@sipemu/anofox-statistics 0.4.1 → 0.4.2

package/README.md

@@ -26,7 +26,7 @@ import init, {
 // Initialize the WASM module
 await init();
 
-// Two-sample t-test
+// Two-sample Welch's t-test
 const result = tTest(
   new Float64Array([1.2, 2.3, 3.1, 4.5, 5.2]),
   new Float64Array([2.1, 3.4, 4.2, 5.6, 6.1]),
@@ -45,7 +45,6 @@ console.log(result);
 ```javascript
 const { tTest, JsTTestKind, JsAlternative } = require('@sipemu/anofox-statistics');
 
-// Use directly (WASM auto-initializes)
 const result = tTest(
   new Float64Array([1.2, 2.3, 3.1]),
   new Float64Array([2.1, 3.4, 4.2]),
@@ -63,6 +62,8 @@ const result = tTest(
 | `tTest` | Student's, Welch's, or paired t-test |
 | `yuenTest` | Yuen's test for trimmed means (robust) |
 | `oneWayAnova` | One-way ANOVA (Fisher or Welch) |
+| `twoWayAnova` | Two-way ANOVA with interaction |
+| `repeatedMeasuresAnova` | One-way RM-ANOVA with sphericity test and GG/HF corrections |
 | `brownForsythe` | Brown-Forsythe test for homogeneity of variances |
 
 ### Nonparametric Tests
@@ -73,44 +74,62 @@ const result = tTest(
 | `wilcoxonSignedRank` | Wilcoxon signed-rank test |
 | `kruskalWallis` | Kruskal-Wallis H test |
 | `brunnerMunzel` | Brunner-Munzel test |
+| `rank` | Rank a sample, with average rank assigned to ties |
 
 ### Distributional Tests
 
 | Function | Description |
 |----------|-------------|
 | `shapiroWilk` | Shapiro-Wilk normality test |
-| `kolmogorovSmirnov` | Kolmogorov-Smirnov test |
-| `andersonDarling` | Anderson-Darling normality test |
-| `jarqueBera` | Jarque-Bera normality test |
+| `dagostinoKSquared` | D'Agostino's K² omnibus normality test |
 
 ### Categorical Tests
 
 | Function | Description |
 |----------|-------------|
-| `chiSquaredTest` | Chi-squared test for independence |
-| `fisherExact` | Fisher's exact test |
-| `binomialTest` | Exact binomial test |
+| `chiSquareTest` | Chi-square test of independence (with optional Yates correction) |
+| `chiSquareGoodnessOfFit` | Chi-square goodness-of-fit test |
+| `gTest` | G-test (likelihood-ratio chi-square) |
+| `fisherExact` | Fisher's exact test (2×2 tables) |
+| `binomTest` | Exact binomial test |
 | `mcnemarTest` | McNemar's test for paired proportions |
+| `mcnemarExact` | Exact McNemar test for small samples |
+| `propTestOne` | One-sample proportion test |
+| `propTestTwo` | Two-sample proportion test |
+| `cohenKappa` | Cohen's kappa for inter-rater agreement |
+| `cramersV` | Cramér's V effect size |
+| `phiCoefficient` | Phi coefficient (2×2 tables) |
+| `contingencyCoefficient` | Pearson's contingency coefficient |
 
 ### Correlation Tests
 
 | Function | Description |
 |----------|-------------|
-| `pearsonCorrelation` | Pearson correlation with significance |
+| `pearsonCorrelation` | Pearson product-moment correlation |
 | `spearmanCorrelation` | Spearman rank correlation |
 | `kendallCorrelation` | Kendall's tau correlation |
-| `icc` | Intraclass correlation coefficient |
+| `partialCorrelation` | Partial correlation (controlling for covariates) |
+| `semiPartialCorrelation` | Semi-partial (part) correlation |
+| `intraclassCorrelation` | Intraclass correlation coefficient |
+| `distanceCorrelation` | Distance correlation (Székely-Rizzo) |
+| `distanceCorrelationTest` | Distance correlation independence test |
 
 ### Equivalence Tests (TOST)
 
 | Function | Description |
 |----------|-------------|
-| `tostTwoSample` | TOST for two independent samples |
-| `tostPaired` | TOST for paired samples |
-| `tostCorrelation` | TOST for correlation equivalence |
-| `tostBootstrap` | Bootstrap TOST |
-
-### Modern/Kernel Tests
+| `tostTTestOneSample` | One-sample TOST |
+| `tostTTestTwoSample` | Two-sample TOST (Welch / Student) |
+| `tostTTestPaired` | Paired TOST |
+| `tostCorrelation` | Correlation equivalence (Pearson / Spearman) |
+| `tostPropOne` | One-proportion TOST |
+| `tostPropTwo` | Two-proportion TOST |
+| `tostWilcoxonPaired` | Paired Wilcoxon TOST |
+| `tostWilcoxonTwoSample` | Two-sample Wilcoxon (Mann-Whitney) TOST |
+| `tostYuen` | Trimmed-mean (Yuen) TOST |
+| `tostBootstrap` | Bootstrap TOST with percentile CI |
+
+### Modern / Kernel Tests
 
 | Function | Description |
 |----------|-------------|
@@ -129,13 +148,11 @@ const result = tTest(
 | `mspeAdjustedSpa` | MSPE-adjusted SPA test |
 | `modelConfidenceSet` | Model Confidence Set procedure |
 
-### Resampling Tests
+### Resampling
 
 | Function | Description |
 |----------|-------------|
 | `permutationTTest` | Permutation t-test |
-| `permutationCorrelation` | Permutation correlation test |
-| `bootstrapCI` | Bootstrap confidence intervals |
 
 ## Enums
 
@@ -149,13 +166,22 @@ enum JsAlternative { TwoSided, Less, Greater }
