Module correction

source code

Correction is an abstract interface for data corrections.

A correction applies to a dataset. Different kinds of datasets support different kinds of corrections. For example, polarized data supports a polarized data correction as well as the usual data corrections on the polarization cross sections.

A correction has a string name that can be stored in a log file.

See properties.py for a discussion of correction parameters.

We may want to support an undo/redo stack, either by providing an undo method for reversing the effect of the correction, or by taking a snapshot of the data before the operation if the operation is not reversable, or by rerunning all corrections up to the current point. We need to be careful how this mechanism is implemented since the correction parameters stored in the object may change over time, and since reversing the effects of the correction may require some information about the contents of the data.

Consider having C1(data) return a copy of the data with the correction applied and data.apply(C1) modify the data set.

Individual corrections must inherit from Correction to support the pipeline syntax:

data |= normalize() | smooth() | attenuate() | ... | plot()