@milaboratories/miplots4 1.0.176 → 1.0.178

package/dist/discrete/layers/stackedBar.js

@@ -1,69 +1,64 @@
-import { NO_GROUPED as k } from "../../DataFrame.js";
-import { getFacetStringKey as z } from "../utils/getFacetStringKey.js";
-import { exhaustive as E } from "../../utils/index.js";
-import M from "../../node_modules/.pnpm/d3-array@3.2.4/node_modules/d3-array/src/mean.js";
-import { quantileSorted as Y } from "../../node_modules/.pnpm/d3-array@3.2.4/node_modules/d3-array/src/quantile.js";
-import g from "../../node_modules/.pnpm/d3-array@3.2.4/node_modules/d3-array/src/min.js";
-import v from "../../node_modules/.pnpm/d3-array@3.2.4/node_modules/d3-array/src/max.js";
-import w from "../../node_modules/.pnpm/d3-array@3.2.4/node_modules/d3-array/src/sum.js";
-function B(t, a, n) {
-  if (t.length === 0)
-    throw Error("Unable to create bar on empty data");
-  switch (n) {
-    case "sum":
-      return w(t, a);
-    case "max":
-      return v(t, a) ?? a(0);
-    case "min":
-      return g(t, a) ?? a(0);
-    case "median": {
-      const u = [...t].map(a).sort((m, i) => m - i);
-      return Y(u, 0.5) ?? a(0);
-    }
-    case "mean":
-      return M(t, a) ?? a(0);
-    default:
-      E(n, `Unknown bar height function ${n}`);
-  }
+import { exhaustive as e } from "../../utils/index.js";
+import { NO_GROUPED as t } from "../../DataFrame.js";
+import { getFacetStringKey as n } from "../utils/getFacetStringKey.js";
+import { max as r, mean as i, min as a, quantileSorted as o, sum as s } from "d3-array";
+function c(t, n, c) {
+	if (t.length === 0) throw Error("Unable to create bar on empty data");
+	switch (c) {
+		case "sum": return s(t, n);
+		case "max": return r(t, n) ?? n(0);
+		case "min": return a(t, n) ?? n(0);
+		case "median": return o([...t].map(n).sort((e, t) => e - t), .5) ?? n(0);
+		case "mean": return i(t, n) ?? n(0);
+		default: e(c, `Unknown bar height function ${c}`);
+	}
 }
-function R(t, a, n, u) {
-  let m = -1 / 0, i = 1 / 0;
-  return {
-    type: "stackedBar",
-    geoms: a.facet.reduce((s, f) => {
-      const l = z(f);
-      return s[l] = [], a.primary.forEach((h) => {
-        let c = 0;
-        const e = {};
-        a.secondary.forEach((r) => {
-          const o = r === k ? "null" : r, b = n.getRowsByGrouping([...f, h, r]);
-          if (!b.length)
-            return;
-          const d = B(b, (x) => Number(n.getColumnValue(u.value, x)), t.height);
-          e[o] = {
-            height: d,
-            key: String(o)
-          }, c += d;
-        });
-        const p = c;
-        t.normalize && a.secondary.forEach((r) => {
-          const o = r === k ? "null" : r;
-          e[o] && (e[o].height = e[o].height / p * 100);
-        }), m = Math.max(m, 0, c), i = Math.min(i, 0, c);
-        const S = t.normalize ? 100 : p;
-        s[l].push({
-          key: `stackedBar_${String(h)}`,
-          valuesMap: e,
-          primaryGrouping: h,
-          boundsY: { min: Math.min(S, 0), max: Math.max(S, 0) }
-        });
-      }), s;
-    }, {}),
-    meta: { minY: t.normalize ? 0 : i, maxY: t.normalize ? 100 : m, normalize: t.normalize },
-    aes: t.aes
-  };
+function l(e, r, i, a) {
+	let o = -Infinity, s = Infinity;
+	return {
+		type: "stackedBar",
+		geoms: r.facet.reduce((l, u) => {
+			let d = n(u);
+			return l[d] = [], r.primary.forEach((n) => {
+				let f = 0, p = {};
+				r.secondary.forEach((r) => {
+					let o = r === t ? "null" : r, s = i.getRowsByGrouping([
+						...u,
+						n,
+						r
+					]);
+					if (!s.length) return;
+					let l = c(s, (e) => Number(i.getColumnValue(a.value, e)), e.height);
+					p[o] = {
+						height: l,
+						key: String(o)
+					}, f += l;
+				});
+				let m = f;
+				e.normalize && r.secondary.forEach((e) => {
+					let n = e === t ? "null" : e;
+					p[n] && (p[n].height = p[n].height / m * 100);
+				}), o = Math.max(o, 0, f), s = Math.min(s, 0, f);
+				let h = e.normalize ? 100 : m;
+				l[d].push({
+					key: `stackedBar_${String(n)}`,
+					valuesMap: p,
+					primaryGrouping: n,
+					boundsY: {
+						min: Math.min(h, 0),
+						max: Math.max(h, 0)
+					}
+				});
+			}), l;
+		}, {}),
+		meta: {
+			minY: e.normalize ? 0 : s,
+			maxY: e.normalize ? 100 : o,
+			normalize: e.normalize
+		},
+		aes: e.aes
+	};
 }
-export {
-  R as getStackedBarsData
-};
-//# sourceMappingURL=stackedBar.js.map
+export { l as getStackedBarsData };
+
+//# sourceMappingURL=stackedBar.js.map

package/dist/discrete/layers/stackedBar.js.map

@@ -1 +1 @@
-{"version":3,"file":"stackedBar.js","sources":["../../../src/discrete/layers/stackedBar.ts"],"sourcesContent":["import { max as d3Max, min as d3Min, mean, quantileSorted, sum } from 'd3-array';\nimport type { DataFrame, GroupKey } from '../../DataFrame';\nimport { NO_GROUPED } from '../../DataFrame';\nimport type { AggregationMethod, ColumnName } from '../../types';\nimport type { StackedBarLayer } from '../DiscreteSettingsImpl';\nimport { getFacetStringKey } from '../utils/getFacetStringKey';\nimport type { StackedBarsData } from './types';\nimport { exhaustive } from '../../utils';\n\ntype BarPart = {\n    height: number;\n    key: string;\n};\n\nexport type StackedBar = {\n    key: string;\n    valuesMap: Record<string, BarPart>;\n    primaryGrouping: string;\n    boundsY: {\n        min: number;\n        max: number;\n    };\n};\n\nfunction calculateHeight(\n    indexes: Uint32Array,\n    getter: (idx: number) => number,\n    heightFunc: AggregationMethod,\n):number {\n    if (indexes.length === 0) {\n        throw Error('Unable to create bar on empty data');\n    }\n\n    switch (heightFunc) {\n        case 'sum':\n            return sum(indexes, getter);\n        case 'max':\n            return d3Max(indexes, getter) ?? getter(0);\n        case 'min':\n            return d3Min(indexes, getter) ?? getter(0);\n        case 'median': {\n            const valuesSorted = [...indexes].map(getter).sort((a, b) => a - b);\n            return quantileSorted(valuesSorted, 0.5) ?? getter(0);\n        }\n        case 'mean': {\n            return mean(indexes, getter) ?? getter(0);\n        }\n        default: exhaustive(heightFunc, `Unknown bar height function ${heightFunc}`);\n    }\n}\n\nexport function getStackedBarsData(\n    layer: StackedBarLayer,\n    groupingKeys: {\n        facet: GroupKey[][],\n        primary: GroupKey[],\n        secondary: GroupKey[]\n    },\n    data: DataFrame,\n    y: ColumnName,\n): StackedBarsData {\n    let maxY = -Infinity;\n    let minY = Infinity;\n\n    return {\n        type: 'stackedBar',\n        geoms: groupingKeys.facet.reduce((res: Record<string, StackedBar[]>, facetKey) => {\n            const strFacetKey = getFacetStringKey(facetKey);\n            res[strFacetKey] = [];\n            groupingKeys.primary.forEach(primaryKey => {\n                let currentHeight = 0;\n                const barParts:Record<string, BarPart> = {};\n                groupingKeys.secondary.forEach(secondaryKey => {\n                    const secondaryKeyStr = secondaryKey === NO_GROUPED ? 