Python Functions¶
This page documents the current registered Python functions.
Summation¶
summation¶
Sums one or more series or expressions. Row-level nulls are treated as zero.
Parameters:
all_required: defaults toTrue. WhenTrueand nocutoffis provided, any entirely-null input series makes the full result null.cutoff: optional fraction from0to1. When provided, the result is returned if at least one input has a non-null fraction greater than or equal tocutoff; otherwise the full result is null.cutofftakes precedence overall_required.
out = apply("summation", a, b, all_required=False)
Multiplication¶
multiply_by_group¶
Multiplies indexed factor_N series or expressions. Set the scalar
invert_N=True to divide by that factor instead. Indices start at 1, are
contiguous, and invert_N defaults to False.
out = apply(
"multiply_by_group",
factor_1=pl.col("a"),
factor_2=pl.col("b"),
factor_3=pl.col("c"),
invert_3=True,
)
Correction And Standardization¶
standardize¶
Computes:
measured * 100 / standard
standardize_creatinine¶
Creatinine-specific wrapper around standardize, using crt as the standard.
normalize_specific_gravity¶
Computes:
measured * (sg_ref - 1) / sg_measured
total_lipid_concentration¶
Computes:
chol * 2.27 + trigl + 62.3
standardize_lipid¶
Lipid-specific wrapper around standardize.
coalesce_by_priority¶
Returns the first non-null value according to a named priority sequence.
out = apply(
"coalesce_by_priority",
primary=df["lab_a"],
secondary=df["lab_b"],
fallback=df["lab_c"],
priority=("primary", "secondary", "fallback"),
)
consolidate_lipid_value¶
Uses the lipid priority order:
lipid_enz_harmlipid_enz_implipid_imp
Imputation¶
lab_sensitivity_dichotomization¶
Produces a boolean expression indicating whether a measurement is below the limit of quantification or, when provided, below the limit of detection.
medium_bound_imputation_scalar_input¶
Uses scalar loq and optional scalar lod thresholds.
medium_bound_imputation¶
Uses series-valued loq and optional series-valued lod thresholds.
bin_decoding¶
Copies from the paired copy_from_N column/expression when values equals
the scalar filter_value_N; otherwise returns values.
random_single_imputation_scalar_input¶
Runs random single imputation with scalar lod and loq thresholds.
random_single_imputation¶
Runs random single imputation with series-valued lod and loq thresholds.
Optional imputation controls include:
min_unique_valuesmin_observed_percentageseed