@bicharts/shape-core 0.1.12 → 0.1.14

package/dist/chunk-NGV56PGT.mjs

@@ -1,787 +0,0 @@
-/*! @bicharts/shape-core — Apache-2.0. Bundled reference data: GeoNames (https://www.geonames.org/) CC BY 4.0; US Census/TIGER (public domain). Full text: NOTICE in this package. */
-import {
-  GET_RANDOM,
-  SIMPLE_STRING_HASH,
-  STR
-} from "./chunk-VLMGGCN4.mjs";
-import {
-  detectGeo
-} from "./chunk-YJVU7G63.mjs";
-import {
-  detectOrdinalDomain,
-  isOrdinalFriendlyName,
-  safeDistinctValuesToShip
-} from "./chunk-VHWC3THS.mjs";
-import {
-  detectFormatSignature
-} from "./chunk-ICDJLWEM.mjs";
-import {
-  monthLookupFor,
-  normalizeMonthKey
-} from "./chunk-RZAZ7PZK.mjs";
-
-// src/indexedText.ts
-import Papa from "papaparse";
-var DISCRETE_NUMERIC_MAX_DISTINCT = 12;
-var CONTINUOUS_ABS_DISTINCT = 100;
-var CONTINUOUS_DISTINCT_RATIO = 0.3;
-var CONTIGUOUS_DENSITY = 0.9;
-var ORDINAL_AXIS_MAX_DISTINCT = 120;
-var ORDINAL_YEAR_MIN = 1900;
-var ORDINAL_YEAR_MAX = 2100;
-var ORDINAL_SMALL_MIN = 0;
-var ORDINAL_SMALL_MAX = 20;
-var IDENTIFIER_NAME_TOKENS = /* @__PURE__ */ new Set([
-  "id",
-  "ids",
-  "uuid",
-  "guid",
-  "key",
-  "zip",
-  "zipcode",
-  "postal",
-  "postalcode",
-  "ssn",
-  "code",
-  "codes",
-  "sku",
-  "isbn",
-  "msisdn",
-  "imei"
-]);
-function isIdentifierName(name) {
-  if (!name) return false;
-  const tokens = name.replace(/([a-z0-9])([A-Z])/g, "$1 $2").replace(/[_\-./]+/g, " ").toLowerCase().split(/\s+/).filter((t) => t.length > 0);
-  if (tokens.length === 0) return false;
-  return IDENTIFIER_NAME_TOKENS.has(tokens[tokens.length - 1]);
-}
-var TEMPORAL_YEAR_NAME = /\b(year|yr|fy|fiscal\s*year|calendar\s*year)\b/i;
-var TEMPORAL_PERIOD_RE = /^(?:(?:19|20)\d{2}(?:[-/ ]?(?:q[1-4]|w(?:0?[1-9]|[1-4]\d|5[0-3])|(?:0?[1-9]|1[0-2])))?|q[1-4][-/ ](?:19|20)\d{2}|(?:jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)[a-z]*[-/ ]?(?:19|20)?\d{2})$/i;
-function isYyyymm(v) {
-  if (!Number.isInteger(v) || v < 190001 || v > 210012) return false;
-  const mm = v % 100;
-  return mm >= 1 && mm <= 12;
-}
-function isYyyymmdd(v) {
-  if (!Number.isInteger(v) || v < 19000101 || v > 21001231) return false;
-  const mm = Math.floor(v / 100) % 100, dd = v % 100;
-  return mm >= 1 && mm <= 12 && dd >= 1 && dd <= 31;
-}
-function looksLikeLocalizedMonthPeriod(s, monthMap) {
-  if (!monthMap) return false;
-  if (!/\b(?:19|20)\d{2}\b/.test(s)) return false;
-  const rest = s.replace(/\b(?:19|20)\d{2}\b/g, "").replace(/[-/.,]/g, " ").trim();
-  if (!rest) return false;
-  const key = normalizeMonthKey(rest);
-  return key.length > 0 && monthMap[key] !== void 0;
-}
-function classifyTemporal(args) {
-  const { dataType, name, isMeasure, distinctCount } = args;
-  if (dataType === "DateTime") return true;
-  if (isMeasure) return false;
-  if (dataType === "Integer" || dataType === "Decimal") {
-    const lo = args.minNum, hi = args.maxNum;
-    const bothInt = lo != null && hi != null && Number.isFinite(lo) && Number.isFinite(hi) && Number.isInteger(lo) && Number.isInteger(hi) && hi >= lo;
-    if (!bothInt) return false;
-    const inYearRange = lo >= 1900 && hi <= 2100;
-    if (inYearRange && distinctCount >= 2 && TEMPORAL_YEAR_NAME.test(name)) return true;
-    if (inYearRange && distinctCount >= 3 && distinctCount === hi - lo + 1) return true;
-    const endpointsYyyymm = isYyyymm(lo) && isYyyymm(hi);
-    const endpointsYyyymmdd = isYyyymmdd(lo) && isYyyymmdd(hi);
-    if (endpointsYyyymm || endpointsYyyymmdd) {
-      if (!args.sampleValues) return true;
-      let n = 0, ok = 0;
-      for (const v of args.sampleValues) {
-        if (v == null || v === "") continue;
-        n++;
-        const num = Number(String(v).trim());
-        if (Number.isFinite(num) && (endpointsYyyymm ? isYyyymm(num) : isYyyymmdd(num))) ok++;
-        if (n >= 60) break;
-      }
-      if (n >= 2 && ok === n) return true;
-    }
-    return false;
