ForgeSPC

Statistical Process Control engine for manufacturing. Pure computation library -- no web framework, no database, no I/O.

Install

pip install forgespc

Quick Start

from forgespc.charts import individuals_moving_range_chart, xbar_r_chart
from forgespc.capability import calculate_capability
from forgespc.rules import check_nelson_rules

result = individuals_moving_range_chart([25.01, 24.99, 25.03, 25.00, 24.98])
cap = calculate_capability([25.01, 24.99, 25.03, 25.00], usl=25.05, lsl=24.95)
violations = check_nelson_rules(result)

Advanced (requires numpy)

from forgespc.gage import gage_rr_crossed
from forgespc.bayesian import bayesian_capability

grr = gage_rr_crossed(measurements, parts, operators)
bcap = bayesian_capability(data, usl=53.0, lsl=47.0)

Modules

Module	Contents
constants	Control chart constants (A2, D3, D4, d2, c4) for subgroups 2-10
models	ControlLimits, ControlChartResult, ProcessCapability, StatisticalSummary
charts	I-MR, X-bar/R, p, c, u, np charts
capability	Cp, Cpk, Pp, Ppk, sigma level, DPMO, yield
rules	Nelson rules (2-8), Western Electric rules
advanced	CUSUM, EWMA, X-bar/S charts
gage	Gage R&R (crossed/nested), attribute agreement, Hotelling T-squared
bayesian	Bayesian capability, changepoint detection, Bayesian control chart
conformal	Conformal prediction control, entropy SPC
calibration	Self-calibration against golden reference files

Dependencies

• Core: Python stdlib only (math, statistics, dataclasses)
• Advanced (gage, bayesian, conformal): numpy, scipy

License

MIT