API Reference¶
The reusable code lives in the utils package. Every public function has a NumPy-style
docstring, and the pages in this section are generated directly from those docstrings
with mkdocstrings, so they never drift from the code.
| Module | What it covers |
|---|---|
stats_tests_utils |
t-tests, Mann-Whitney, Wilcoxon, normality and variance tests, effect sizes, BH correction |
distribution_utils |
MLE fitting, PDF/CDF evaluation, KS and Anderson goodness-of-fit |
pdf_ecdf_utils |
ECDF computation and PDF/ECDF overlays |
sim_utils |
bootstrap, stratified/weighted sampling, Mahalanobis outliers |
inference_utils |
SEM, confidence intervals, margin of error, sample size |
power_utils |
power of z- and t-tests, Cohen's d |
optimization_utils |
cost functions, constraints, minimize wrappers, loss surfaces |
linear_algebra_utils |
eigen, SVD, least squares, matrix diagnostics |
interpolation_utils |
1-D/2-D interpolation, RBF, curve_fit models |
viz_utils |
every plotting helper |
scipymasterpro.cli |
the command line interface |
import numpy as np
from scipy import stats
from utils.distribution_utils import fit_distribution, perform_ks_test
from utils.stats_tests_utils import cohens_d_independent, run_two_sample_ttest
a = np.random.default_rng(0).normal(5, 2, 200)
b = np.random.default_rng(1).normal(6, 2, 200)
print(run_two_sample_ttest(a, b)) # {'t_stat': ..., 'p_value': ...}
print(cohens_d_independent(a, b)) # ~ -0.5
params = fit_distribution(a, stats.norm) # (loc, scale)
print(perform_ks_test(a, stats.norm, params)) # {'KS_stat': ..., 'p_value': ...}