-  }
-  if (args.sampleValues) {
-    const monthMap = args.locale ? monthLookupFor(args.locale) : null;
-    let n = 0, hit = 0;
-    for (const v of args.sampleValues) {
-      if (v == null || v === "") continue;
-      n++;
-      const s = String(v).trim();
-      if (TEMPORAL_PERIOD_RE.test(s) || looksLikeLocalizedMonthPeriod(s, monthMap)) hit++;
-      if (n >= 60) break;
-    }
-    if (n >= 2 && hit / n >= 0.8) return true;
-  }
-  return false;
-}
-function hostAggHint(name) {
-  const n = (name || "").trim().toLowerCase();
-  if (n.startsWith("sum ") || n.startsWith("total ")) return "sum";
-  if (n.startsWith("average ") || n.startsWith("avg. ")) return "avg";
-  if (n.startsWith("count ") || n.startsWith("distinct count ") || n.startsWith("count (distinct) ")) return "count";
-  if (n.startsWith("min ") || n.startsWith("minimum ")) return "min";
-  if (n.startsWith("max ") || n.startsWith("maximum ")) return "max";
-  return null;
-}
-function classifyAdditivity(args) {
-  const { rows, measureColIdx, measureName, candidateDims, minNum, maxNum, discourageAgg } = args;
-  const SERIES_CARD_CAP = 100;
-  const TOL_FRAC = 0.01;
-  const SEP = "
";
-  for (const c of candidateDims) {
-    if (c.distinct < 2 || c.distinct > SERIES_CARD_CAP) continue;
-    const otherDims = candidateDims.filter((d) => d.idx !== c.idx).map((d) => d.idx);
-    const groups = /* @__PURE__ */ new Map();
-    for (const row of rows) {
-      const mv = row[measureColIdx];
-      if (typeof mv !== "number" || !Number.isFinite(mv)) continue;
-      const key = otherDims.length ? otherDims.map((di) => String(row[di])).join(SEP) : "";
-      let g = groups.get(key);
-      if (!g) {
-        g = { sum: 0, cset: /* @__PURE__ */ new Set() };
-        groups.set(key, g);
-      }
-      g.sum += mv;
-      g.cset.add(String(row[c.idx]));
-    }
-    let n100 = 0, n1 = 0, tot = 0;
-    for (const g of groups.values()) {
-      if (g.cset.size < 2) continue;
-      tot++;
-      if (Math.abs(g.sum - 100) <= 100 * TOL_FRAC) n100++;
-      else if (Math.abs(g.sum - 1) <= 1 * TOL_FRAC) n1++;
-    }
-    if (tot >= 2) {
-      if (n100 / tot >= 0.9 || n1 / tot >= 0.9) {
-        return { additivity: "part_of_whole", partOfWholeDim: c.name };
-      }
-    }
-  }
-  if (discourageAgg) return { additivity: "intensive_rate" };
-  const agg = hostAggHint(measureName);
-  if (agg === "avg" || agg === "min" || agg === "max") return { additivity: "intensive_rate" };
-  if (agg === "sum" || agg === "count") return { additivity: "additive" };
-  if (minNum != null && maxNum != null && Number.isFinite(minNum) && Number.isFinite(maxNum) && minNum >= 0 && maxNum <= 1.0001 && maxNum > 0) {
-    return { additivity: "intensive_rate" };
-  }
-  return { additivity: "unknown" };
-}
-function classifyNumericValueNature(i) {
-  const { dataType, isMeasure, name, distinct, nonblank, prec, maxval, minval } = i;
-  const span = maxval - minval + 1;
-  const density = span > 0 ? distinct / span : 1;
-  const ratio = nonblank > 0 ? distinct / nonblank : 0;
-  const fractional = dataType === "Decimal" || prec > 0;
-  if (isIdentifierName(name)) return "Categorical";
-  if (density >= CONTIGUOUS_DENSITY) {
-    if (isOrdinalFriendlyName(name) && distinct <= ORDINAL_AXIS_MAX_DISTINCT) return "Ordinal";
-    if (minval >= ORDINAL_YEAR_MIN && maxval <= ORDINAL_YEAR_MAX) return "Ordinal";
-    if (minval >= ORDINAL_SMALL_MIN && maxval <= ORDINAL_SMALL_MAX) return "Ordinal";
-  }
-  if (isMeasure) return "Continuous";
-  if (distinct > DISCRETE_NUMERIC_MAX_DISTINCT && (fractional || distinct >= CONTINUOUS_ABS_DISTINCT || ratio >= CONTINUOUS_DISTINCT_RATIO)) {
-    return "Continuous";
-  }
-  return "Categorical";
-}
-var IndexedText = class {
-  _computedStatsForLevel = null;
-  _cols = [];
-  _rows = [];
-  _origIndices = [];
-  _rowHashes = /* @__PURE__ */ new Set();
-  STR(v) {