'null' : secondaryKey;\n                    const indexes = data.getRowsByGrouping([...facetKey, primaryKey, secondaryKey]);\n                    if (!indexes.length) {\n                        return;\n                    }\n                    const height = calculateHeight(indexes, (i) => Number(data.getColumnValue(y.value, i)), layer.height);\n                    barParts[secondaryKeyStr] = {\n                        height,\n                        key: String(secondaryKeyStr),\n                    };\n                    currentHeight += height;\n                });\n                const maxStackHeight = currentHeight;\n                if (layer.normalize) { // show in percents of every stack\n                    groupingKeys.secondary.forEach(secondaryKey => {\n                        const secondaryKeyStr = secondaryKey === NO_GROUPED ? 'null' : secondaryKey;\n                        if (!barParts[secondaryKeyStr]) {\n                            return;\n                        }\n                        barParts[secondaryKeyStr].height = barParts[secondaryKeyStr].height /  maxStackHeight * 100;\n                    });\n                }\n                maxY = Math.max(maxY, 0, currentHeight);\n                minY = Math.min(minY, 0, currentHeight);\n                const bound = layer.normalize ? 100 : maxStackHeight;\n                res[strFacetKey].push({\n                    key: `stackedBar_${String(primaryKey)}`,\n                    valuesMap: barParts,\n                    primaryGrouping: primaryKey,\n                    boundsY: {min: Math.min(bound, 0), max: Math.max(bound, 0)},\n                } as StackedBar);\n            });\n            return res;\n        }, {}),\n        meta: {minY: layer.normalize ? 0 : minY, maxY: layer.normalize ? 100 : maxY, normalize: layer.normalize},\n        aes: layer.aes,\n    };\n}\n"],"names":["calculateHeight","indexes","getter","heightFunc","sum","d3Max","d3Min","valuesSorted","a","b","quantileSorted","mean","exhaustive","getStackedBarsData","layer","groupingKeys","data","y","maxY","minY","res","facetKey","strFacetKey","getFacetStringKey","primaryKey","currentHeight","barParts","secondaryKey","secondaryKeyStr","NO_GROUPED","height","i","maxStackHeight","bound"],"mappings":";;;;;;;;AAwBA,SAASA,EACLC,GACAC,GACAC,GACK;AACL,MAAIF,EAAQ,WAAW;AACnB,UAAM,MAAM,oCAAoC;AAGpD,UAAQE,GAAA;AAAA,IACJ,KAAK;AACD,aAAOC,EAAIH,GAASC,CAAM;AAAA,IAC9B,KAAK;AACD,aAAOG,EAAMJ,GAASC,CAAM,KAAKA,EAAO,CAAC;AAAA,IAC7C,KAAK;AACD,aAAOI,EAAML,GAASC,CAAM,KAAKA,EAAO,CAAC;AAAA,IAC7C,KAAK,UAAU;AACX,YAAMK,IAAe,CAAC,GAAGN,CAAO,EAAE,IAAIC,CAAM,EAAE,KAAK,CAACM,GAAGC,MAAMD,IAAIC,CAAC;AAClE,aAAOC,EAAeH,GAAc,GAAG,KAAKL,EAAO,CAAC;AAAA,IACxD;AAAA,IACA,KAAK;AACD,aAAOS,EAAKV,GAASC,CAAM,KAAKA,EAAO,CAAC;AAAA,IAE5C;AAAS,MAAAU,EAAWT,GAAY,+BAA+BA,CAAU,EAAE;AAAA,EAAA;AAEnF;AAEO,SAASU,EACZC,GACAC,GAKAC,GACAC,GACe;AACf,MAAIC,IAAO,QACPC,IAAO;AAEX,SAAO;AAAA,IACH,MAAM;AAAA,IACN,OAAOJ,EAAa,MAAM,OAAO,CAACK,GAAmCC,MAAa;AAC9E,YAAMC,IAAcC,EAAkBF,CAAQ;AAC9C,aAAAD,EAAIE,CAAW,IAAI,CAAA,GACnBP,EAAa,QAAQ,QAAQ,CAAAS,MAAc;AACvC,YAAIC,IAAgB;AACpB,cAAMC,IAAmC,CAAA;AACzC,QAAAX,EAAa,UAAU,QAAQ,CAAAY,MAAgB;AAC3C,gBAAMC,IAAkBD,MAAiBE,IAAa,SAASF,GACzD1B,IAAUe,EAAK,kBAAkB,CAAC,GAAGK,GAAUG,GAAYG,CAAY,CAAC;AAC9E,cAAI,CAAC1B,EAAQ;AACT;AAEJ,gBAAM6B,IAAS9B,EAAgBC,GAAS,CAAC8B,MAAM,OAAOf,EAAK,eAAeC,EAAE,OAAOc,CAAC,CAAC,GAAGjB,EAAM,MAAM;AACpG,UAAAY,EAASE,CAAe,IAAI;AAAA,YACxB,QAAAE;AAAA,YACA,KAAK,OAAOF,CAAe;AAAA,UAAA,GAE/BH,KAAiBK;AAAA,QACrB,CAAC;AACD,cAAME,IAAiBP;AACvB,QAAIX,EAAM,aACNC,EAAa,UAAU,QAAQ,CAAAY,MAAgB;AAC3C,gBAAMC,IAAkBD,MAAiBE,IAAa,SAASF;AAC/D,UAAKD,EAASE,CAAe,MAG7BF,EAASE,CAAe,EAAE,SAASF,EAASE,CAAe,EAAE,SAAUI,IAAiB;AAAA,QAC5F,CAAC,GAELd,IAAO,KAAK,IAAIA,GAAM,GAAGO,CAAa,GACtCN,IAAO,KAAK,IAAIA,GAAM,GAAGM,CAAa;AACtC,cAAMQ,IAAQnB,EAAM,YAAY,MAAMkB;AACtC,QAAAZ,EAAIE,CAAW,EAAE,KAAK;AAAA,UAClB,KAAK,cAAc,OAAOE,CAAU,CAAC;AAAA,UACrC,WAAWE;AAAA,UACX,iBAAiBF;AAAA,UACjB,SAAS,EAAC,KAAK,KAAK,IAAIS,GAAO,CAAC,GAAG,KAAK,KAAK,IAAIA,GAAO,CAAC,EAAA;AAAA,QAAC,CAC/C;AAAA,MACnB,CAAC,GACMb;AAAA,IACX,GAAG,CAAA,CAAE;AAAA,IACL,MAAM,EAAC,MAAMN,EAAM,YAAY,IAAIK,GAAM,MAAML,EAAM,YAAY,MAAMI,GAAM,WAAWJ,EAAM,UAAA;AAAA,IAC9F,KAAKA,EAAM;AAAA,EAAA;AAEnB;"}
+{"version":3,"file":"stackedBar.js","names":[],"sources":["../../../src/discrete/layers/stackedBar.ts"],"sourcesContent":["import { max as d3Max, min as d3Min, mean, quantileSorted, sum } from 'd3-array';\nimport type { DataFrame, GroupKey } from '../../DataFrame';\nimport { NO_GROUPED } from '../../DataFrame';\nimport type { AggregationMethod, ColumnName } from '../../types';\nimport type { StackedBarLayer } from '../DiscreteSettingsImpl';\nimport { getFacetStringKey } from '../utils/getFacetStringKey';\nimport type { StackedBarsData } from './types';\nimport { exhaustive } from '../../utils';\n\ntype BarPart = {\n    height: number;\n    key: string;\n};\n\nexport type StackedBar = {\n    key: string;\n    valuesMap: Record<string, BarPart>;\n    primaryGrouping: string;\n    boundsY: {\n        min: number;\n        max: number;\n    };\n};\n\nfunction calculateHeight(\n    indexes: Uint32Array,\n    getter: (idx: number) => number,\n    heightFunc: AggregationMethod,\n):number {\n    if (indexes.length === 0) {\n        throw Error('Unable to create bar on empty data');\n    }\n\n    switch (heightFunc) {\n        case 'sum':\n            return sum(indexes, getter);\n        case 'max':\n            return d3Max(indexes, getter) ?? getter(0);\n        case 'min':\n            return d3Min(indexes, getter) ?? getter(0);\n        case 'median': {\n            const valuesSorted = [...indexes].map(getter).sort((a, b) => a - b);\n            return quantileSorted(valuesSorted, 0.5) ?? getter(0);\n        }\n        case 'mean': {\n            return mean(indexes, getter) ?? getter(0);\n        }\n        default: exhaustive(heightFunc, `Unknown bar height function ${heightFunc}`);\n    }\n}\n\nexport function getStackedBarsData(\n    layer: StackedBarLayer,\n    groupingKeys: {\n        facet: GroupKey[][],\n        primary: GroupKey[],\n        secondary: GroupKey[]\n    },\n    data: DataFrame,\n    y: ColumnName,\n): StackedBarsData {\n    let maxY = -Infinity;\n    let minY = Infinity;\n\n    return {\n        type: 'stackedBar',\n        geoms: groupingKeys.facet.reduce((res: Record<string, StackedBar[]>, facetKey) => {\n            const strFacetKey = getFacetStringKey(facetKey);\n            res[strFacetKey] = [];\n            groupingKeys.primary.forEach(primaryKey => {\n                let currentHeight = 0;\n                const barParts:Record<string, BarPart> = {};\n                groupingKeys.secondary.forEach(secondaryKey => {\n                    const secondaryKeyStr = secondaryKey === NO_GROUPED ? 'null' : secondaryKey;\n                    const indexes = data.getRowsByGrouping([...facetKey, primaryKey, secondaryKey]);\n                    if (!indexes.length) {\n                        return;\n                    }\n                    const height = calculateHeight(indexes, (i) => Number(data.getColumnValue(y.value, i)), layer.height);\n                    barParts[secondaryKeyStr] = {\n                        height,\n                        key: String(secondaryKeyStr),\n                    };\n                    currentHeight += height;\n                });\n                const maxStackHeight = currentHeight;\n                if (layer.normalize) { // show in percents of every stack\n                    groupingKeys.secondary.forEach(secondaryKey => {\n                        const secondaryKeyStr = secondaryKey === NO_GROUPED ? 