 // ANOVA variants
 enum JsAnovaKind { Fisher, Welch }
 
+// Kendall correlation variants
+enum JsKendallVariant { TauA, TauB, TauC }
+
+// Intraclass correlation forms
+enum JsICCType { ICC1, ICC2, ICC3, ICC1k, ICC2k, ICC3k }
+
+// Method for correlation TOST
+enum JsCorrelationTostMethod { Pearson, Spearman }
+
 // Kernel types for MMD
 enum JsKernel { Gaussian, Laplacian, Linear }
 
 // Loss functions for forecast comparison
 enum JsLossFunction { SquaredError, AbsoluteError }
 
-// Variance estimators
+// Variance estimators (Diebold-Mariano)
 enum JsVarEstimator { Acf, Bartlett }
 
 // MCS statistics
@@ -167,7 +193,7 @@ enum JsMCSStatistic { Max, Range }
 ### Normality Testing
 
 ```javascript
-import init, { shapiroWilk, andersonDarling } from '@sipemu/anofox-statistics';
+import init, { shapiroWilk, dagostinoKSquared } from '@sipemu/anofox-statistics';
 
 await init();
 
@@ -176,11 +202,11 @@ const data = new Float64Array([2.3, 3.1, 2.8, 3.5, 2.9, 3.2, 2.7]);
 const sw = shapiroWilk(data);
 console.log(`Shapiro-Wilk: W=${sw.statistic.toFixed(4)}, p=${sw.p_value.toFixed(4)}`);
 
-const ad = andersonDarling(data);
-console.log(`Anderson-Darling: A=${ad.statistic.toFixed(4)}, p=${ad.p_value.toFixed(4)}`);
+const dk = dagostinoKSquared(data);
+console.log(`D'Agostino K²: K²=${dk.statistic.toFixed(4)}, p=${dk.p_value.toFixed(4)}`);
 ```
 
-### ANOVA with Post-hoc
+### One-Way ANOVA
 
 ```javascript
 import init, { oneWayAnova, JsAnovaKind } from '@sipemu/anofox-statistics';
@@ -198,20 +224,61 @@ console.log(`F(${result.df_between}, ${result.df_within}) = ${result.statistic.t
 console.log(`p-value: ${result.p_value.toFixed(6)}`);
 ```
 
+### Two-Way ANOVA
+
+```javascript
+import init, { twoWayAnova } from '@sipemu/anofox-statistics';
+
+await init();
+
+// 2 (factor A) × 3 (factor B) design, 2 reps per cell
+const values   = new Float64Array([10, 12,  20, 22,  30, 32,  15, 17,  25, 27,  35, 37]);
+const factorA  = new Uint32Array([   0,  0,   0,  0,   0,  0,   1,  1,   1,  1,   1,  1]);
+const factorB  = new Uint32Array([   0,  0,   1,  1,   2,  2,   0,  0,   1,  1,   2,  2]);
+
+const result = twoWayAnova(values, factorA, factorB);
+console.log(`Factor A: F=${result.factor_a.f_statistic.toFixed(2)}, p=${result.factor_a.p_value.toFixed(4)}`);
+console.log(`Factor B: F=${result.factor_b.f_statistic.toFixed(2)}, p=${result.factor_b.p_value.toFixed(4)}`);
+console.log(`A×B:      F=${result.interaction.f_statistic.toFixed(2)}, p=${result.interaction.p_value.toFixed(4)}`);
+```
+
+### Repeated-Measures ANOVA
+
+```javascript
+import init, { repeatedMeasuresAnova } from '@sipemu/anofox-statistics';
+
+await init();
+
+// Rows = subjects, columns = conditions
+const data = [
+  new Float64Array([10, 12, 15]),
+  new Float64Array([11, 14, 17]),
+  new Float64Array([ 9, 13, 16]),
+  new Float64Array([12, 15, 18]),
+];
+
+const result = repeatedMeasuresAnova(data, /* computeSphericity */ true);
+console.log(`Within-subjects: F=${result.within_subjects.f_statistic.toFixed(2)}, p=${result.within_subjects.p_value.toFixed(4)}`);
+if (result.sphericity) {
+  console.log(`Mauchly W=${result.sphericity.w.toFixed(4)}, p=${result.sphericity.p_value.toFixed(4)}`);
+  console.log(`Greenhouse-Geisser p=${result.greenhouse_geisser.p_value.toFixed(4)}`);
+}
+```
+
 ### Equivalence Testing (TOST)
 
 ```javascript
-import init, { tostTwoSample, JsAlternative } from '@sipemu/anofox-statistics';
+import init, { tostTTestTwoSample, JsTTestKind } from '@sipemu/anofox-statistics';
 
 await init();
 
 const treatment = new Float64Array([10.2, 11.1, 9.8, 10.5, 10.9]);
-const control = new Float64Array([10.0, 10.8, 9.9, 10.3, 10.7]);
+const control   = new Float64Array([10.0, 10.8, 9.9, 10.3, 10.7]);
 
-// Test equivalence within bounds of -1 to +1
-const result = tostTwoSample(treatment, control, -1.0, 1.0, 0.95);
+// Test equivalence within ±1.0 of zero difference
+const result = tostTTestTwoSample(treatment, control, -1.0, 1.0, JsTTestKind.Welch, 0.05);
 console.log(`Equivalent: ${result.equivalent}`);
-console.log(`TOST p-value: ${result.p_value.toFixed(4)}`);
+console.log(`TOST p-value: ${result.tost_p_value.toFixed(4)}`);
 ```
 
 ### Forecast Model Comparison
@@ -233,7 +300,7 @@ const dm = dieboldMariano(
   errors1,
   errors2,
   JsLossFunction.SquaredError,
-  1,  // horizon
+  1,                       // horizon
   JsAlternative.TwoSided,
   JsVarEstimator.Bartlett
 );
@@ -268,11 +335,26 @@ console.log(result.p_value);
 
 ## Validation
 
-All statistical tests are validated against R's implementation using extensive test suites. Results match R's output within numerical precision (typically 10^-10).