-    if (v === null || v === void 0) {
-      return "";
-    }
-    return v.toString();
-  }
-  getPrecision(value) {
-    if (!isFinite(value)) return 0;
-    const valueStr = value.toString();
-    const decimalPart = valueStr.split(".")[1];
-    return decimalPart ? decimalPart.length : 0;
-  }
-  getColumns() {
-    return this._cols;
-  }
-  hasTimeComponent(date) {
-    if (!(date instanceof Date) || isNaN(date.getTime())) {
-      return false;
-    }
-    return date.getHours() !== 0 || date.getMinutes() !== 0 || date.getSeconds() !== 0 || date.getMilliseconds() !== 0;
-  }
-  getColumnsWithStats(privacyLevel, locale) {
-    if (this._computedStatsForLevel == privacyLevel) {
-      return this._cols;
-    }
-    let cidx = 0;
-    const pl = parseInt(privacyLevel);
-    const colValueSets = [];
-    for (const col of this._cols) {
-      const vals = /* @__PURE__ */ new Map();
-      const arr = [];
-      let datalen = 0;
-      let nonblank = 0;
-      let prec = 0;
-      let minval = null;
-      let maxval = null;
-      let sumval = null;
-      let hastime = false;
-      let monoAsc = true;
-      let monoDesc = true;
-      let prevForOrder = null;
-      let sawPrev = false;
-      for (const row of this._rows) {
-        const val = row[cidx];
-        const strVal = this.STR(val);
-        if (strVal != "") {
-          datalen += strVal.length;
-          vals.set(strVal, (vals.get(strVal) || 0) + 1);
-          nonblank++;
-          if (col.dataType == "Integer" || col.dataType == "Decimal" || col.dataType == "DateTime") {
-            if (minval === null || val < minval) {
-              minval = val;
-            }
-            if (maxval === null || val > maxval) {
-              maxval = val;
-            }
-            if (col.dataType == "Integer" || col.dataType == "Decimal") {
-              if (sumval === null) {
-                sumval = val;
-              } else {
-                sumval += val;
-              }
-              if (col.dataType == "Decimal") {
-                const dp = this.getPrecision(val);
-                if (dp > prec) {
-                  prec = dp;
-                }
-              }
-              arr.push(val);
-            } else {
-              hastime = hastime || this.hasTimeComponent(val);
-            }
-          }
-          if (sawPrev) {
-            if (val < prevForOrder) monoAsc = false;
-            else if (val > prevForOrder) monoDesc = false;
-          }
-          prevForOrder = val;
-          sawPrev = true;
-        }
-      }
-      if (nonblank >= 3) {
-        if (monoAsc && !monoDesc) col.sortedDirection = "asc";
-        else if (monoDesc && !monoAsc) col.sortedDirection = "desc";
-        else col.sortedDirection = "none";
-      }
-      if (nonblank > 0 && vals.size > 0 && !col.isMeasure) {
-        let topCount = 0;
-        let minCount = Number.MAX_SAFE_INTEGER;
-        for (const v of vals.values()) {
-          if (v > topCount) topCount = v;
-          if (v < minCount) minCount = v;
-        }
-        col.modalShare = topCount / nonblank;
-        col.minGroupCount = minCount === Number.MAX_SAFE_INTEGER ? null : minCount;
-        if (vals.size === 2) {
-          const ks = [...vals.keys()].map((k) => STR(k).trim().toLowerCase()).sort();
-          const boolPairs = [["false", "true"], ["no", "yes"], ["n", "y"], ["0", "1"], ["f", "t"]];
-          const valuesBoolean = boolPairs.some((p) => ks[0] === p[0] && ks[1] === p[1]);
-          const nameOutcome = /^(is|has)[_a-z0-9]|(?:default|churn|fraud|approv|convert|active|cancel|delinquen|flag|paid|win|pass|fail)/i.test(col.name);
-          if (valuesBoolean || nameOutcome) col.isBinaryFlag = true;
-        }
-      }
-      if (col.dataType == "DateTime") {
-        col.dateWithTime = hastime;
-      }
-      col.isTemporal = classifyTemporal({
-        dataType: col.dataType,
