Arithmetic and normalization

MoDaCor provides two arithmetic patterns.

Source-based Divide, Multiply, and Subtract obtain their second operand from IoSources. They suit scalar metadata or calibration arrays that should remain external to the processing workspace.

DivideDatabundles, MultiplyDatabundles, and SubtractDatabundles operate on two prepared BaseData entries in ProcessingData. Their with_processing_keys order is significant and they update the first bundle.

All use BaseData unit algebra, uncertainty propagation, broadcasting, and metadata checks. Do not strip values to arrays to reimplement these semantics in a pipeline step.

FindScaleFactor1D estimates a one-dimensional overlap scale. Its lognormal estimator uses a caller-selected named uncertainty component rather than silently inventing a combined uncertainty. Exact duplicate fit coordinates are inverse-variance averaged before interpolation.

SubtractInterpolated1D subtracts a one-dimensional background on a different coordinate grid. Its nearest mode is an explicit nearest-neighbour remap without extrapolation; linear propagates background uncertainty through the interpolation coefficients. Negate changes the sign of a selected BaseData entry while retaining its uncertainty magnitudes.

UnitsLabelUpdate changes a unit label only under its documented contract; it is not a substitute for numerical conversion.