'null' : secondaryKey;\n                        if (!barParts[secondaryKeyStr]) {\n                            return;\n                        }\n                        barParts[secondaryKeyStr].height = barParts[secondaryKeyStr].height /  maxStackHeight * 100;\n                    });\n                }\n                maxY = Math.max(maxY, 0, currentHeight);\n                minY = Math.min(minY, 0, currentHeight);\n                const bound = layer.normalize ? 100 : maxStackHeight;\n                res[strFacetKey].push({\n                    key: `stackedBar_${String(primaryKey)}`,\n                    valuesMap: barParts,\n                    primaryGrouping: primaryKey,\n                    boundsY: {min: Math.min(bound, 0), max: Math.max(bound, 0)},\n                } as StackedBar);\n            });\n            return res;\n        }, {}),\n        meta: {minY: layer.normalize ? 0 : minY, maxY: layer.normalize ? 100 : maxY, normalize: layer.normalize},\n        aes: layer.aes,\n    };\n}\n"],"mappings":";;;;AAwBA,SAAS,EACL,GACA,GACA,GACK;AACL,KAAI,EAAQ,WAAW,EACnB,OAAM,MAAM,qCAAqC;AAGrD,SAAQ,GAAR;EACI,KAAK,MACD,QAAO,EAAI,GAAS,EAAO;EAC/B,KAAK,MACD,QAAO,EAAM,GAAS,EAAO,IAAI,EAAO,EAAE;EAC9C,KAAK,MACD,QAAO,EAAM,GAAS,EAAO,IAAI,EAAO,EAAE;EAC9C,KAAK,SAED,QAAO,EADc,CAAC,GAAG,EAAQ,CAAC,IAAI,EAAO,CAAC,MAAM,GAAG,MAAM,IAAI,EAAE,EAC/B,GAAI,IAAI,EAAO,EAAE;EAEzD,KAAK,OACD,QAAO,EAAK,GAAS,EAAO,IAAI,EAAO,EAAE;EAE7C,QAAS,GAAW,GAAY,+BAA+B,IAAa;;;AAIpF,SAAgB,EACZ,GACA,GAKA,GACA,GACe;CACf,IAAI,IAAO,WACP,IAAO;AAEX,QAAO;EACH,MAAM;EACN,OAAO,EAAa,MAAM,QAAQ,GAAmC,MAAa;GAC9E,IAAM,IAAc,EAAkB,EAAS;AAsC/C,UArCA,EAAI,KAAe,EAAE,EACrB,EAAa,QAAQ,SAAQ,MAAc;IACvC,IAAI,IAAgB,GACd,IAAmC,EAAE;AAC3C,MAAa,UAAU,SAAQ,MAAgB;KAC3C,IAAM,IAAkB,MAAiB,IAAa,SAAS,GACzD,IAAU,EAAK,kBAAkB;MAAC,GAAG;MAAU;MAAY;MAAa,CAAC;AAC/E,SAAI,CAAC,EAAQ,OACT;KAEJ,IAAM,IAAS,EAAgB,IAAU,MAAM,OAAO,EAAK,eAAe,EAAE,OAAO,EAAE,CAAC,EAAE,EAAM,OAAO;AAKrG,KAJA,EAAS,KAAmB;MACxB;MACA,KAAK,OAAO,EAAgB;MAC/B,EACD,KAAiB;MACnB;IACF,IAAM,IAAiB;AAWvB,IAVI,EAAM,aACN,EAAa,UAAU,SAAQ,MAAgB;KAC3C,IAAM,IAAkB,MAAiB,IAAa,SAAS;AAC1D,OAAS,OAGd,EAAS,GAAiB,SAAS,EAAS,GAAiB,SAAU,IAAiB;MAC1F,EAEN,IAAO,KAAK,IAAI,GAAM,GAAG,EAAc,EACvC,IAAO,KAAK,IAAI,GAAM,GAAG,EAAc;IACvC,IAAM,IAAQ,EAAM,YAAY,MAAM;AACtC,MAAI,GAAa,KAAK;KAClB,KAAK,cAAc,OAAO,EAAW;KACrC,WAAW;KACX,iBAAiB;KACjB,SAAS;MAAC,KAAK,KAAK,IAAI,GAAO,EAAE;MAAE,KAAK,KAAK,IAAI,GAAO,EAAE;MAAC;KAC9D,CAAe;KAClB,EACK;KACR,EAAE,CAAC;EACN,MAAM;GAAC,MAAM,EAAM,YAAY,IAAI;GAAM,MAAM,EAAM,YAAY,MAAM;GAAM,WAAW,EAAM;GAAU;EACxG,KAAK,EAAM;EACd"}

package/dist/discrete/layers/stats/MannWhitneyU.js

@@ -1,149 +1,115 @@
-var P = Object.defineProperty;
-var w = (i, t, e) => t in i ? P(i, t, { enumerable: !0, configurable: !0, writable: !0, value: e }) : i[t] = e;
-var f = (i, t, e) => w(i, typeof t != "symbol" ? t + "" : t, e);
-import m from "../../../_virtual/lodash.js";
-function U(i) {
-  const t = Math.abs(i), e = 1 / (1 + t / 2), n = e * Math.exp(-t * t - 1.26551223 + e * (1.00002368 + e * (0.37409196 + e * (0.09678418 + e * (-0.18628806 + e * (0.27886807 + e * (-1.13520398 + e * (1.48851587 + e * (-0.82215223 + e * 0.17087277)))))))));
-  return i >= 0 ? n : 2 - n;
+import e from "lodash";
+function t(e) {
+	let t = Math.abs(e), n = 1 / (1 + t / 2), r = n * Math.exp(-t * t - 1.26551223 + n * (1.00002368 + n * (.37409196 + n * (.09678418 + n * (-.18628806 + n * (.27886807 + n * (-1.13520398 + n * (1.48851587 + n * (-.82215223 + n * .17087277)))))))));
+	return e >= 0 ? r : 2 - r;
 }
-function b(i) {
-  return 0.5 * U(-i / Math.sqrt(2));
+function n(e) {
+	return .5 * t(-e / Math.sqrt(2));
 }
-class N {
-  constructor(t, e) {
-    this.value = t, this.position = e;
-  }
-  getValue() {
-    return this.value;
-  }
-  getPosition() {
-    return this.position;
-  }
-}
-function k(i) {
-  const t = new Array(i.length);
-  for (let s = 0; s < i.length; s++)
-    t[s] = new N(i[s], s);
-  const e = E(t);
-  t.sort((s, h) => s.getValue() - h.getValue());
-  const n = new Array(t.length);
-  let r = 1;
-  n[t[0].getPosition()] = r;
-  let o = [];
-  o.push(t[0].getPosition());
-  for (let s = 1; s < t.length; s++)
-    t[s].getValue() > t[s - 1].getValue() && (r = s + 1, o.length > 1 && p(n, o), o = []), o.push(t[s].getPosition()), n[t[s].getPosition()] = r;
-  return o.length > 1 && p(n, o), A(n, e), n;
-}
-function p(i, t) {
-  const n = (2 * i[t[0]] + t.length - 1) / 2;
-  t.forEach((r) => {
-    i[r] = n;
-  });
+var r = class {
+	constructor(e, t) {
+		this.value = e, this.position = t;
+	}
+	getValue() {
+		return this.value;
+	}
+	getPosition() {
+		return this.position;
+	}
+};
+function i(e) {
+	let t = Array(e.length);
+	for (let n = 0; n < e.length; n++) t[n] = new r(e[n], n);
+	let n = s(t);
+	t.sort((e, t) => e.getValue() - t.getValue());
+	let i = Array(t.length), c = 1;
+	i[t[0].getPosition()] = c;
+	let l = [];
+	l.push(t[0].getPosition());
+	for (let e = 1; e < t.length; e++) t[e].getValue() > t[e - 1].getValue() ? (c = e + 1, l.length > 1 && a(i, l), l = [], l.push(t[e].getPosition())) : l.push(t[e].getPosition()), i[t[e].getPosition()] = c;
+	return l.length > 1 && a(i, l), o(i, n), i;
 }
-function A(i, t) {
-  t.forEach((e) => {
-    i[e] = NaN;
-  });
+function a(e, t) {
+	let n = (2 * e[t[0]] + t.length - 1) / 2;
+	t.forEach((t) => {
+		e[t] = n;
+	});
 }
-function E(i) {
-  const t = [];
-  for (let e = 0; e < i.length; e++)
-    isNaN(i[e].getValue()) && t.push(e);
-  return t;
+function o(e, t) {
+	t.forEach((t) => {
+		e[t] = NaN;
+	});
 }
-class L {
-  constructor() {
-    f(this, "SMALL_SAMPLE_SIZE", 50);
-  }
-  mannWhitneyU(t, e) {
-    const n = [...t, ...e], r = k(n);
-    let o = 0;
-    for (let l = 0; l < t.length; l++)
-      o += r[l];
-    const s = o - t.length * (t.length + 1) / 2, h = t.length * e.length - s;
-    return Math.min(s, h);
-  }
-  mannWhitneyUTest(t, e, n) {
-    if (typeof n > "u")
-      return t.length + e.length <= this.SMALL_SAMPLE_SIZE && Object.keys(this.tiesMap(t, e)).length === 0 ? this.mannWhitneyUTest(t, e, !0) : this.mannWhitneyUTest(t, e, !1);
-    const r = this.tiesMap(t, e), o = this.mannWhitneyU(t, e);
-    if (n) {
-      if (Object.keys(r).length !== 0)
-        throw new Error();
-      return this.exactP(t.length, e.length, o);
-    }
-    return this.approximateP(
-      o,
-      t.length,
-      e.length,
-      this.varU(t.length, e.length, r)
-    );
-  }
-  approximateP(t, e, n, r) {
-    const o = e * n / 2;
-    if (o == t)
-      return 1;
-    const s = -Math.abs(t - o + 0.5) / Math.sqrt(r);
-    return 2 * b(s);
-  }
-  exactP(t, e, n) {
-    const r = e * t;
-    if (n > r)
-      return 1;
-    const o = n < r / 2 ? n : r / 2 - n;
-    let s = 0, h = 0;
-    for (; h <= o; )
-      s += this.uDensity(t, e, h), h++;
-    return 2 * s;
-  }