-        name: col.name,
-        isMeasure: col.isMeasure,
-        distinctCount: vals.size,
-        minNum: typeof minval === "number" ? minval : null,
-        maxNum: typeof maxval === "number" ? maxval : null,
-        sampleValues: vals.keys(),
-        locale
-      });
-      this.updateColumnStats10(pl, col, nonblank, vals, prec, maxval, minval);
-      this.updateColumnStats20(pl, arr, nonblank, sumval, col, datalen, hastime, minval, maxval, prec, vals, locale);
-      if (!col.isMeasure && (col.dataType === "Integer" || col.dataType === "Decimal")) {
-        const nature = classifyNumericValueNature({
-          dataType: col.dataType,
-          isMeasure: false,
-          name: col.name,
-          distinct: vals.size,
-          nonblank,
-          prec,
-          maxval: typeof maxval === "number" ? maxval : 0,
-          minval: typeof minval === "number" ? minval : 0
-        });
-        if (nature === "Continuous") col.isMeasure = true;
-      }
-      colValueSets[cidx] = col.isMeasure || vals.size > 2e3 ? null : new Set(vals.keys());
-      cidx++;
-    }
-    const dimCols = [];
-    this._cols.forEach((c, i) => {
-      if (!c.isMeasure) dimCols.push({ idx: i, distinct: c.distinctCount ?? 0, name: c.name });
-    });
-    this._cols.forEach((c, i) => {
-      if (!c.isMeasure) return;
-      const lo = typeof c.lowValue === "number" ? c.lowValue : null;
-      const hi = typeof c.highValue === "number" ? c.highValue : null;
-      const r = classifyAdditivity({
-        rows: this._rows,
-        measureColIdx: i,
-        measureName: c.name,
-        candidateDims: dimCols,
-        minNum: lo,
-        maxNum: hi,
-        discourageAgg: !!c.discourageAggregationAcrossGroups
-      });
-      c.additivity = r.additivity;
-      if (r.partOfWholeDim) c.partOfWholeDim = r.partOfWholeDim;
-    });
-    const DISCRIM_DIM_CAP = 50;
-    const discrimDims = dimCols.filter((d) => d.distinct >= 2 && d.distinct <= DISCRIM_DIM_CAP);
-    this._cols.forEach((c, mi) => {
-      if (!c.isMeasure) return;
-      const xs = [];
-      for (const row of this._rows) {
-        const v = row[mi];
-        if (v === null || v === void 0 || v === "") continue;
-        const n = typeof v === "number" ? v : parseFloat(v);
-        if (!isNaN(n)) xs.push(n);
-      }
-      if (xs.length === 0) return;
-      const sorted = [...xs].sort((a, b) => a - b);
-      const q = (p) => sorted[Math.min(sorted.length - 1, Math.max(0, Math.floor(p * (sorted.length - 1))))];
-      const median = q(0.5), p10 = q(0.1), p90 = q(0.9);
-      const denom = Math.abs(median) > 1e-9 ? Math.abs(median) : Math.abs(sorted[sorted.length - 1]) > 1e-9 ? Math.abs(sorted[sorted.length - 1]) : 1;
-      c.relativeDispersion = Math.round((p90 - p10) / denom * 1e3) / 1e3;
-      let grand = 0;
-      for (const x of xs) grand += x;
-      grand /= xs.length;
-      let ssTotal = 0;
-      for (const x of xs) ssTotal += (x - grand) * (x - grand);
-      if (ssTotal <= 1e-12) return;
-      const out = [];
-      for (const d of discrimDims) {
-        const groups = /* @__PURE__ */ new Map();
-        for (const row of this._rows) {
-          const mv = row[mi];
-          if (mv === null || mv === void 0 || mv === "") continue;
-          const n = typeof mv === "number" ? mv : parseFloat(mv);
-          if (isNaN(n)) continue;
-          const key = this.STR(row[d.idx]) + "";
-          const g = groups.get(key);
-          if (g) {
-            g.s += n;
-            g.n++;
-          } else groups.set(key, { s: n, n: 1 });
-        }
-        let ssBetween = 0;
-        for (const g of groups.values()) {
-          const gm = g.s / g.n;
-          ssBetween += g.n * (gm - grand) * (gm - grand);
-        }
-        out.push({ otherColumn: d.name, eta2: Math.round(ssBetween / ssTotal * 1e4) / 1e4 });
-      }
-      if (out.length > 0) c.groupDiscrimination = out;
