Python API

Python API

Public root imports after import loqculate:

from loqculate import (
    CalibrationData,
    EmpiricalCV,
    MODEL_REGISTRY,
    PiecewiseCF,
    PiecewiseWLS,
    read_calibration_data,
    __version__,
)
from loqculate.compat import OriginalCV, OriginalWLS

MODEL_REGISTRY keys: piecewise_cf, piecewise_wls, cv_empirical, original_wls, original_cv. Compat classes are registered at package import, not inside loqculate.models.

Arrays, not an iterator

import numpy as np
from loqculate import PiecewiseCF, read_calibration_data

data = read_calibration_data("data.tsv", "conc_map.csv")
# data.peptide, data.concentration, data.area
# data.metadata["format"], ["n_peptides"], ["n_measurements"]

for peptide in np.unique(data.peptide):
    mask = data.peptide == peptide
    m = PiecewiseCF().fit(data.concentration[mask], data.area[mask])
    print(peptide, m.lod(), m.loq())

There is no iter_peptides(). fmt= defaults to "auto". The map path is required even for generic CSV.

apply_multiplier(data, path) is in loqculate.io. Columns peptide, multiplier. Inner join.

Per-class methods

The math is on the per-model pages. PiecewiseCF documents loq(method=...), loq_delta(), covariance(), and --fast. PiecewiseWLS loq() has no method. OriginalWLS and OriginalCV compute LOQ during fit() and ignore later loq(cv_thresh) arguments.