+All statistical tests are validated against R's implementation using extensive test suites. Results match R's output within numerical precision (typically 10⁻¹⁰).
+
+## Changelog
+
+### 0.4.2
+
+- **Fixed:** p-values smaller than ~1.4×10⁻¹⁴ no longer underflow to `0.0`. The
+  underlying Rust library now uses each distribution's survival function (`sf`)
+  instead of `1 - cdf(x)` everywhere a tail probability is computed. Affects all
+  tests that use a normal, Student's t, chi-square, or F approximation
+  (Mann-Whitney, Wilcoxon, t/Welch/paired, Yuen, every ANOVA variant, Levene,
+  Brunner-Munzel, Kruskal-Wallis, chi-square, McNemar, Shapiro-Wilk, D'Agostino,
+  Pearson/Spearman/Kendall/partial correlation, ICC, proportion tests, Cohen's
+  kappa, Diebold-Mariano, Clark-West, and all TOST variants).
+- **Added:** `rank`, `twoWayAnova`, `repeatedMeasuresAnova` bindings — the JS
+  package now exposes the same statistical surface as the Rust crate.
 
 ## License
 
-MIT License - see [LICENSE](LICENSE) for details.
+MIT License — see [LICENSE](LICENSE) for details.
 
 ## Links
 

package/anofox_statistics_js.d.ts

@@ -442,6 +442,28 @@ export function propTestOne(successes: number, trials: number, p0: number, alter
  */
 export function propTestTwo(successes1: number, trials1: number, successes2: number, trials2: number, alternative: JsAlternative, correction?: boolean | null): any;
 
+/**
+ * Rank a sample, assigning the average rank to ties.
+ *
+ * @param data - Input sample as Float64Array
+ * @returns Float64Array of ranks (length matches input)
+ */
+export function rank(data: Float64Array): Float64Array;
+
+/**
+ * One-way repeated-measures ANOVA.
+ *
+ * @param data - Array of subjects, each subject is a Float64Array whose
+ *   entries are that subject's measurements across conditions. All inner
+ *   arrays must have the same length.
+ * @param computeSphericity - If true, also compute Mauchly's test and
+ *   Greenhouse-Geisser / Huynh-Feldt corrections (requires k >= 3 conditions).
+ * @returns Object with within_subjects, subjects, error, total rows; optional
+ *   sphericity, greenhouse_geisser, huynh_feldt; grand_mean, condition_means,
+ *   subject_means.
+ */
+export function repeatedMeasuresAnova(data: Array<any>, compute_sphericity?: boolean | null): any;
+
 /**
  * Semi-partial (part) correlation.
  *
@@ -628,6 +650,22 @@ export function tostWilcoxonTwoSample(x: Float64Array, y: Float64Array, lower_bo
  */
 export function tostYuen(x: Float64Array, y: Float64Array, lower_bound: number, upper_bound: number, trim?: number | null, alpha?: number | null): any;
 
+/**
+ * Two-way ANOVA with interaction.
+ *
+ * All three input arrays must have the same length. `factorA` and `factorB`
+ * hold the factor-level index for each observation (0-based; the number of
+ * levels is inferred from the maximum index plus one).
+ *
+ * @param values - Observations as Float64Array
+ * @param factorA - Factor A level index per observation as Uint32Array
+ * @param factorB - Factor B level index per observation as Uint32Array
+ * @returns Object with factor_a, factor_b, interaction, residual, total rows
+ *   (each: ss, df, ms, f_statistic, p_value), plus levels_a, levels_b, n,
+ *   grand_mean, cell_means, marginal_means_a, marginal_means_b.
+ */
+export function twoWayAnova(values: Float64Array, factor_a: Uint32Array, factor_b: Uint32Array): any;
+
 /**
  * Wilcoxon signed-rank test.