-    });
-    const COLLINEAR_R = 0.97;
-    const MIN_CORR_ROWS = 5;
-    const measureIdx = this._cols.map((c, i) => ({ c, i })).filter((x) => x.c.isMeasure);
-    for (let a = 0; a < measureIdx.length; a++) {
-      for (let b = a + 1; b < measureIdx.length; b++) {
-        const ia = measureIdx[a].i, ib = measureIdx[b].i;
-        let n = 0, sx = 0, sy = 0, sxx = 0, syy = 0, sxy = 0;
-        for (const row of this._rows) {
-          const va = row[ia], vb = row[ib];
-          if (va === null || va === void 0 || va === "" || vb === null || vb === void 0 || vb === "") continue;
-          const x = typeof va === "number" ? va : parseFloat(va);
-          const y = typeof vb === "number" ? vb : parseFloat(vb);
-          if (isNaN(x) || isNaN(y)) continue;
-          n++;
-          sx += x;
-          sy += y;
-          sxx += x * x;
-          syy += y * y;
-          sxy += x * y;
-        }
-        if (n < MIN_CORR_ROWS) continue;
-        const vx = n * sxx - sx * sx, vy = n * syy - sy * sy;
-        if (vx <= 1e-12 || vy <= 1e-12) continue;
-        const r = (n * sxy - sx * sy) / Math.sqrt(vx * vy);
-        if (Math.abs(r) >= COLLINEAR_R) {
-          const ca = measureIdx[a].c, cb = measureIdx[b].c;
-          if (!ca.collinearWithMeasures) ca.collinearWithMeasures = [];
-          if (!cb.collinearWithMeasures) cb.collinearWithMeasures = [];
-          ca.collinearWithMeasures.push(cb.name);
-          cb.collinearWithMeasures.push(ca.name);
-        }
-      }
-    }
-    for (let a = 0; a < this._cols.length; a++) {
-      const setA = colValueSets[a];
-      if (!setA || setA.size === 0) continue;
-      const ratios = [];
-      for (let b = 0; b < this._cols.length; b++) {
-        if (b === a) continue;
-        const setB = colValueSets[b];
-        if (!setB || setB.size === 0) continue;
-        let shared = 0;
-        for (const v of setA) if (setB.has(v)) shared++;
-        const pct = Math.round(shared / setA.size * 100);
-        if (pct > 0) ratios.push({ otherColumn: this._cols[b].name, percentShared: pct });
-      }
-      if (ratios.length > 0) this._cols[a].sharedValueRatioWithOthers = ratios;
-    }
-    const PAIR_CAP = 50;
-    const candIdx = [];
-    for (let i = 0; i < this._cols.length; i++) {
-      const s = colValueSets[i];
-      if (s && s.size >= 2 && s.size <= PAIR_CAP) candIdx.push(i);
-    }
-    if (candIdx.length >= 2) {
-      const pairCounts = /* @__PURE__ */ new Map();
-      const pk = (a, b) => a + ":" + b;
-      for (let x = 0; x < candIdx.length; x++)
-        for (let y = x + 1; y < candIdx.length; y++)
-          pairCounts.set(pk(candIdx[x], candIdx[y]), /* @__PURE__ */ new Map());
-      for (const row of this._rows) {
-        for (let x = 0; x < candIdx.length; x++) {
-          const va = this.STR(row[candIdx[x]]) + "
";
-          for (let y = x + 1; y < candIdx.length; y++) {
-            const vb = String.fromCharCode(1) + this.STR(row[candIdx[y]]);
-            const cm = pairCounts.get(pk(candIdx[x], candIdx[y]));
-            const ck = va + "
" + vb;
-            cm.set(ck, (cm.get(ck) ?? 0) + 1);
-          }
-        }
-      }
-      const parentCand = /* @__PURE__ */ new Map();
-      for (let x = 0; x < candIdx.length; x++) {
-        for (let y = x + 1; y < candIdx.length; y++) {
-          const a = candIdx[x], b = candIdx[y];
-          const cardA = colValueSets[a].size, cardB = colValueSets[b].size;
-          const cellMap = pairCounts.get(pk(a, b));
-          const distinctPairs = cellMap.size;
-          let minCellCount = 0;
-          for (const cval of cellMap.values()) {
-            if (minCellCount === 0 || cval < minCellCount) minCellCount = cval;
-          }
-          const fillPct = Math.round(distinctPairs / (cardA * cardB) * 100);
-          const aDet = distinctPairs === cardA;
-          const bDet = distinctPairs === cardB;
