clophfit.fitting.pipeline#

Pipeline orchestrators for fitting multistage workflows (e.g., FGLS).

Functions#

fgls_plate_fit(datasets, sigma_floor, *[, ...])

Two-stage Feasible Generalized Least Squares (FGLS) plate fit.

fit_plate(datasets[, method])

Run a single-pass fit on an entire plate of datasets.

Module Contents#

clophfit.fitting.pipeline.fgls_plate_fit(datasets, sigma_floor, *, first_pass_method='huber', second_pass_method='lm')#

Two-stage Feasible Generalized Least Squares (FGLS) plate fit.

  1. First-pass fit (typically robust like ‘huber’) on each well.

  2. Extract residuals globally and calibrate the comprehensive error model, anchoring the constant noise term to the provided sigma_floor.

  3. Second-pass fit using the exact pooled weights derived from the model.

Parameters:

  • datasets (dict[str, Dataset]) – The dataset dictionary keyed by well name.

  • sigma_floor (dict[str, float]) – Known read-noise floor per label (e.g. from buffer wells).

  • first_pass_method (str) – Method for the first-pass fit (default ‘huber’).

  • second_pass_method (str) – Method for the second-pass, calibrated fit (default ‘lm’).

Returns:

Final fit results and the calibrated error model parameters (sigma_read, gain, alpha) for each label.

Return type:

tuple[dict[str, FitResult[Any]], dict[str, tuple[float, float, float]]]

clophfit.fitting.pipeline.fit_plate(datasets, method='lm', **kwargs)#

Run a single-pass fit on an entire plate of datasets.

Parameters:

  • datasets (dict[str, Dataset]) – A mapping of well keys (e.g. ‘A01’) to Dataset objects.

  • method (str) – The fitting method to use: ‘lm’ (default), ‘odr’, or ‘mcmc’.

  • **kwargs (Any) – Additional keyword arguments passed to the specific fitting function.

Returns:

A dictionary mapping well keys to their corresponding FitResult.

Return type:

dict[str, FitResult[Any]]