the correction is always applied.

  • corrected_units (str, optional) – Units string written to the variable’s attributes after a correction is applied (e.g. "mS/cm"). Left untouched if omitted or if no correction runs.

Examples

Example usage in a pipeline configuration:

steps:
  - name: Correct Values
    parameters:
      target_variable: CNDC
      slope: 10.0
      intercept: 0.0
      expected_range: [20, 45]
      corrected_units: mS/cm
    diagnostics: false
On this page