-          (this._cols[a].categoricalPairStats ||= []).push(
-            { otherColumn: this._cols[b].name, fillPct, determinesOther: aDet, minCellCount }
-          );
-          (this._cols[b].categoricalPairStats ||= []).push(
-            { otherColumn: this._cols[a].name, fillPct, determinesOther: bDet, minCellCount }
-          );
-          if (aDet && cardA > cardB) {
-            const cur = parentCand.get(a);
-            if (!cur || cardB < cur.card)
-              parentCand.set(a, { name: this._cols[b].name, card: cardB, ratio: cardB / cardA });
-          }
-          if (bDet && cardB > cardA) {
-            const cur = parentCand.get(b);
-            if (!cur || cardA < cur.card)
-              parentCand.set(b, { name: this._cols[a].name, card: cardA, ratio: cardA / cardB });
-          }
-        }
-      }
-      for (const [i, p] of parentCand) {
-        this._cols[i].primaryParentColumn = p.name;
-        this._cols[i].nestingRatio = Math.round(p.ratio * 100) / 100;
-      }
-    }
-    this._computedStatsForLevel = privacyLevel;
-    return this._cols;
-  }
-  updateColumnStats20(pl, arr, nonblank, sumval, col, datalen, hastime, minval, maxval, prec, vals, locale) {
-    if (pl >= 20) {
-      const sorted = [...arr].sort((a, b) => a - b);
-      const mid = Math.floor(sorted.length / 2);
-      if (nonblank > 0 && sumval != null) {
-        col.avgValue = col.dataType == "Integer" ? Math.round(sumval / nonblank) : sumval / nonblank;
-      } else {
-        col.avgValue = null;
-      }
-      if (nonblank > 0) {
-        col.avgLength = datalen / nonblank;
-      }
-      if (this._rows.length > 0) {
-        if (col.dataType != "DateTime" || hastime || !minval) {
-          col.lowValue = minval;
-        } else {
-          col.lowValue = new Date(minval).toISOString().slice(0, 10) + "T00:00:00.000Z";
-        }
-        if (col.dataType != "DateTime" || hastime || !maxval) {
-          col.highValue = maxval;
-        } else {
-          col.highValue = new Date(maxval).toISOString().slice(0, 10) + "T00:00:00.000Z";
-        }
-        col.medianValue = sorted.length === 0 ? null : sorted.length % 2 === 0 ? col.dataType == "Integer" ? Math.round((sorted[mid - 1] + sorted[mid]) / 2) : (sorted[mid - 1] + sorted[mid]) / 2 : sorted[mid];
-        col.numericPrecision = prec;
-      }
-      if (sorted.length > 0) {
-        const mean = col.avgValue ?? 0;
-        const n = sorted.length;
-        let m2 = 0, m3 = 0;
-        for (const v of sorted) {
-          const d = v - mean;
-          m2 += d * d;
-          m3 += d * d * d;
-        }
-        const variance = m2 / n;
-        col.skewness = variance === 0 ? 0 : m3 / n / Math.pow(Math.sqrt(variance), 3);
-        col.stdDev = Math.sqrt(variance);
-      }
-      if (sumval !== null) {
-        col.sum = sumval;
-      }
-      const topcatEntries = Array.from(vals.entries()).sort((a, b) => b[1] - a[1]);
-      col.topCategories = topcatEntries.slice(0, 10).map((e) => e[1]);
-      if (!col.isMeasure && topcatEntries.length > 0) {
-        if (col.dataType === "String") {
-          const allDistinct = topcatEntries.map((e) => e[0]);
-          col.formatSignature = detectFormatSignature(allDistinct);
-          const nonWordFmt = /^(UUID|EMAIL_LIKE|URL_LIKE|PHONE_LIKE|NUMERIC_ID|HEX_ID|OPAQUE_ID_ALPHANUMERIC|DATE_STR|BOOLEAN)$/;
-          if (col.formatSignature && !nonWordFmt.test(col.formatSignature) && (col.avgLength === void 0 || col.avgLength >= 3))
-            col.isFreeText = true;
-        } else {
-          col.topCategoryValues = topcatEntries.slice(0, 5).map((e) => e[0]);
-        }
-        if (pl >= 20) {
-          const allDistinctRaw = topcatEntries.map((e) => e[0]);
-          const detection = detectOrdinalDomain(allDistinctRaw, locale);
-          if (detection !== null) {
-            col.ordinalPattern = detection.pattern;
