Struct RuleSet 

pub struct RuleSet { /* private fields */ }

A chosen subset of the run tests.

Each test reports exactly one ViolationType, so a rule is named by what it would report rather than by a parallel enum that would have to be kept in step with this one.

Three sets have names – shewhart (the 3-sigma test alone, the default), western_electric and nelson – and RuleSet::named looks them up by the names a wire request uses; any other combination is built from the tests a process actually warrants. A rule that a process is known to trigger for a benign reason – a deliberately drifting tool, a bimodal fixture – otherwise has to be filtered out downstream, after it has already been counted as an out-of-control signal.

Examples

use u_analytics::spc::{RuleSet, RunRule, ChartPoint, ControlLimits, ViolationType};

// Only the 3-sigma test, and nothing else.
let rules = RuleSet::from_iter([ViolationType::BeyondLimits]);

let limits = ControlLimits { ucl: 28.0, cl: 25.0, lcl: 22.0 };
let points: Vec<ChartPoint> = (0..12)
    .map(|i| ChartPoint { value: 26.0, index: i, violations: vec![] })
    .collect();

// Twelve points above the centre line would trip Nelson rule 2, but that
// test is not in this set.
assert!(rules.check(&points, &limits).is_empty());
assert!(RuleSet::nelson().check(&points, &limits).len() > 0);

Implementations

impl RuleSet

pub const NAMES: [&'static str; 3]

The names RuleSet::named takes, in the order a refusal lists them.

pub fn shewhart() -> Self

The 3-sigma test alone – Shewhart’s criterion, and the default.

In routine monitoring each added test raises the false alarm rate of an in-control process (ISO 7870-2:2023, 8.2.7), so the supplementary tests are chosen, not assumed.

pub fn nelson() -> Self

Nelson’s eight tests (1984).

pub fn western_electric() -> Self

The four Western Electric tests (1956): Nelson rules 1, 5 and 6, and eight points in a row on one side of the centre line.

pub fn named(name: &str) -> Option<Self>

The set a name in RuleSet::NAMES stands for.

pub fn none() -> Self

The empty set: no run test is applied.

Useful as a base to add to, and as the honest way to say “limits only”.

pub fn contains(&self, rule: ViolationType) -> bool

Whether this set applies the test that reports rule.

pub fn with(self, rule: ViolationType) -> Self

This set with rule added.

pub fn without(self, rule: ViolationType) -> Self

This set with rule removed.

pub fn iter(&self) -> impl Iterator<Item = ViolationType> + '_

The tests in this set, in the order of ViolationType::ALL.

pub fn len(&self) -> usize

How many tests this set applies.

pub fn is_empty(&self) -> bool

Whether no test is applied.

Trait Implementations

impl Clone for RuleSet

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for RuleSet

impl Debug for RuleSet

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for RuleSet

fn default() -> Self

RuleSet::shewhart – the 3-sigma test alone.

impl Eq for RuleSet

impl FromIterator<ViolationType> for RuleSet

fn from_iter<I: IntoIterator<Item = ViolationType>>(iter: I) -> Self

Creates a value from an iterator.

impl PartialEq for RuleSet

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl RunRule for RuleSet

fn check( &self, points: &[ChartPoint], limits: &ControlLimits, ) -> Vec<(usize, ViolationType)>

Check points against this rule set and return violations per point index.

impl StructuralPartialEq for RuleSet