  *
@@ -671,6 +709,12 @@ export interface InitOutput {
     readonly phiCoefficient: (a: any) => [number, number, number];
     readonly propTestOne: (a: number, b: number, c: number, d: number) => [number, number, number];
     readonly propTestTwo: (a: number, b: number, c: number, d: number, e: number, f: number) => [number, number, number];
+    readonly brownForsythe: (a: any) => [number, number, number];
+    readonly oneWayAnova: (a: any, b: number) => [number, number, number];
+    readonly repeatedMeasuresAnova: (a: any, b: number) => [number, number, number];
+    readonly tTest: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number) => [number, number, number];
+    readonly twoWayAnova: (a: number, b: number, c: number, d: number, e: number, f: number) => [number, number, number];
+    readonly yuenTest: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number) => [number, number, number];
     readonly tostBootstrap: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: bigint) => [number, number, number];
     readonly tostCorrelation: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number) => [number, number, number];
     readonly tostPropOne: (a: number, b: number, c: number, d: number, e: number, f: number, g: number) => [number, number, number];
@@ -689,28 +733,25 @@ export interface InitOutput {
     readonly pearsonCorrelation: (a: number, b: number, c: number, d: number, e: number, f: number) => [number, number, number];
     readonly semiPartialCorrelation: (a: number, b: number, c: number, d: number, e: any) => [number, number, number];
     readonly spearmanCorrelation: (a: number, b: number, c: number, d: number, e: number, f: number) => [number, number, number];
+    readonly getVersion: () => [number, number];
+    readonly start: () => void;
     readonly clarkWest: (a: number, b: number, c: number, d: number, e: number) => [number, number, number];
     readonly dieboldMariano: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => [number, number, number];
     readonly modelConfidenceSet: (a: any, b: number, c: number, d: number, e: number, f: number, g: number, h: bigint) => [number, number, number];
     readonly mspeAdjustedSpa: (a: number, b: number, c: any, d: number, e: number, f: number, g: bigint) => [number, number, number];
     readonly spaTest: (a: number, b: number, c: any, d: number, e: number, f: number, g: bigint) => [number, number, number];
-    readonly brownForsythe: (a: any) => [number, number, number];
-    readonly oneWayAnova: (a: any, b: number) => [number, number, number];
-    readonly tTest: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number) => [number, number, number];
-    readonly yuenTest: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number) => [number, number, number];
     readonly brunnerMunzel: (a: number, b: number, c: number, d: number, e: number, f: number, g: number) => [number, number, number];
     readonly kruskalWallis: (a: any) => [number, number, number];
     readonly mannWhitneyU: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number) => [number, number, number];
+    readonly rank: (a: number, b: number) => [number, number, number, number];
     readonly wilcoxonSignedRank: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number) => [number, number, number];
+    readonly dagostinoKSquared: (a: number, b: number) => [number, number, number];
+    readonly permutationTTest: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: bigint) => [number, number, number];
+    readonly shapiroWilk: (a: number, b: number) => [number, number, number];
     readonly energyDistanceTest: (a: any, b: any, c: number, d: number, e: bigint) => [number, number, number];
     readonly energyDistanceTest1d: (a: number, b: number, c: number, d: number, e: number, f: number, g: bigint) => [number, number, number];
     readonly mmdTest: (a: any, b: any, c: number, d: number, e: number, f: number, g: number, h: bigint) => [number, number, number];
     readonly mmdTest1d: (a: number, b: number, c: number, d: number, e: number, f: number, g: bigint) => [number, number, number];
-    readonly dagostinoKSquared: (a: number, b: number) => [number, number, number];
-    readonly getVersion: () => [number, number];
-    readonly permutationTTest: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: bigint) => [number, number, number];
-    readonly shapiroWilk: (a: number, b: number) => [number, number, number];
-    readonly start: () => void;
     readonly __wbindgen_malloc: (a: number, b: number) => number;
     readonly __wbindgen_realloc: (a: number, b: number, c: number, d: number) => number;
     readonly __wbindgen_exn_store: (a: number) => void;

package/anofox_statistics_js.js

@@ -367,7 +367,7 @@ export function distanceCorrelationTest(x, y, n_permutations, seed) {
     const len0 = WASM_VECTOR_LEN;
     const ptr1 = passArrayF64ToWasm0(y, wasm.__wbindgen_malloc);
     const len1 = WASM_VECTOR_LEN;
-    const ret = wasm.distanceCorrelationTest(ptr0, len0, ptr1, len1, isLikeNone(n_permutations) ? 0x100000001 : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.distanceCorrelationTest(ptr0, len0, ptr1, len1, isLikeNone(n_permutations) ? Number.MAX_SAFE_INTEGER : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -389,7 +389,7 @@ export function distanceCorrelationTest(x, y, n_permutations, seed) {
  * @returns {any}
  */
 export function energyDistanceTest(x, y, n_permutations, seed) {