-            col.orderedDomain = detection.orderedDomain;
-          } else if (col.dataType === "String") {
-            const safe = safeDistinctValuesToShip(allDistinctRaw);
-            if (safe !== null) {
-              col.safeDistinctValues = safe;
-            }
-          }
-        }
-        const geo = detectGeo(topcatEntries.map((e) => e[0]), col.name, locale);
-        if (geo !== null) {
-          col.geoKind = geo.geoKind;
-          col.geoMatchPct = geo.geoMatchPct;
-          if (geo.geoAmbiguous) col.geoAmbiguous = true;
-        }
-      }
-      if (sorted.length >= 8 && (col.dataType === "Integer" || col.dataType === "Decimal")) {
-        const pick = (p) => sorted[Math.floor((sorted.length - 1) * p)];
-        col.quantiles = {
-          p05: pick(0.05),
-          p25: pick(0.25),
-          p75: pick(0.75),
-          p95: pick(0.95)
-        };
-      }
-    }
-  }
-  updateColumnStats10(pl, col, nonblank, vals, prec, maxval, minval) {
-    if (pl >= 10) {
-      col.blankCount = this._rows.length - nonblank;
-      col.distinctCount = vals.size + (col.blankCount != 0 ? 1 : 0);
-      col.valueNature = "Categorical";
-      if (nonblank > 0) {
-        if (col.dataType === "Integer" || col.dataType === "Decimal") {
-          col.valueNature = classifyNumericValueNature({
-            dataType: col.dataType,
-            isMeasure: !!col.isMeasure,
-            name: col.name,
-            distinct: vals.size,
-            nonblank,
-            prec,
-            maxval,
-            minval
-          });
-        } else if (col.dataType === "DateTime") {
-          col.valueNature = col.dateWithTime ? "Continuous" : "Ordinal";
-        }
-      }
-    }
-  }
-  getRowCount() {
-    return this._rows.length;
-  }
-  // LEAF CARDINALITY (2026-07-13) — distinct count of the tuple of NON-MEASURE
-  // (grouping) columns: the finest grain at which the data would be counted. The
-  // server only has per-column marginals, never this JOINT distinctness — so the
-  // client measures it here and ships it, letting AggregationAnalyzer decide whether
-  // an implicit row COUNT is meaningful (rows repeat -> RowCount > leaf) or degenerate
-  // (each row a unique tuple -> leaf ~ RowCount, a count is all-1s). One pass over the
-  // accumulated rows; cached until reset().
-  getLeafCardinality() {
-    if (this._leafCardinality !== null) return this._leafCardinality;
-    const dimIdx = [];
-    this._cols.forEach((c, i) => {
-      if (!c.isMeasure) dimIdx.push(i);
-    });
-    if (dimIdx.length === 0) {
-      this._leafCardinality = this._rows.length;
-      return this._leafCardinality;
-    }
-    const seen = /* @__PURE__ */ new Set();
-    for (const row of this._rows) {
-      seen.add(JSON.stringify(dimIdx.map((di) => row[di])));
-    }
-    this._leafCardinality = seen.size;
-    return this._leafCardinality;
-  }
-  _leafCardinality = null;
-  obfuscateString(input) {
-    const lowers = "abcdefghijklmnopqrstuvwxyz";
-    const getRandomLower = () => lowers[Math.floor(GET_RANDOM() * lowers.length)];
-    const getRandomUpper = () => getRandomLower().toUpperCase();
-    const getRandomDigit = () => Math.floor(GET_RANDOM() * 10).toString();
-    const substituteSameClass = (origChar) => {
-      if (/[A-Z]/.test(origChar)) return getRandomUpper();
-      if (/[a-z]/.test(origChar)) return getRandomLower();
-      if (/[0-9]/.test(origChar)) return getRandomDigit();
-      return origChar;
-    };
-    const isEmail = input.includes("@") && input.includes(".");
-    const emailParts = isEmail ? input.split("@") : [input];
-    const domain = isEmail ? emailParts[1] : "";
-    const preserveSuffix = domain.endsWith(".com");
-    const suffix = preserveSuffix ? ".com" : "";
-    const domainName = preserveSuffix ? domain.slice(0, -4) : domain;
-    const obfuscatePart = (str) => {
-      let result = "";
-      for (const char of str) {