-  uDensity(t, e, n) {
-    if (n < 0 || n > e * t)
-      return 0;
-    const r = this.uFrequencies(t, e);
-    return r[Math.round(n + 1)] / m.sum(r);
-  }
-  uFrequencies(t, e) {
-    const n = Math.max(e, t);
-    if (n > 100)
-      throw Error("n or m > 100");
-    const r = Math.min(e, t), o = new Array(t * e + 2), s = new Array(t * e + 2);
-    for (let l = 1; l < o.length; l++)
-      o[l] = l <= n + 1 ? 1 : 0;
-    s[1] = 0;
-    let h = n;
-    for (let l = 2; l <= r; l++) {
-      s[l] = 0, h = h + n;
-      let u = h + 2;
-      const M = 1 + h / 2;
-      let g = l;
-      for (let a = 1; a <= M; a++) {
-        g++, u = u - 1;
-        const c = o[a] + s[a];
-        o[a] = c, s[g] = c - o[u], o[u] = c;
-      }
-    }
-    return o;
-  }
-  varU(t, e, n) {
-    const r = t * e;
-    if (Object.keys(n).length === 0)
-      return r * (t + e + 1) / 12;
-    const o = m.sum(Object.values(n).map((h) => h * h * h - h)), s = t + e;
-    return r / 12 * (s + 1 - o / (s * (s - 1)));
-  }
-  tiesMap(t, e) {
-    const n = {};
-    for (let r = 0; r < t.length; r++)
-      n[t[r]] = n[t[r]] ? n[t[r]] + 1 : 1;
-    for (let r = 0; r < e.length; r++)
-      n[e[r]] = n[e[r]] ? n[e[r]] + 1 : 1;
-    for (const r in n)
-      n[r] === 1 && delete n[r];
-    return n;
-  }
+function s(e) {
+	let t = [];
+	for (let n = 0; n < e.length; n++) isNaN(e[n].getValue()) && t.push(n);
+	return t;
 }
-export {
-  L as MannWhitneyU
+var c = class {
+	SMALL_SAMPLE_SIZE = 50;
+	mannWhitneyU(e, t) {
+		let n = i([...e, ...t]), r = 0;
+		for (let t = 0; t < e.length; t++) r += n[t];
+		let a = r - e.length * (e.length + 1) / 2, o = e.length * t.length - a;
+		return Math.min(a, o);
+	}
+	mannWhitneyUTest(e, t, n) {
+		if (n === void 0) return e.length + t.length <= this.SMALL_SAMPLE_SIZE && Object.keys(this.tiesMap(e, t)).length === 0 ? this.mannWhitneyUTest(e, t, !0) : this.mannWhitneyUTest(e, t, !1);
+		let r = this.tiesMap(e, t), i = this.mannWhitneyU(e, t);
+		if (n) {
+			if (Object.keys(r).length !== 0) throw Error();
+			return this.exactP(e.length, t.length, i);
+		}
+		return this.approximateP(i, e.length, t.length, this.varU(e.length, t.length, r));
+	}
+	approximateP(e, t, r, i) {
+		let a = t * r / 2;
+		return a == e ? 1 : 2 * n(-Math.abs(e - a + .5) / Math.sqrt(i));
+	}
+	exactP(e, t, n) {
+		let r = t * e;
+		if (n > r) return 1;
+		let i = n < r / 2 ? n : r / 2 - n, a = 0, o = 0;
+		for (; o <= i;) a += this.uDensity(e, t, o), o++;
+		return 2 * a;
+	}
+	uDensity(t, n, r) {
+		if (r < 0 || r > n * t) return 0;
+		let i = this.uFrequencies(t, n);
+		return i[Math.round(r + 1)] / e.sum(i);
+	}
+	uFrequencies(e, t) {
+		let n = Math.max(t, e);
+		if (n > 100) throw Error("n or m > 100");
+		let r = Math.min(t, e), i = Array(e * t + 2), a = Array(e * t + 2);
+		for (let e = 1; e < i.length; e++) i[e] = e <= n + 1 ? 1 : 0;
+		a[1] = 0;
+		let o = n;
+		for (let e = 2; e <= r; e++) {
+			a[e] = 0, o += n;
+			let t = o + 2, r = 1 + o / 2, s = e;
+			for (let e = 1; e <= r; e++) {
+				s++, --t;
+				let n = i[e] + a[e];
+				i[e] = n, a[s] = n - i[t], i[t] = n;
+			}
+		}
+		return i;
+	}
+	varU(t, n, r) {
+		let i = t * n;
+		if (Object.keys(r).length === 0) return i * (t + n + 1) / 12;
+		let a = e.sum(Object.values(r).map((e) => e * e * e - e)), o = t + n;
+		return i / 12 * (o + 1 - a / (o * (o - 1)));
+	}
+	tiesMap(e, t) {
+		let n = {};
+		for (let t = 0; t < e.length; t++) n[e[t]] = n[e[t]] ? n[e[t]] + 1 : 1;
+		for (let e = 0; e < t.length; e++) n[t[e]] = n[t[e]] ? n[t[e]] + 1 : 1;
+		for (let e in n) n[e] === 1 && delete n[e];
+		return n;
+	}
 };
-//# sourceMappingURL=MannWhitneyU.js.map
+export { c as MannWhitneyU };
+
+//# sourceMappingURL=MannWhitneyU.js.map

package/dist/discrete/layers/stats/MannWhitneyU.js.map

@@ -1 +1 @@
-{"version":3,"file":"MannWhitneyU.js","sources":["../../../../src/discrete/layers/stats/MannWhitneyU.ts"],"sourcesContent":["import lodash from 'lodash';\n\n// Normal distribution - complementary error function\n// https://en.wikipedia.org/wiki/Error_function#Complementary_error_function\n// https://github.com/errcw/gaussian/blob/master/lib/gaussian.js#L5\nfunction erfc(x:number):number {\n    const z = Math.abs(x);\n    const t = 1 / (1 + z / 2);\n    const r = t * Math.exp(-z * z - 1.26551223 + t * (1.00002368 +\n        t * (0.37409196 + t * (0.09678418 + t * (-0.18628806 +\n            t * (0.27886807 + t * (-1.13520398 + t * (1.48851587 +\n                t * (-0.82215223 + t * 0.17087277)))))))));\n    return x >= 0 ? r : 2 - r;\n}\n// Normal distribution - cumulative probability, for mean = 0, standardDeviation = 1\n// https://github.com/errcw/gaussian/blob/master/lib/gaussian.js#L54\nfunction standardNormalCumulativeProbability (x:number):number {\n    return 0.5 * erfc(-x / (Math.sqrt(2)));\n}\n\n// Natural ranking\n// https://github.com/apache/commons-math/blob/ece74cff78ae7858827f1913bb63854601d88c06/commons-math-legacy/src/main/java/org/apache/commons/math4/legacy/stat/ranking/NaturalRanking.java#L195\nclass IntDoublePair {\n    constructor(private value:number, private position:number) {}\n    public getValue() {\n        return this.value;\n    }\n    public getPosition() {\n        return this.position;\n    }\n}\n// with NaNStrategy.FIXED, TiesStrategy.AVERAGE https://github.com/milaboratory/miplots/blob/559b13065f17dfc7c4dd332247dd1ce0183affcc/src/main/kotlin/com/milaboratory/miplots/stat/util/MannWhitneyU.kt#L36\nfunction rank(data:number[]) {\n    // Array recording initial positions of data to be ranked\n    const ranks:IntDoublePair[] = new Array(data.length);\n    for (let i = 0; i < data.length; i++) {\n        ranks[i] = new IntDoublePair(data[i], i);\n    }\n\n    // Recode, remove or record positions of NaNs\n    const nanPositions = getNanPositions(ranks);\n    // Sort the IntDoublePairs\n    ranks.sort((a, b) => a.getValue() - b.getValue());\n\n    // Walk the sorted array, filling output array using sorted positions,\n    // resolving ties as we go\n    const out:number[] = new Array(ranks.length);\n    let pos = 1;  // position in sorted array\n    out[ranks[0].getPosition()] = pos;\n    let tiesTrace = [];\n    tiesTrace.push(ranks[0].getPosition());\n\n    for (let i = 1; i < ranks.length; i++) {\n        if (ranks[i].getValue() > ranks[i - 1].getValue()) {\n            // tie sequence has ended (or had length 1)\n            pos = i + 1;\n            if (tiesTrace.length > 1) {  // if seq is nontrivial, resolve\n                resolveTie(out, tiesTrace);\n            }\n            tiesTrace = [];\n            tiesTrace.push(ranks[i].getPosition());\n        } else {\n            // tie sequence continues\n            tiesTrace.push(ranks[i].getPosition());\n        }\n        out[ranks[i].getPosition()] = pos;\n    }\n\n    if (tiesTrace.length > 1) {  // handle tie sequence at end\n        resolveTie(out, tiesTrace);\n    }\n    restoreNaNs(out, nanPositions);\n    return out;\n}\n\nfunction resolveTie(ranks:number[], tiesTrace:number[]) {\n    // constant value of ranks over tiesTrace\n    const c = ranks[tiesTrace[0]];\n    const value = (2 * c + tiesTrace.length - 1) / 2;\n\n    tiesTrace.forEach((trace) => {\n        ranks[trace] = value;\n    });\n}\n\nfunction restoreNaNs(ranks:number[], nanPositions:number[]) {\n    nanPositions.forEach((position) => {\n        ranks[position] = NaN;\n    });\n}\n\nfunction getNanPositions(ranks:IntDoublePair[] ) {\n    const out = [];\n    for (let i = 0; i < ranks.length; i++) {\n        if (isNaN(ranks[i].getValue())){\n            out.push(i);\n        }\n    }\n    return out;\n}\n\n// https://github.com/milaboratory/miplots/blob/559b13065f17dfc7c4dd332247dd1ce0183affcc/src/main/kotlin/com/milaboratory/miplots/stat/util/MannWhitneyU.kt#L36\nexport class MannWhitneyU {\n    SMALL_SAMPLE_SIZE = 50;\n\n    mannWhitneyU (x: number[], y: number[]): number {\n        const z = [...x, ...y];\n        const ranks = rank(z);\n        let sumRankX = 0;\n\n        /*\n         * The ranks for x is in the first x.length entries in ranks because x\n         * is in the first x.length entries in z\n         */for (let i = 0; i < x.length; i++) {\n            sumRankX += ranks[i];\n        }\n\n        /*\n         * U1 = R1 - (n1 * (n1 + 1)) / 2 where R1 is sum of ranks for sample 1,\n         * e.g. x, n1 is the number of observations in sample 1.\n         */\n        const U1 = sumRankX - x.length * (x.length + 1) / 2;\n\n        /*\n         * U1 + U2 = n1 * n2\n         */\n        const U2 = x.length * y.length - U1;\n        return Math.min(U1, U2);\n    }\n\n    mannWhitneyUTest(x: number[], y: number[], exact?:boolean): number {\n        if (typeof exact === 'undefined') {\n            // If samples are both small and there are no ties, perform exact test\n            if (x.length + y.length <= this.SMALL_SAMPLE_SIZE && Object.keys(this.tiesMap(x, y)).length === 0) {\n                return this.mannWhitneyUTest(x, y, true);\n            }\n            // Normal approximation\n            return this.mannWhitneyUTest(x, y, false);\n        }\n\n        const tiesMap = this.tiesMap(x, y);\n        const u = this.mannWhitneyU(x, y);\n        if (exact) {\n            if (Object.keys(tiesMap).length !== 0) {\n                throw new Error();\n            }\n            return this.exactP(x.length, y.length, u);\n        }\n        return this.approximateP(\n            u, x.length, y.length,\n            this.varU(x.length, y.length, tiesMap)\n        );\n    }\n\n    approximateP(u: number, n1: number, n2: number, varU: number): number {\n        const mu = n1 * n2 / 2;\n\n        // If u == mu, return 1\n        if (mu == u) {\n            return 1.0;\n        }\n\n        // Force z <= 0 so we get tail probability. Also apply continuity\n        // correction\n        const z = -Math.abs(u - mu + 0.5) / Math.sqrt(varU);\n        return 2 * standardNormalCumulativeProbability(z);\n    }\n\n    exactP(n: number, m: number, u: number): number {\n        const nm = m * n;\n        if (u > nm) { // Quick exit if u is out of range\n            return 1.0;\n        }\n        // Need to convert u to a mean deviation, so cumulative probability is\n        // tail probability\n        const crit = (u < nm / 2) ? u : nm / 2 - u;\n        let cum = 0;\n        let ct = 0;\n        while (ct <= crit) {\n            cum += this.uDensity(n, m, ct);\n            ct++;\n        }\n        return 2 * cum;\n    }\n\n    uDensity(n: number, m: number, u: number): number {\n        if (u < 0 || u > m * n) {\n            return 0;\n        }\n        const freq = this.uFrequencies(n, m);\n        return freq[Math.round(u + 1)] / lodash.sum(freq);\n    }\n\n    uFrequencies(n: number, m: number): number[] {\n        const max = Math.max(m, n);\n        if (max > 100) {\n            throw Error('n or m > 100');\n        }\n        const min = Math.min(m, n);\n        const out = new Array(n * m + 2);\n        const work = new Array(n * m + 2);\n        for (let i = 1; i < out.length; i++) {\n            out[i] = (i <= max + 1) ? 1 : 0;\n        }\n        work[1] = 0;\n        let in_ = max;\n        for (let i = 2; i <= min; i++) {\n            work[i] = 0;\n            in_ = in_ + max;\n            let n1 = in_ + 2;\n            const l = (1 + in_ / 2);\n            let k = i;\n            for (let j = 1; j <= l; j++) {\n                k++;\n                n1 = n1 - 1;\n                const sum_ = out[j] + work[j];\n                out[j] = sum_;\n                work[k] = sum_ - out[n1];\n                out[n1] = sum_;\n            }\n        }\n        return out;\n    }\n\n    varU(n: number, m: number, tiesMap: Record<number, number>): number {\n        const nm = n * m;\n        if (Object.keys(tiesMap).length === 0) {\n            return nm * (n + m + 1) / 12.0;\n        }\n        const tSum = lodash.sum(Object.values(tiesMap).map(( value) => value * value * value - value));\n        const totalN = n + m;\n        return nm / 12 * (totalN + 1 - tSum / (totalN * (totalN - 1)));\n    }\n\n    tiesMap(x: number[], y: number[]): Record<number, number> {\n        const tiesMap: Record<number, number> = {};\n        for (let i = 0; i < x.length; i++) {\n            tiesMap[x[i]] = tiesMap[x[i]] ? tiesMap[x[i]] + 1 : 1;\n        }\n        for (let i = 0; i < y.length; i++) {\n            tiesMap[y[i]] = tiesMap[y[i]] ? tiesMap[y[i]] + 1 : 1;\n        }\n        for (const key in tiesMap) {\n            if (tiesMap[key] === 1) {\n                delete tiesMap[key];\n            }\n        }\n        return tiesMap;\n    }\n}\n"],"names":["erfc","x","z","t","r","standardNormalCumulativeProbability","IntDoublePair","value","position","rank","data","ranks","i","nanPositions","getNanPositions","a","b","out","pos","tiesTrace","resolveTie","restoreNaNs","trace","MannWhitneyU","__publicField","y","sumRankX","U1","U2","exact","tiesMap","u","n1","n2","varU","mu","n","m","nm","crit","cum","ct","freq","lodash","max","min","work","in_","l","k","j","sum_","tSum","totalN","key"],"mappings":";;;;AAKA,SAASA,EAAKC,GAAiB;AAC3B,QAAMC,IAAI,KAAK,IAAID,CAAC,GACdE,IAAI,KAAK,IAAID,IAAI,IACjBE,IAAID,IAAI,KAAK,IAAI,CAACD,IAAIA,IAAI,aAAaC,KAAK,aAC9CA,KAAK,aAAaA,KAAK,aAAaA,KAAK,cACrCA,KAAK,aAAaA,KAAK,cAAcA,KAAK,aACtCA,KAAK,cAAcA,IAAI,kBAAkB;AACrD,SAAOF,KAAK,IAAIG,IAAI,IAAIA;AAC5B;AAGA,SAASC,EAAqCJ,GAAiB;AAC3D,SAAO,MAAMD,EAAK,CAACC,IAAK,KAAK,KAAK,CAAC,CAAE;AACzC;AAIA,MAAMK,EAAc;AAAA,EAChB,YAAoBC,GAAsBC,GAAiB;AAAvC,SAAA,QAAAD,GAAsB,KAAA,WAAAC;AAAA,EAAkB;AAAA,EACrD,WAAW;AACd,WAAO,KAAK;AAAA,EAChB;AAAA,EACO,cAAc;AACjB,WAAO,KAAK;AAAA,EAChB;AACJ;AAEA,SAASC,EAAKC,GAAe;AAEzB,QAAMC,IAAwB,IAAI,MAAMD,EAAK,MAAM;AACnD,WAASE,IAAI,GAAGA,IAAIF,EAAK,QAAQE;AAC7B,IAAAD,EAAMC,CAAC,IAAI,IAAIN,EAAcI,EAAKE,CAAC,GAAGA,CAAC;AAI3C,QAAMC,IAAeC,EAAgBH,CAAK;AAE1C,EAAAA,EAAM,KAAK,CAACI,GAAGC,MAAMD,EAAE,SAAA,IAAaC,EAAE,UAAU;AAIhD,QAAMC,IAAe,IAAI,MAAMN,EAAM,MAAM;AAC3C,MAAIO,IAAM;AACV,EAAAD,EAAIN,EAAM,CAAC,EAAE,YAAA,CAAa,IAAIO;AAC9B,MAAIC,IAAY,CAAA;AAChB,EAAAA,EAAU,KAAKR,EAAM,CAAC,EAAE,aAAa;AAErC,WAASC,IAAI,GAAGA,IAAID,EAAM,QAAQC;AAC9B,IAAID,EAAMC,CAAC,EAAE,SAAA,IAAaD,EAAMC,IAAI,CAAC,EAAE,eAEnCM,IAAMN,IAAI,GACNO,EAAU,SAAS,KACnBC,EAAWH,GAAKE,CAAS,GAE7BA,IAAY,CAAA,IACZA,EAAU,KAAKR,EAAMC,CAAC,