-    const ret = wasm.energyDistanceTest(x, y, isLikeNone(n_permutations) ? 0x100000001 : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.energyDistanceTest(x, y, isLikeNone(n_permutations) ? Number.MAX_SAFE_INTEGER : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -417,7 +417,7 @@ export function energyDistanceTest1d(x, y, n_permutations, seed) {
     const len0 = WASM_VECTOR_LEN;
     const ptr1 = passArrayF64ToWasm0(y, wasm.__wbindgen_malloc);
     const len1 = WASM_VECTOR_LEN;
-    const ret = wasm.energyDistanceTest1d(ptr0, len0, ptr1, len1, isLikeNone(n_permutations) ? 0x100000001 : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.energyDistanceTest1d(ptr0, len0, ptr1, len1, isLikeNone(n_permutations) ? Number.MAX_SAFE_INTEGER : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -617,7 +617,7 @@ export function mcnemarTest(table, correction) {
  * @returns {any}
  */
 export function mmdTest(x, y, kernel, bandwidth, n_permutations, seed) {
-    const ret = wasm.mmdTest(x, y, kernel, !isLikeNone(bandwidth), isLikeNone(bandwidth) ? 0 : bandwidth, isLikeNone(n_permutations) ? 0x100000001 : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.mmdTest(x, y, kernel, !isLikeNone(bandwidth), isLikeNone(bandwidth) ? 0 : bandwidth, isLikeNone(n_permutations) ? Number.MAX_SAFE_INTEGER : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -645,7 +645,7 @@ export function mmdTest1d(x, y, n_permutations, seed) {
     const len0 = WASM_VECTOR_LEN;
     const ptr1 = passArrayF64ToWasm0(y, wasm.__wbindgen_malloc);
     const len1 = WASM_VECTOR_LEN;
-    const ret = wasm.mmdTest1d(ptr0, len0, ptr1, len1, isLikeNone(n_permutations) ? 0x100000001 : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.mmdTest1d(ptr0, len0, ptr1, len1, isLikeNone(n_permutations) ? Number.MAX_SAFE_INTEGER : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -673,7 +673,7 @@ export function mmdTest1d(x, y, n_permutations, seed) {
  * @returns {any}
  */
 export function modelConfidenceSet(losses, alpha, statistic, n_bootstrap, block_length, seed) {
-    const ret = wasm.modelConfidenceSet(losses, !isLikeNone(alpha), isLikeNone(alpha) ? 0 : alpha, statistic, isLikeNone(n_bootstrap) ? 0x100000001 : (n_bootstrap) >>> 0, block_length, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.modelConfidenceSet(losses, !isLikeNone(alpha), isLikeNone(alpha) ? 0 : alpha, statistic, isLikeNone(n_bootstrap) ? Number.MAX_SAFE_INTEGER : (n_bootstrap) >>> 0, block_length, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -699,7 +699,7 @@ export function modelConfidenceSet(losses, alpha, statistic, n_bootstrap, block_
 export function mspeAdjustedSpa(benchmark_errors, model_errors, n_bootstrap, block_length, seed) {
     const ptr0 = passArrayF64ToWasm0(benchmark_errors, wasm.__wbindgen_malloc);
     const len0 = WASM_VECTOR_LEN;
-    const ret = wasm.mspeAdjustedSpa(ptr0, len0, model_errors, isLikeNone(n_bootstrap) ? 0x100000001 : (n_bootstrap) >>> 0, block_length, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.mspeAdjustedSpa(ptr0, len0, model_errors, isLikeNone(n_bootstrap) ? Number.MAX_SAFE_INTEGER : (n_bootstrap) >>> 0, block_length, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -793,7 +793,7 @@ export function permutationTTest(x, y, alternative, n_permutations, seed) {
     const len0 = WASM_VECTOR_LEN;
     const ptr1 = passArrayF64ToWasm0(y, wasm.__wbindgen_malloc);
     const len1 = WASM_VECTOR_LEN;
-    const ret = wasm.permutationTTest(ptr0, len0, ptr1, len1, alternative, isLikeNone(n_permutations) ? 0x100000001 : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.permutationTTest(ptr0, len0, ptr1, len1, alternative, isLikeNone(n_permutations) ? Number.MAX_SAFE_INTEGER : (n_permutations) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -864,6 +864,49 @@ export function propTestTwo(successes1, trials1, successes2, trials2, alternativ
     return takeFromExternrefTable0(ret[0]);
 }
 
+/**
+ * Rank a sample, assigning the average rank to ties.
+ *
+ * @param data - Input sample as Float64Array
+ * @returns Float64Array of ranks (length matches input)
+ * @param {Float64Array} data
+ * @returns {Float64Array}
+ */
+export function rank(data) {
+    const ptr0 = passArrayF64ToWasm0(data, wasm.__wbindgen_malloc);
+    const len0 = WASM_VECTOR_LEN;
+    const ret = wasm.rank(ptr0, len0);
+    if (ret[3]) {
+        throw takeFromExternrefTable0(ret[2]);
+    }
+    var v2 = getArrayF64FromWasm0(ret[0], ret[1]).slice();
+    wasm.__wbindgen_free(ret[0], ret[1] * 8, 8);
+    return v2;
+}
+
+/**
+ * One-way repeated-measures ANOVA.
+ *
+ * @param data - Array of subjects, each subject is a Float64Array whose
+ *   entries are that subject's measurements across conditions. All inner
+ *   arrays must have the same length.
+ * @param computeSphericity - If true, also compute Mauchly's test and
+ *   Greenhouse-Geisser / Huynh-Feldt corrections (requires k >= 3 conditions).
+ * @returns Object with within_subjects, subjects, error, total rows; optional
+ *   sphericity, greenhouse_geisser, huynh_feldt; grand_mean, condition_means,
+ *   subject_means.
+ * @param {Array<any>} data
+ * @param {boolean | null} [compute_sphericity]
+ * @returns {any}
+ */
+export function repeatedMeasuresAnova(data, compute_sphericity) {
+    const ret = wasm.repeatedMeasuresAnova(data, isLikeNone(compute_sphericity) ? 0xFFFFFF : compute_sphericity ? 1 : 0);
+    if (ret[2]) {
+        throw takeFromExternrefTable0(ret[1]);
+    }
+    return takeFromExternrefTable0(ret[0]);
+}
+
 /**
  * Semi-partial (part) correlation.