-        result += substituteSameClass(char);
-        if (/[a-zA-Z0-9]/.test(char) && GET_RANDOM() < 0.2) {
-          result += substituteSameClass(char);
-        }
-      }
-      if (result.length > 6) {
-        const numToRemove = Math.floor(GET_RANDOM() * Math.min(3, result.length - 6));
-        for (let i = 0; i < numToRemove; i++) {
-          let attempts = 0;
-          while (attempts < 5) {
-            const removeIndex = Math.floor(GET_RANDOM() * result.length);
-            if (/[a-zA-Z0-9]/.test(result[removeIndex])) {
-              result = result.slice(0, removeIndex) + result.slice(removeIndex + 1);
-              break;
-            }
-            attempts++;
-          }
-        }
-      }
-      return result;
-    };
-    if (isEmail) {
-      const localPart = obfuscatePart(emailParts[0]);
-      const obfuscatedDomain = obfuscatePart(domainName);
-      return `${localPart}@${obfuscatedDomain}${suffix}`;
-    } else {
-      return obfuscatePart(input);
-    }
-  }
-  toObjectArray() {
-    return this._rows.map((row) => {
-      const obj = {};
-      this._cols.forEach((col, index) => {
-        obj[col.name] = row[index];
-      });
-      return obj;
-    });
-  }
-  async getCSVAsync(forSample) {
-    const rows = [];
-    let cnt = forSample ? 25 : 1e7;
-    let len = forSample ? 4e3 : 1e7;
-    const strMap = /* @__PURE__ */ new Map();
-    for (const r of this._rows) {
-      const obj = {};
-      for (let c = 0; c < this._cols.length; c++) {
-        const col = this._cols[c];
-        let v = r[c];
-        if (this.STR(v) != "") {
-          if (forSample && col.dataType == "String") {
-            if (strMap.has(v)) {
-              v = strMap.get(v);
-            } else {
-              const nv = this.obfuscateString(v);
-              strMap.set(v, nv);
-              v = nv;
-            }
-          } else {
-            if (col.dataType == "DateTime" && !col.dateWithTime) {
-              v = new Date(v).toISOString().slice(0, 10) + "T00:00:00.000Z";
-            }
-          }
-          len -= this.STR(v).length;
-        } else {
-          v = null;
-        }
-        obj[c] = v;
-        --len;
-      }
-      rows.push(obj);
-      --cnt;
-      if (cnt <= 0 || len <= 0) {
-        break;
-      }
-    }
-    const csv = await Papa.unparse(rows, { header: false });
-    return csv;
-  }
-  // Column-name header line matching getCSVAsync's column ORDER (this._cols, by index).
-  // getCSVAsync emits a HEADERLESS body; legacy cached code reads CSV_STRING with a bare
-  // pd.read_csv (no names), so the host prepends this line at injection time — otherwise
-  // pandas uses row 0 as the header and every column reference KeyErrors (v1->v2 upgrade
-  // running cached v1 Python code). Built via Papa so escaping matches the data rows.
-  getCSVHeaderLine() {
-    return Papa.unparse([this._cols.map((c) => this.STR(c.name))]);
-  }
-  addRow(colvals, originalIdx) {
-    let str = "";
-    for (const v of colvals) {
-      str += this.STR(v) + "~";
-    }
-    const hash = SIMPLE_STRING_HASH(str);
-    if (this._rowHashes.has(hash)) {
-      return false;
-    }
-    this._rows.push(colvals);
-    this._origIndices.push(originalIdx === void 0 || originalIdx === null ? -1 : originalIdx);
-    this._rowHashes.add(hash);
-    return true;
-  }
-  getOriginalRowIndex(i) {
-    if (i < 0 || i >= this._origIndices.length) return -1;
-    return this._origIndices[i];
-  }
-  reset() {
-    this._cols = [];
-    this._rows = [];
-    this._origIndices = [];
-    this._rowHashes = /* @__PURE__ */ new Set();
-    this._leafCardinality = null;
-  }
-  setColumns(cols) {
-    this._cols = cols;
-  }
-};
-
-export {
-  isIdentifierName,
-  classifyTemporal,
-  hostAggHint,
-  classifyAdditivity,
-  classifyNumericValueNature,
-  IndexedText
-};
-//# sourceMappingURL=chunk-NGV56PGT.mjs.map