EAAE,aAAa,GAKzCK,EAAIN,EAAMC,CAAC,EAAE,YAAA,CAAa,IAAIM;AAGlC,SAAIC,EAAU,SAAS,KACnBC,EAAWH,GAAKE,CAAS,GAE7BE,EAAYJ,GAAKJ,CAAY,GACtBI;AACX;AAEA,SAASG,EAAWT,GAAgBQ,GAAoB;AAGpD,QAAMZ,KAAS,IADLI,EAAMQ,EAAU,CAAC,CAAC,IACLA,EAAU,SAAS,KAAK;AAE/C,EAAAA,EAAU,QAAQ,CAACG,MAAU;AACzB,IAAAX,EAAMW,CAAK,IAAIf;AAAA,EACnB,CAAC;AACL;AAEA,SAASc,EAAYV,GAAgBE,GAAuB;AACxD,EAAAA,EAAa,QAAQ,CAACL,MAAa;AAC/B,IAAAG,EAAMH,CAAQ,IAAI;AAAA,EACtB,CAAC;AACL;AAEA,SAASM,EAAgBH,GAAwB;AAC7C,QAAMM,IAAM,CAAA;AACZ,WAASL,IAAI,GAAGA,IAAID,EAAM,QAAQC;AAC9B,IAAI,MAAMD,EAAMC,CAAC,EAAE,SAAA,CAAU,KACzBK,EAAI,KAAKL,CAAC;AAGlB,SAAOK;AACX;AAGO,MAAMM,EAAa;AAAA,EAAnB;AACH,IAAAC,EAAA,2BAAoB;AAAA;AAAA,EAEpB,aAAcvB,GAAawB,GAAqB;AAC5C,UAAMvB,IAAI,CAAC,GAAGD,GAAG,GAAGwB,CAAC,GACfd,IAAQF,EAAKP,CAAC;AACpB,QAAIwB,IAAW;AAKZ,aAASd,IAAI,GAAGA,IAAIX,EAAE,QAAQW;AAC7B,MAAAc,KAAYf,EAAMC,CAAC;AAOvB,UAAMe,IAAKD,IAAWzB,EAAE,UAAUA,EAAE,SAAS,KAAK,GAK5C2B,IAAK3B,EAAE,SAASwB,EAAE,SAASE;AACjC,WAAO,KAAK,IAAIA,GAAIC,CAAE;AAAA,EAC1B;AAAA,EAEA,iBAAiB3B,GAAawB,GAAaI,GAAwB;AAC/D,QAAI,OAAOA,IAAU;AAEjB,aAAI5B,EAAE,SAASwB,EAAE,UAAU,KAAK,qBAAqB,OAAO,KAAK,KAAK,QAAQxB,GAAGwB,CAAC,CAAC,EAAE,WAAW,IACrF,KAAK,iBAAiBxB,GAAGwB,GAAG,EAAI,IAGpC,KAAK,iBAAiBxB,GAAGwB,GAAG,EAAK;AAG5C,UAAMK,IAAU,KAAK,QAAQ7B,GAAGwB,CAAC,GAC3BM,IAAI,KAAK,aAAa9B,GAAGwB,CAAC;AAChC,QAAII,GAAO;AACP,UAAI,OAAO,KAAKC,CAAO,EAAE,WAAW;AAChC,cAAM,IAAI,MAAA;AAEd,aAAO,KAAK,OAAO7B,EAAE,QAAQwB,EAAE,QAAQM,CAAC;AAAA,IAC5C;AACA,WAAO,KAAK;AAAA,MACRA;AAAA,MAAG9B,EAAE;AAAA,MAAQwB,EAAE;AAAA,MACf,KAAK,KAAKxB,EAAE,QAAQwB,EAAE,QAAQK,CAAO;AAAA,IAAA;AAAA,EAE7C;AAAA,EAEA,aAAaC,GAAWC,GAAYC,GAAYC,GAAsB;AAClE,UAAMC,IAAKH,IAAKC,IAAK;AAGrB,QAAIE,KAAMJ;AACN,aAAO;AAKX,UAAM7B,IAAI,CAAC,KAAK,IAAI6B,IAAII,IAAK,GAAG,IAAI,KAAK,KAAKD,CAAI;AAClD,WAAO,IAAI7B,EAAoCH,CAAC;AAAA,EACpD;AAAA,EAEA,OAAOkC,GAAWC,GAAWN,GAAmB;AAC5C,UAAMO,IAAKD,IAAID;AACf,QAAIL,IAAIO;AACJ,aAAO;AAIX,UAAMC,IAAQR,IAAIO,IAAK,IAAKP,IAAIO,IAAK,IAAIP;AACzC,QAAIS,IAAM,GACNC,IAAK;AACT,WAAOA,KAAMF;AACT,MAAAC,KAAO,KAAK,SAASJ,GAAGC,GAAGI,CAAE,GAC7BA;AAEJ,WAAO,IAAID;AAAA,EACf;AAAA,EAEA,SAASJ,GAAWC,GAAWN,GAAmB;AAC9C,QAAIA,IAAI,KAAKA,IAAIM,IAAID;AACjB,aAAO;AAEX,UAAMM,IAAO,KAAK,aAAaN,GAAGC,CAAC;AACnC,WAAOK,EAAK,KAAK,MAAMX,IAAI,CAAC,CAAC,IAAIY,EAAO,IAAID,CAAI;AAAA,EACpD;AAAA,EAEA,aAAaN,GAAWC,GAAqB;AACzC,UAAMO,IAAM,KAAK,IAAIP,GAAGD,CAAC;AACzB,QAAIQ,IAAM;AACN,YAAM,MAAM,cAAc;AAE9B,UAAMC,IAAM,KAAK,IAAIR,GAAGD,CAAC,GACnBnB,IAAM,IAAI,MAAMmB,IAAIC,IAAI,CAAC,GACzBS,IAAO,IAAI,MAAMV,IAAIC,IAAI,CAAC;AAChC,aAASzB,IAAI,GAAGA,IAAIK,EAAI,QAAQL;AAC5B,MAAAK,EAAIL,CAAC,IAAKA,KAAKgC,IAAM,IAAK,IAAI;AAElC,IAAAE,EAAK,CAAC,IAAI;AACV,QAAIC,IAAMH;AACV,aAAShC,IAAI,GAAGA,KAAKiC,GAAKjC,KAAK;AAC3B,MAAAkC,EAAKlC,CAAC,IAAI,GACVmC,IAAMA,IAAMH;AACZ,UAAIZ,IAAKe,IAAM;AACf,YAAMC,IAAK,IAAID,IAAM;AACrB,UAAIE,IAAIrC;AACR,eAASsC,IAAI,GAAGA,KAAKF,GAAGE,KAAK;AACzB,QAAAD,KACAjB,IAAKA,IAAK;AACV,cAAMmB,IAAOlC,EAAIiC,CAAC,IAAIJ,EAAKI,CAAC;AAC5B,QAAAjC,EAAIiC,CAAC,IAAIC,GACTL,EAAKG,CAAC,IAAIE,IAAOlC,EAAIe,CAAE,GACvBf,EAAIe,CAAE,IAAImB;AAAA,MACd;AAAA,IACJ;AACA,WAAOlC;AAAA,EACX;AAAA,EAEA,KAAKmB,GAAWC,GAAWP,GAAyC;AAChE,UAAMQ,IAAKF,IAAIC;AACf,QAAI,OAAO,KAAKP,CAAO,EAAE,WAAW;AAChC,aAAOQ,KAAMF,IAAIC,IAAI,KAAK;AAE9B,UAAMe,IAAOT,EAAO,IAAI,OAAO,OAAOb,CAAO,EAAE,IAAI,CAAEvB,MAAUA,IAAQA,IAAQA,IAAQA,CAAK,CAAC,GACvF8C,IAASjB,IAAIC;AACnB,WAAOC,IAAK,MAAMe,IAAS,IAAID,KAAQC,KAAUA,IAAS;AAAA,EAC9D;AAAA,EAEA,QAAQpD,GAAawB,GAAqC;AACtD,UAAMK,IAAkC,CAAA;AACxC,aAASlB,IAAI,GAAGA,IAAIX,EAAE,QAAQW;AAC1B,MAAAkB,EAAQ7B,EAAEW,CAAC,CAAC,IAAIkB,EAAQ7B,EAAEW,CAAC,CAAC,IAAIkB,EAAQ7B,EAAEW,CAAC,CAAC,IAAI,IAAI;AAExD,aAASA,IAAI,GAAGA,IAAIa,EAAE,QAAQb;AAC1B,MAAAkB,EAAQL,EAAEb,CAAC,CAAC,IAAIkB,EAAQL,EAAEb,CAAC,CAAC,IAAIkB,EAAQL,EAAEb,CAAC,CAAC,IAAI,IAAI;AAExD,eAAW0C,KAAOxB;AACd,MAAIA,EAAQwB,CAAG,MAAM,KACjB,OAAOxB,EAAQwB,CAAG;AAG1B,WAAOxB;AAAA,EACX;AACJ;"}
+{"version":3,"file":"MannWhitneyU.js","names":[],"sources":["../../../../src/discrete/layers/stats/MannWhitneyU.ts"],"sourcesContent":["import lodash from 'lodash';\n\n// Normal distribution - complementary error function\n// https://en.wikipedia.org/wiki/Error_function#Complementary_error_function\n// https://github.com/errcw/gaussian/blob/master/lib/gaussian.js#L5\nfunction erfc(x:number):number {\n    const z = Math.abs(x);\n    const t = 1 / (1 + z / 2);\n    const r = t * Math.exp(-z * z - 1.26551223 + t * (1.00002368 +\n        t * (0.37409196 + t * (0.09678418 + t * (-0.18628806 +\n            t * (0.27886807 + t * (-1.13520398 + t * (1.48851587 +\n                t * (-0.82215223 + t * 0.17087277)))))))));\n    return x >= 0 ? r : 2 - r;\n}\n// Normal distribution - cumulative probability, for mean = 0, standardDeviation = 1\n// https://github.com/errcw/gaussian/blob/master/lib/gaussian.js#L54\nfunction standardNormalCumulativeProbability (x:number):number {\n    return 0.5 * erfc(-x / (Math.sqrt(2)));\n}\n\n// Natural ranking\n// https://github.com/apache/commons-math/blob/ece74cff78ae7858827f1913bb63854601d88c06/commons-math-legacy/src/main/java/org/apache/commons/math4/legacy/stat/ranking/NaturalRanking.java#L195\nclass IntDoublePair {\n    constructor(private value:number, private position:number) {}\n    public getValue() {\n        return this.value;\n    }\n    public getPosition() {\n        return this.position;\n    }\n}\n// with NaNStrategy.FIXED, TiesStrategy.AVERAGE https://github.com/milaboratory/miplots/blob/559b13065f17dfc7c4dd332247dd1ce0183affcc/src/main/kotlin/com/milaboratory/miplots/stat/util/MannWhitneyU.kt#L36\nfunction rank(data:number[]) {\n    // Array recording initial positions of data to be ranked\n    const ranks:IntDoublePair[] = new Array(data.length);\n    for (let i = 0; i < data.length; i++) {\n        ranks[i] = new IntDoublePair(data[i], i);\n    }\n\n    // Recode, remove or record positions of NaNs\n    const nanPositions = getNanPositions(ranks);\n    // Sort the IntDoublePairs\n    ranks.sort((a, b) => a.getValue() - b.getValue());\n\n    // Walk the sorted array, filling output array using sorted positions,\n    // resolving ties as we go\n    const out:number[] = new Array(ranks.length);\n    let pos = 1;  // position in sorted array\n    out[ranks[0].getPosition()] = pos;\n    let tiesTrace = [];\n    tiesTrace.push(ranks[0].getPosition());\n\n    for (let i = 1; i < ranks.length; i++) {\n        if (ranks[i].getValue() > ranks[i - 1].getValue()) {\n            // tie sequence has ended (or had length 1)\n            pos = i + 1;\n            if (tiesTrace.length > 1) {  // if seq is nontrivial, resolve\n                resolveTie(out, tiesTrace);\n            }\n            tiesTrace = [];\n            tiesTrace.push(ranks[i].getPosition());\n        } else {\n            // tie sequence continues\n            tiesTrace.push(ranks[i].getPosition());\n        }\n        out[ranks[i].getPosition()] = pos;\n    }\n\n    if (tiesTrace.length > 1) {  // handle tie sequence at end\n        resolveTie(out, tiesTrace);\n    }\n    restoreNaNs(out, nanPositions);\n    return out;\n}\n\nfunction resolveTie(ranks:number[], tiesTrace:number[]) {\n    // constant value of ranks over tiesTrace\n    const c = ranks[tiesTrace[0]];\n    const value = (2 * c + tiesTrace.length - 1) / 2;\n\n    tiesTrace.forEach((trace) => {\n        