  *
@@ -927,7 +970,7 @@ export function shapiroWilk(x) {
 export function spaTest(benchmark_losses, model_losses, n_bootstrap, block_length, seed) {
     const ptr0 = passArrayF64ToWasm0(benchmark_losses, wasm.__wbindgen_malloc);
     const len0 = WASM_VECTOR_LEN;
-    const ret = wasm.spaTest(ptr0, len0, model_losses, isLikeNone(n_bootstrap) ? 0x100000001 : (n_bootstrap) >>> 0, block_length, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.spaTest(ptr0, len0, model_losses, isLikeNone(n_bootstrap) ? Number.MAX_SAFE_INTEGER : (n_bootstrap) >>> 0, block_length, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -1017,7 +1060,7 @@ export function tostBootstrap(x, y, lower_bound, upper_bound, alpha, n_bootstrap
     const len0 = WASM_VECTOR_LEN;
     const ptr1 = passArrayF64ToWasm0(y, wasm.__wbindgen_malloc);
     const len1 = WASM_VECTOR_LEN;
-    const ret = wasm.tostBootstrap(ptr0, len0, ptr1, len1, lower_bound, upper_bound, !isLikeNone(alpha), isLikeNone(alpha) ? 0 : alpha, isLikeNone(n_bootstrap) ? 0x100000001 : (n_bootstrap) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
+    const ret = wasm.tostBootstrap(ptr0, len0, ptr1, len1, lower_bound, upper_bound, !isLikeNone(alpha), isLikeNone(alpha) ? 0 : alpha, isLikeNone(n_bootstrap) ? Number.MAX_SAFE_INTEGER : (n_bootstrap) >>> 0, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed);
     if (ret[2]) {
         throw takeFromExternrefTable0(ret[1]);
     }
@@ -1280,6 +1323,38 @@ export function tostYuen(x, y, lower_bound, upper_bound, trim, alpha) {
     return takeFromExternrefTable0(ret[0]);
 }
 
+/**
+ * Two-way ANOVA with interaction.
+ *
+ * All three input arrays must have the same length. `factorA` and `factorB`
+ * hold the factor-level index for each observation (0-based; the number of
+ * levels is inferred from the maximum index plus one).
+ *
+ * @param values - Observations as Float64Array
+ * @param factorA - Factor A level index per observation as Uint32Array
+ * @param factorB - Factor B level index per observation as Uint32Array
+ * @returns Object with factor_a, factor_b, interaction, residual, total rows
+ *   (each: ss, df, ms, f_statistic, p_value), plus levels_a, levels_b, n,
+ *   grand_mean, cell_means, marginal_means_a, marginal_means_b.
+ * @param {Float64Array} values
+ * @param {Uint32Array} factor_a
+ * @param {Uint32Array} factor_b
+ * @returns {any}
+ */
+export function twoWayAnova(values, factor_a, factor_b) {
+    const ptr0 = passArrayF64ToWasm0(values, wasm.__wbindgen_malloc);
+    const len0 = WASM_VECTOR_LEN;
+    const ptr1 = passArray32ToWasm0(factor_a, wasm.__wbindgen_malloc);
+    const len1 = WASM_VECTOR_LEN;
+    const ptr2 = passArray32ToWasm0(factor_b, wasm.__wbindgen_malloc);
+    const len2 = WASM_VECTOR_LEN;
+    const ret = wasm.twoWayAnova(ptr0, len0, ptr1, len1, ptr2, len2);
+    if (ret[2]) {
+        throw takeFromExternrefTable0(ret[1]);
+    }
+    return takeFromExternrefTable0(ret[0]);
+}
+
 /**
  * Wilcoxon signed-rank test.
  *
@@ -1339,154 +1414,149 @@ export function yuenTest(x, y, trim, alternative, conf_level) {
     }
     return takeFromExternrefTable0(ret[0]);
 }
-
 function __wbg_get_imports() {
     const import0 = {
         __proto__: null,
-        __wbg_Error_8c4e43fe74559d73: function(arg0, arg1) {
+        __wbg_Error_ef53bc310eb298a0: function(arg0, arg1) {
             const ret = Error(getStringFromWasm0(arg0, arg1));
             return ret;
         },
-        __wbg_String_8f0eb39a4a4c2f66: function(arg0, arg1) {
+        __wbg_String_8564e559799eccda: function(arg0, arg1) {
             const ret = String(arg1);
             const ptr1 = passStringToWasm0(ret, wasm.__wbindgen_malloc, wasm.__wbindgen_realloc);
             const len1 = WASM_VECTOR_LEN;
             getDataViewMemory0().setInt32(arg0 + 4 * 1, len1, true);
             getDataViewMemory0().setInt32(arg0 + 4 * 0, ptr1, true);
         },
-        __wbg___wbindgen_is_function_0095a73b8b156f76: function(arg0) {
+        __wbg___wbindgen_is_function_754e9f305ff6029e: function(arg0) {
             const ret = typeof(arg0) === 'function';
             return ret;
         },
-        __wbg___wbindgen_is_object_5ae8e5880f2c1fbd: function(arg0) {
+        __wbg___wbindgen_is_object_56732c2bc353f41d: function(arg0) {
             const val = arg0;
             const ret = typeof(val) === 'object' && val !== null;
             return ret;
         },
-        __wbg___wbindgen_is_string_cd444516edc5b180: function(arg0) {
+        __wbg___wbindgen_is_string_c236cabd84a4d769: function(arg0) {
             const ret = typeof(arg0) === 'string';
             return ret;
         },
-        __wbg___wbindgen_is_undefined_9e4d92534c42d778: function(arg0) {
+        __wbg___wbindgen_is_undefined_67b456be8673d3d7: function(arg0) {
             const ret = arg0 === undefined;
             return ret;
         },
-        __wbg___wbindgen_number_get_8ff4255516ccad3e: function(arg0, arg1) {
+        __wbg___wbindgen_number_get_9bb1761122181af2: function(arg0, arg1) {
             const obj = arg1;
             const ret = typeof(obj) === 'number' ? obj : undefined;
             getDataViewMemory0().setFloat64(arg0 + 8 * 1, isLikeNone(ret) ? 0 : ret, true);
             getDataViewMemory0().setInt32(arg0 + 4 * 0, !isLikeNone(ret), true);
         },
-        __wbg___wbindgen_throw_be289d5034ed271b: function(arg0, arg1) {
+        __wbg___wbindgen_throw_1506f2235d1bdba0: function(arg0, arg1) {
             throw new Error(getStringFromWasm0(arg0, arg1));
         },
-        __wbg_call_389efe28435a9388: function() { return handleError(function (arg0, arg1) {
-            const ret = arg0.call(arg1);
-            return ret;
-        }, arguments); },
-        __wbg_call_4708e0c13bdc8e95: function() { return handleError(function (arg0, arg1, arg2) {
+        __wbg_call_9c758de292015997: function() { return handleError(function (arg0, arg1, arg2) {
             const ret = arg0.call(arg1, arg2);
             return ret;
         }, arguments); },
-        __wbg_crypto_86f2631e91b51511: function(arg0) {
+        __wbg_crypto_38df2bab126b63dc: function(arg0) {
             const ret = arg0.crypto;
             return ret;
         },
-        __wbg_from_bddd64e7d5ff6941: function(arg0) {
+        __wbg_from_d300fe49deab18f5: function(arg0) {
             const ret = Array.from(arg0);
             return ret;
         },
-        __wbg_getRandomValues_b3f15fcbfabb0f8b: function() { return handleError(function (arg0, arg1) {
+        __wbg_getRandomValues_c44a50d8cfdaebeb: function() { return handleError(function (arg0, arg1) {
             arg0.getRandomValues(arg1);
         }, arguments); },
-        __wbg_get_9b94d73e6221f75c: function(arg0, arg1) {
+        __wbg_get_2b48c7d0d006a781: function(arg0, arg1) {
             const ret = arg0[arg1 >>> 0];
             return ret;
         },
-        __wbg_length_32ed9a279acd054c: function(arg0) {
-            const ret = arg0.length;
+        __wbg_get_unchecked_33f6e5c9e2f2d6b2: function(arg0, arg1) {
+            const ret = arg0[arg1 >>> 0];
             return ret;
         },
-        __wbg_length_35a7bace40f36eac: function(arg0) {
+        __wbg_length_33096ac1966bb961: function(arg0) {
             const ret = arg0.length;
             return ret;
         },
-        __wbg_length_f7386240689107f3: function(arg0) {
+        __wbg_length_4a591ecaa01354d9: function(arg0) {
             const ret = arg0.length;
             return ret;
         },
-        __wbg_msCrypto_d562bbe83e0d4b91: function(arg0) {
-            const ret = arg0.msCrypto;
+        __wbg_length_66f1a4b2e9026940: function(arg0) {
+            const ret = arg0.length;
             return ret;
         },
-        __wbg_new_361308b2356cecd0: function() {
-            const ret = new Object();
+        __wbg_msCrypto_bd5a034af96bcba6: function(arg0) {
+            const ret = arg0.msCrypto;
             return ret;
         },
-        __wbg_new_3eb36ae241fe6f44: function() {
-            const ret = new Array();
+        __wbg_new_8d67bd740527e37f: function(arg0) {
+            const ret = new Float64Array(arg0);
             return ret;
         },
-        __wbg_new_d17d15ba1d2042a1: function(arg0) {
-            const ret = new Float64Array(arg0);
+        __wbg_new_ce1ab61c1c2b300d: function() {
+            const ret = new Object();
             return ret;
         },
-        __wbg_new_no_args_1c7c842f08d00ebb: function(arg0, arg1) {
-            const ret = new Function(getStringFromWasm0(arg0, arg1));
+        __wbg_new_d90091b82fdf5b91: function() {
+            const ret = new Array();
             return ret;
         },
-        __wbg_new_with_length_a2c39cbe88fd8ff1: function(arg0) {
+        __wbg_new_with_length_36a4998e27b014c5: function(arg0) {
             const ret = new Uint8Array(arg0 >>> 0);
             return ret;
         },
-        __wbg_node_e1f24f89a7336c2e: function(arg0) {
+        __wbg_node_84ea875411254db1: function(arg0) {
             const ret = arg0.node;
             return ret;
         },
-        __wbg_process_3975fd6c72f520aa: function(arg0) {
+        __wbg_process_44c7a14e11e9f69e: function(arg0) {
             const ret = arg0.process;
             return ret;
         },
-        __wbg_prototypesetcall_aefe6319f589ab4b: function(arg0, arg1, arg2) {
-            Float64Array.prototype.set.call(getArrayF64FromWasm0(arg0, arg1), arg2);