ranks[trace] = value;\n    });\n}\n\nfunction restoreNaNs(ranks:number[], nanPositions:number[]) {\n    nanPositions.forEach((position) => {\n        ranks[position] = NaN;\n    });\n}\n\nfunction getNanPositions(ranks:IntDoublePair[] ) {\n    const out = [];\n    for (let i = 0; i < ranks.length; i++) {\n        if (isNaN(ranks[i].getValue())){\n            out.push(i);\n        }\n    }\n    return out;\n}\n\n// https://github.com/milaboratory/miplots/blob/559b13065f17dfc7c4dd332247dd1ce0183affcc/src/main/kotlin/com/milaboratory/miplots/stat/util/MannWhitneyU.kt#L36\nexport class MannWhitneyU {\n    SMALL_SAMPLE_SIZE = 50;\n\n    mannWhitneyU (x: number[], y: number[]): number {\n        const z = [...x, ...y];\n        const ranks = rank(z);\n        let sumRankX = 0;\n\n        /*\n         * The ranks for x is in the first x.length entries in ranks because x\n         * is in the first x.length entries in z\n         */for (let i = 0; i < x.length; i++) {\n            sumRankX += ranks[i];\n        }\n\n        /*\n         * U1 = R1 - (n1 * (n1 + 1)) / 2 where R1 is sum of ranks for sample 1,\n         * e.g. x, n1 is the number of observations in sample 1.\n         */\n        const U1 = sumRankX - x.length * (x.length + 1) / 2;\n\n        /*\n         * U1 + U2 = n1 * n2\n         */\n        const U2 = x.length * y.length - U1;\n        return Math.min(U1, U2);\n    }\n\n    mannWhitneyUTest(x: number[], y: number[], exact?:boolean): number {\n        if (typeof exact === 'undefined') {\n            // If samples are both small and there are no ties, perform exact test\n            if (x.length + y.length <= this.SMALL_SAMPLE_SIZE && Object.keys(this.tiesMap(x, y)).length === 0) {\n                return this.mannWhitneyUTest(x, y, true);\n            }\n            // Normal approximation\n            return this.mannWhitneyUTest(x, y, false);\n        }\n\n        const tiesMap = this.tiesMap(x, y);\n        const u = this.mannWhitneyU(x, y);\n        if (exact) {\n            if (Object.keys(tiesMap).length !== 0) {\n                throw new Error();\n            }\n            return this.exactP(x.length, y.length, u);\n        }\n        return this.approximateP(\n            u, x.length, y.length,\n            this.varU(x.length, y.length, tiesMap)\n        );\n    }\n\n    approximateP(u: number, n1: number, n2: number, varU: number): number {\n        const mu = n1 * n2 / 2;\n\n        // If u == mu, return 1\n        if (mu == u) {\n            return 1.0;\n        }\n\n        // Force z <= 0 so we get tail probability. Also apply continuity\n        // correction\n        const z = -Math.abs(u - mu + 0.5) / Math.sqrt(varU);\n        return 2 * standardNormalCumulativeProbability(z);\n    }\n\n    exactP(n: number, m: number, u: number): number {\n        const nm = m * n;\n        if (u > nm) { // Quick exit if u is out of range\n            return 1.0;\n        }\n        // Need to convert u to a mean deviation, so cumulative probability is\n        // tail probability\n        const crit = (u < nm / 2) ? u : nm / 2 - u;\n        let cum = 0;\n        let ct = 0;\n        while (ct <= crit) {\n            cum += this.uDensity(n, m, ct);\n            ct++;\n        }\n        return 2 * cum;\n    }\n\n    uDensity(n: number, m: number, u: number): number {\n        if (u < 0 || u > m * n) {\n            return 0;\n        }\n        const freq = this.uFrequencies(n, m);\n        return freq[Math.round(u + 1)] / lodash.sum(freq);\n    }\n\n    uFrequencies(n: number, m: number): number[] {\n        const max = Math.max(m, n);\n        if (max > 100) {\n            throw Error('n or m > 100');\n        }\n        const min = Math.min(m, n);\n        const out = new Array(n * m + 2);\n        const work = new Array(n * m + 2);\n        for (let i = 1; i < out.length; i++) {\n            out[i] = (i <= max + 1) ? 1 : 0;\n        }\n        work[1] = 0;\n        let in_ = max;\n        for (let i = 2; i <= min; i++) {\n            work[i] = 0;\n            in_ = in_ + max;\n            let n1 = in_ + 2;\n            const l = (1 + in_ / 2);\n            let k = i;\n            for (let j = 1; j <= l; j++) {\n                k++;\n                n1 = n1 - 1;\n                const sum_ = out[j] + work[j];\n                out[j] = sum_;\n                work[k] = sum_ - out[n1];\n                out[n1] = sum_;\n            }\n        }\n        return out;\n    }\n\n    varU(n: number, m: number, tiesMap: Record<number, number>): number {\n        const nm = n * m;\n        if (Object.keys(tiesMap).length === 0) {\n            return nm * (n + m + 1) / 12.0;\n        }\n        const tSum = lodash.sum(Object.values(tiesMap).map(( value) => value * value * value - value));\n        const totalN = n + m;\n        return nm / 12 * (totalN + 1 - tSum / (totalN * (totalN - 1)));\n    }\n\n    tiesMap(x: number[], y: number[]): Record<number, number> {\n        const tiesMap: Record<number, number> = {};\n        for (let i = 0; i < x.length; i++) {\n            tiesMap[x[i]] = tiesMap[x[i]] ? tiesMap[x[i]] + 1 : 1;\n        }\n        for (let i = 0; i < y.length; i++) {\n            tiesMap[y[i]] = tiesMap[y[i]] ? tiesMap[y[i]] + 1 : 1;\n        }\n        for (const key in tiesMap) {\n            if (tiesMap[key] === 1) {\n                delete tiesMap[key];\n            }\n        }\n        return tiesMap;\n    }\n}\n"],"mappings":";AAKA,SAAS,EAAK,GAAiB;CAC3B,IAAM,IAAI,KAAK,IAAI,EAAE,EACf,IAAI,KAAK,IAAI,IAAI,IACjB,IAAI,IAAI,KAAK,IAAI,CAAC,IAAI,IAAI,aAAa,KAAK,aAC9C,KAAK,YAAa,KAAK,YAAa,KAAK,aACrC,KAAK,YAAa,KAAK,cAAc,KAAK,aACtC,KAAK,aAAc,IAAI,kBAAmB;AACtD,QAAO,KAAK,IAAI,IAAI,IAAI;;AAI5B,SAAS,EAAqC,GAAiB;AAC3D,QAAO,KAAM,EAAK,CAAC,IAAK,KAAK,KAAK,EAAE,CAAE;;AAK1C,IAAM,IAAN,MAAoB;CAChB,YAAY,GAAsB,GAAyB;AAAjB,EAAtB,KAAA,QAAA,GAAsB,KAAA,WAAA;;CAC1C,WAAkB;AACd,SAAO,KAAK;;CAEhB,cAAqB;AACjB,SAAO,KAAK;;;AAIpB,SAAS,EAAK,GAAe;CAEzB,IAAM,IAA4B,MAAM,EAAK,OAAO;AACpD,MAAK,IAAI,IAAI,GAAG,IAAI,EAAK,QAAQ,IAC7B,GAAM,KAAK,IAAI,EAAc,EAAK,IAAI,EAAE;CAI5C,IAAM,IAAe,EAAgB,EAAM;AAE3C,GAAM,MAAM,GAAG,MAAM,EAAE,UAAU,GAAG,EAAE,UAAU,CAAC;CAIjD,IAAM,IAAmB,MAAM,EAAM,OAAO,EACxC,IAAM;AACV,GAAI,EAAM,GAAG,aAAa,IAAI;CAC9B,IAAI,IAAY,EAAE;AAClB,GAAU,KAAK,EAAM,GAAG,aAAa,CAAC;AAEtC,MAAK,IAAI,IAAI,GAAG,IAAI,EAAM,QAAQ,IAa9B,CAZI,EAAM,GAAG,UAAU,GAAG,EAAM,IAAI,GAAG,UAAU,IAE7C,IAAM,IAAI,GACN,EAAU,SAAS,KACnB,EAAW,GAAK,EAAU,EAE9B,IAAY,EAAE,EACd,EAAU,KAAK,EAAM,GAAG,aAAa,CAAC,IAGtC,EAAU,KAAK,EAAM,GAAG,aAAa,CAAC,EAE1C,EAAI,EAAM,GAAG,aAAa,IAAI;AAOlC,QAJI,EAAU,SAAS,KACnB,EAAW,GAAK,EAAU,EAE9B,EAAY,GAAK,EAAa,EACvB;;AAGX,SAAS,EAAW,GAAgB,GAAoB;CAGpD,IAAM,KAAS,IADL,EAAM,EAAU,MACH,EAAU,SAAS,KAAK;AAE/C,GAAU,SAAS,MAAU;AACzB,IAAM,KAAS;GACjB;;AAGN,SAAS,EAAY,GAAgB,GAAuB;AACxD,GAAa,SAAS,MAAa;AAC/B,IAAM,KAAY;GACpB;;AAGN,SAAS,EAAgB,GAAwB;CAC7C,IAAM,IAAM,EAAE;AACd,MAAK,IAAI,IAAI,GAAG,IAAI,EAAM,QAAQ,IAC9B,CAAI,MAAM,EAAM,GAAG,UAAU,CAAC,IAC1B,EAAI,KAAK,EAAE;AAGnB,QAAO;;AAIX,IAAa,IAAb,MAA0B;CACtB,oBAAoB;CAEpB,aAAc,GAAa,GAAqB;EAE5C,IAAM,IAAQ,EADJ,CAAC,GAAG,GAAG,GAAG,EAAE,CACD,EACjB,IAAW;AAKZ,OAAK,IAAI,IAAI,GAAG,IAAI,EAAE,QAAQ,IAC7B,MAAY,EAAM;EAOtB,IAAM,IAAK,IAAW,EAAE,UAAU,EAAE,SAAS,KAAK,GAK5C,IAAK,EAAE,SAAS,EAAE,SAAS;AACjC,SAAO,KAAK,IAAI,GAAI,EAAG;;CAG3B,iBAAiB,GAAa,GAAa,GAAwB;AAC/D,MAAW,MAAU,OAMjB,QAJI,EAAE,SAAS,EAAE,UAAU,KAAK,qBAAqB,OAAO,KAAK,KAAK,QAAQ,GAAG,EAAE,CAAC,CAAC,WAAW,IACrF,KAAK,iBAAiB,GAAG,GAAG,GAAK,GAGrC,KAAK,iBAAiB,GAAG,GAAG,GAAM;EAG7C,IAAM,IAAU,KAAK,QAAQ,GAAG,EAAE,EAC5B,IAAI,KAAK,aAAa,GAAG,EAAE;AACjC,MAAI,GAAO;AACP,OAAI,OAAO,KAAK,EAAQ,CAAC,WAAW,EAChC,OAAU,OAAO;AAErB,UAAO,KAAK,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE;;AAE7C,SAAO,KAAK,aACR,GAAG,EAAE,QAAQ,EAAE,QACf,KAAK,KAAK,EAAE,QAAQ,EAAE,QAAQ,EAAQ,CACzC;;CAGL,aAAa,GAAW,GAAY,GAAY,GAAsB;EAClE,IAAM,IAAK,IAAK,IAAK;AAUrB,SAPI,KAAM,IACC,IAMJ,IAAI,EADD,CAAC,KAAK,IAAI,IAAI,IAAK,GAAI,GAAG,KAAK,KAAK,EAAK,CACF;;CAGrD,OAAO,GAAW,GAAW,GAAmB;EAC5C,IAAM,IAAK,IAAI;AACf,MAAI,IAAI,EACJ,QAAO;EAIX,IAAM,IAAQ,IAAI,IAAK,IAAK,IAAI,IAAK,IAAI,GACrC,IAAM,GACN,IAAK;AACT,SAAO,KAAM,GAET,CADA,KAAO,KAAK,SAAS,GAAG,GAAG,EAAG,EAC9B;AAEJ,SAAO,IAAI;;CAGf,SAAS,GAAW,GAAW,GAAmB;AAC9C,MAAI,IAAI,KAAK,IAAI,IAAI,EACjB,QAAO;EAEX,IAAM,IAAO,KAAK,aAAa,GAAG,EAAE;AACpC,SAAO,EAAK,KAAK,MAAM,IAAI,EAAE,IAAI,EAAO,IAAI,EAAK;;CAGrD,aAAa,GAAW,GAAqB;EACzC,IAAM,IAAM,KAAK,IAAI,GAAG,EAAE;AAC1B,MAAI,IAAM,IACN,OAAM,MAAM,eAAe;EAE/B,IAAM,IAAM,KAAK,IAAI,GAAG,EAAE,EACpB,IAAU,MAAM,IAAI,IAAI,EAAE,EAC1B,IAAW,MAAM,IAAI,IAAI,EAAE;AACjC,OAAK,IAAI,IAAI,GAAG,IAAI,EAAI,QAAQ,IAC5B,GAAI,KAAM,KAAK,IAAM,IAAK,IAAI;AAElC,IAAK,KAAK;EACV,IAAI,IAAM;AACV,OAAK,IAAI,IAAI,GAAG,KAAK,GAAK,KAAK;AAE3B,GADA,EAAK,KAAK,GACV,KAAY;GACZ,IAAI,IAAK,IAAM,GACT,IAAK,IAAI,IAAM,GACjB,IAAI;AACR,QAAK,IAAI,IAAI,GAAG,KAAK,GAAG,KAAK;AAEzB,IADA,KACA;IACA,IAAM,IAAO,EAAI,KAAK,EAAK;AAG3B,IAFA,EAAI,KAAK,GACT,EAAK,KAAK,IAAO,EAAI,IACrB,EAAI,KAAM;;;AAGlB,SAAO;;CAGX,KAAK,GAAW,GAAW,GAAyC;EAChE,IAAM,IAAK,IAAI;AACf,MAAI,OAAO,KAAK,EAAQ,CAAC,WAAW,EAChC,QAAO,KAAM,IAAI,IAAI,KAAK;EAE9B,IAAM,IAAO,EAAO,IAAI,OAAO,OAAO,EAAQ,CAAC,KAAM,MAAU,IAAQ,IAAQ,IAAQ,EAAM,CAAC,EACxF,IAAS,IAAI;AACnB,SAAO,IAAK,MAAM,IAAS,IAAI,KAAQ,KAAU,IAAS;;CAG9D,QAAQ,GAAa,GAAqC;EACtD,IAAM,IAAkC,EAAE;AAC1C,OAAK,IAAI,IAAI,GAAG,IAAI,EAAE,QAAQ,IAC1B,GAAQ,EAAE,MAAM,EAAQ,EAAE,MAAM,EAAQ,EAAE,MAAM,IAAI;AAExD,OAAK,IAAI,IAAI,GAAG,IAAI,EAAE,QAAQ,IAC1B,GAAQ,EAAE,MAAM,EAAQ,EAAE,MAAM,EAAQ,EAAE,MAAM,IAAI;AAExD,OAAK,IAAM,KAAO,EACd,CAAI,EAAQ,OAAS,KACjB,OAAO,EAAQ;AAGvB,SAAO"}

package/dist/discrete/layers/stats/index.js

@@ -1,84 +1,56 @@
-import { NO_GROUPED as h } from "../../../DataFrame.js";
-import { calculatePValue as f } from "./pValueCalculation.js";
-import { adjustPValues as d } from "./pValueCorrection.js";
-import { getFacetStringKey as G } from "../../utils/getFacetStringKey.js";
-function k(e) {
-  return e.slice(0, -1).reduce(
-    (a, c, u) => a.concat(e.slice(u + 1).map((i) => [c, i])),
-    []
-  );
+import { NO_GROUPED as e } from "../../../DataFrame.js";
+import { getFacetStringKey as t } from "../../utils/getFacetStringKey.js";
+import { calculatePValue as n } from "./pValueCalculation.js";
+import { adjustPValues as r } from "./pValueCorrection.js";
+function i(e) {
+	return e.slice(0, -1).reduce((t, n, r) => t.concat(e.slice(r + 1).map((e) => [n, e])), []);
 }
-function B(e, a, c, u, i) {
-  return {
-    type: "stats",
-    values: a.reduce((m, p) => {
-      let g = [];
-      switch (e.statType) {
-        case "overall": {
-          const n = c.map((l) => i.getColumnByGrouping([...p, l], u.value)).filter((l) => l.length > 0);
-          if (n.length < 2)
-            break;
-          const r = f(n, e.testMethod, e.paired);
-          g = [[["overall"], r]];
-          break;
-        }
-        case "referenceGroup": {
-          const n = e.referenceGroup === null ? h : String(e.referenceGroup);
-          if (n !== h && !c.includes(n))
-            throw new Error(`Reference group ${String(n)} for ${e.testMethod} test is not in available groups`);
-          const r = i.getColumnByGrouping([...p, n], u.value);
-          if (r.length === 0) {
-            g = [];
-            break;
-          }
-          const t = c.filter((o) => o !== n).map((o) => [o, i.getColumnByGrouping([...p, o], u.value)]).filter(([, o]) => o.length > 0);
-          if (t.length < 2 && e.correctionMethod !== null)
-            break;
-          const s = d(
-            t.map(
-              ([, o]) => f(
-                [r, o],
-                e.testMethod,
-                e.paired
-              )
-            ),
-            e.correctionMethod
-          );
-          g = t.map(([o], v) => [[o], s[v]]);
-          break;
-        }
-        case "pairwise": {
-          const r = k(c).map(([t, s]) => [
-            [t, s],
-            [
-              i.getColumnByGrouping([...p, t], u.value),
-              i.getColumnByGrouping([...p, s], u.value)
-            ]
-          ]).filter(([, [t, s]]) => t.length && s.length);
-          if (r.length < 1)
-            break;
-          const l = d(
-            r.map(([, t]) => f(t, e.testMethod, e.paired)),
-            e.correctionMethod
-          );
-          g = r.map(([t], s) => [t, l[s]]);
-          break;
-        }
-        default:
-          throw new Error("Unknown statType");
-      }
-      return m[G(p)] = g, m;
-    }, {}),
-    meta: {
-      statType: e.statType,
-      testMethod: e.testMethod,
-      format: e.format,
-      ns: e.ns,
-      significanceLevels: e.significanceLevels
-    }
-  };
+function a(a, o, s, c, l) {
+	return {
+		type: "stats",
+		values: o.reduce((o, u) => {
+			let d = [];
+			switch (a.statType) {
+				case "overall": {
+					let e = s.map((e) => l.getColumnByGrouping([...u, e], c.value)).filter((e) => e.length > 0);
+					if (e.length < 2) break;
+					d = [[["overall"], n(e, a.testMethod, a.paired)]];
+					break;
+				}
+				case "referenceGroup": {
+					let t = a.referenceGroup === null ? e : String(a.referenceGroup);
+					if (t !== e && !s.includes(t)) throw Error(`Reference group ${String(t)} for ${a.testMethod} test is not in available groups`);
+					let i = l.getColumnByGrouping([...u, t], c.value);
+					if (i.length === 0) {
+						d = [];
+						break;
+					}
+					let o = s.filter((e) => e !== t).map((e) => [e, l.getColumnByGrouping([...u, e], c.value)]).filter(([, e]) => e.length > 0);
+					if (o.length < 2 && a.correctionMethod !== null) break;
+					let f = r(o.map(([, e]) => n([i, e], a.testMethod, a.paired)), a.correctionMethod);
+					d = o.map(([e], t) => [[e], f[t]]);
+					break;
+				}
+				case "pairwise": {
+					let e = i(s).map(([e, t]) => [[e, t], [l.getColumnByGrouping([...u, e], c.value), l.getColumnByGrouping([...u, t], c.value)]]).filter(([, [e, t]]) => e.length && t.length);
+					if (e.length < 1) break;
+					let t = r(e.map(([, e]) => n(e, a.testMethod, a.paired)), a.correctionMethod);
+					d = e.map(([e], n) => [e, t[n]]);
+					break;
+				}
+				default: throw Error("Unknown statType");
+			}
+			return o[t(u)] = d, o;
+		}, {}),
+		meta: {
+			statType: a.statType,
+			testMethod: a.testMethod,
+			format: a.format,
+			ns: a.ns,
+			significanceLevels: a.significanceLevels
+		}
+	};
 }
-export {
-  B as getStatsData
-};
-//# sourceMappingURL=index.js.map
+export { a as getStatsData };
+
+//# sourceMappingURL=index.js.map