-        },
-        __wbg_prototypesetcall_bdcdcc5842e4d77d: function(arg0, arg1, arg2) {
+        __wbg_prototypesetcall_3249fc62a0fafa30: function(arg0, arg1, arg2) {
             Uint8Array.prototype.set.call(getArrayU8FromWasm0(arg0, arg1), arg2);
         },
-        __wbg_randomFillSync_f8c153b79f285817: function() { return handleError(function (arg0, arg1) {
+        __wbg_prototypesetcall_d1ae8885a2e9d458: function(arg0, arg1, arg2) {
+            Float64Array.prototype.set.call(getArrayF64FromWasm0(arg0, arg1), arg2);
+        },
+        __wbg_randomFillSync_6c25eac9869eb53c: function() { return handleError(function (arg0, arg1) {
             arg0.randomFillSync(arg1);
         }, arguments); },
-        __wbg_require_b74f47fc2d022fd6: function() { return handleError(function () {
+        __wbg_require_b4edbdcf3e2a1ef0: function() { return handleError(function () {
             const ret = module.require;
             return ret;
         }, arguments); },
-        __wbg_set_3f1d0b984ed272ed: function(arg0, arg1, arg2) {
+        __wbg_set_6be42768c690e380: function(arg0, arg1, arg2) {
             arg0[arg1] = arg2;
         },
-        __wbg_set_f43e577aea94465b: function(arg0, arg1, arg2) {
+        __wbg_set_dca99999bba88a9a: function(arg0, arg1, arg2) {
             arg0[arg1 >>> 0] = arg2;
         },
-        __wbg_static_accessor_GLOBAL_12837167ad935116: function() {
+        __wbg_static_accessor_GLOBAL_9d53f2689e622ca1: function() {
             const ret = typeof global === 'undefined' ? null : global;
             return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
         },
-        __wbg_static_accessor_GLOBAL_THIS_e628e89ab3b1c95f: function() {
+        __wbg_static_accessor_GLOBAL_THIS_a1a35cec07001a8a: function() {
             const ret = typeof globalThis === 'undefined' ? null : globalThis;
             return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
         },
-        __wbg_static_accessor_SELF_a621d3dfbb60d0ce: function() {
+        __wbg_static_accessor_SELF_4c59f6c7ea29a144: function() {
             const ret = typeof self === 'undefined' ? null : self;
             return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
         },
-        __wbg_static_accessor_WINDOW_f8727f0cf888e0bd: function() {
+        __wbg_static_accessor_WINDOW_e70ae9f2eb052253: function() {
             const ret = typeof window === 'undefined' ? null : window;
             return isLikeNone(ret) ? 0 : addToExternrefTable0(ret);
         },
-        __wbg_subarray_a96e1fef17ed23cb: function(arg0, arg1, arg2) {
+        __wbg_subarray_4aa221f6a4f5ab22: function(arg0, arg1, arg2) {
             const ret = arg0.subarray(arg1 >>> 0, arg2 >>> 0);
             return ret;
         },
-        __wbg_versions_4e31226f5e8dc909: function(arg0) {
+        __wbg_versions_276b2795b1c6a219: function(arg0) {
             const ret = arg0.versions;
             return ret;
         },
@@ -1559,8 +1629,15 @@ function getFloat64ArrayMemory0() {
 }
 
 function getStringFromWasm0(ptr, len) {
-    ptr = ptr >>> 0;
-    return decodeText(ptr, len);
+    return decodeText(ptr >>> 0, len);
+}
+
+let cachedUint32ArrayMemory0 = null;
+function getUint32ArrayMemory0() {
+    if (cachedUint32ArrayMemory0 === null || cachedUint32ArrayMemory0.byteLength === 0) {
+        cachedUint32ArrayMemory0 = new Uint32Array(wasm.memory.buffer);
+    }
+    return cachedUint32ArrayMemory0;
 }
 
 let cachedUint8ArrayMemory0 = null;
@@ -1584,6 +1661,13 @@ function isLikeNone(x) {
     return x === undefined || x === null;
 }
 
+function passArray32ToWasm0(arg, malloc) {
+    const ptr = malloc(arg.length * 4, 4) >>> 0;
+    getUint32ArrayMemory0().set(arg, ptr / 4);
+    WASM_VECTOR_LEN = arg.length;
+    return ptr;
+}
+
 function passArrayF64ToWasm0(arg, malloc) {
     const ptr = malloc(arg.length * 8, 8) >>> 0;
     getFloat64ArrayMemory0().set(arg, ptr / 8);
@@ -1663,12 +1747,14 @@ if (!('encodeInto' in cachedTextEncoder)) {
 
 let WASM_VECTOR_LEN = 0;
 
-let wasmModule, wasm;
+let wasmModule, wasmInstance, wasm;
 function __wbg_finalize_init(instance, module) {
+    wasmInstance = instance;
     wasm = instance.exports;
     wasmModule = module;
     cachedDataViewMemory0 = null;
     cachedFloat64ArrayMemory0 = null;
+    cachedUint32ArrayMemory0 = null;
     cachedUint8ArrayMemory0 = null;
     wasm.__wbindgen_start();
     return wasm;

package/package.json

@@ -5,7 +5,7 @@
     "Simon Müller <code@santic.info>"
   ],
   "description": "WebAssembly bindings for anofox-statistics hypothesis testing library",
-  "version": "0.4.1",
+  "version": "0.4.2",
   "license": "MIT",
   "repository": {
     "type": "git",