Fix Perl vs TypeScript comparison gaps

package.json

@@ -33,6 +33,7 @@
   "dependencies": {
     "@types/react": "^19.2.14",
     "better-sqlite3": "^12.6.2",
+    "chrono-node": "^2.9.0",
     "exceljs": "^4.4.0",
     "fast-xml-parser": "^5.3.7",
     "ink": "^6.8.0",
@@ -47,6 +48,7 @@
     "papaparse": "^5.5.3",
     "pg": "^8.18.0",
     "react": "^19.2.4",
-    "string-width": "^8.2.0"
+    "string-width": "^8.2.0",
+    "woothee": "^1.11.1"
   }
 }

src/aggregators/Ord2Bivariate.ts

@@ -0,0 +1,97 @@
+import type { Aggregator } from "../Aggregator.ts";
+import type { Record } from "../Record.ts";
+import type { JsonValue } from "../types/json.ts";
+import { findKey } from "../KeySpec.ts";
+import { aggregatorRegistry } from "../Aggregator.ts";
+
+// [sum1, sumX, sumY, sumXY, sumX2, sumY2]
+type Ord2BivState = [number, number, number, number, number, number];
+
+/**
+ * Second-order bivariate statistics aggregator.
+ * Computes covariance, correlation, and linear regression parameters
+ * between two fields using a single pass.
+ *
+ * Analogous to App::RecordStream::Aggregator::Ord2Bivariate in Perl.
+ */
+export class Ord2BivariateAggregator implements Aggregator<Ord2BivState | null> {
+  fieldX: string;
+  fieldY: string;
+
+  constructor(fieldX: string, fieldY: string) {
+    this.fieldX = fieldX;
+    this.fieldY = fieldY;
+  }
+
+  initial(): Ord2BivState | null {
+    return null;
+  }
+
+  combine(state: Ord2BivState | null, record: Record): Ord2BivState | null {
+    const vx = findKey(record.dataRef(), this.fieldX, true);
+    const vy = findKey(record.dataRef(), this.fieldY, true);
+    if (vx === undefined || vx === null || vy === undefined || vy === null) return state;
+    const x = Number(vx);
+    const y = Number(vy);
+    const mapped: Ord2BivState = [1, x, y, x * y, x * x, y * y];
+    if (state === null) return mapped;
+    return [
+      state[0] + mapped[0],
+      state[1] + mapped[1],
+      state[2] + mapped[2],
+      state[3] + mapped[3],
+      state[4] + mapped[4],
+      state[5] + mapped[5],
+    ];
+  }
+
+  squish(state: Ord2BivState | null): JsonValue {
+    if (state === null) return null;
+    const [n, sumX, sumY, sumXY, sumX2, sumY2] = state;
+
+    const meanX = sumX / n;
+    const meanY = sumY / n;
+
+    // Covariance: E[XY] - E[X]*E[Y]
+    const covariance = sumXY / n - meanX * meanY;
+
+    // Variances
+    const varX = sumX2 / n - meanX * meanX;
+    const varY = sumY2 / n - meanY * meanY;
+
+    // Correlation: cov / (stdX * stdY)
+    const denominator = Math.sqrt(varX * varY);
+    const correlation = denominator > 0
+      ? (sumXY * n - sumX * sumY) / Math.sqrt((sumX2 * n - sumX ** 2) * (sumY2 * n - sumY ** 2))
+      : null;
+
+    // Linear regression: y = alpha + beta * x
+    const betaDenom = sumX2 * n - sumX ** 2;
+    const beta = betaDenom !== 0 ? (sumXY * n - sumX * sumY) / betaDenom : null;
+    const alpha = beta !== null ? (sumY - beta * sumX) / n : null;
+
+    const result: { [key: string]: JsonValue } = {
+      count: n,
+      covariance,
+      correlation,
+    };
+
+    if (alpha !== null && beta !== null) {
+      result["alpha"] = alpha;
+      result["beta"] = beta;
+    }
+
+    return result;
+  }
+}
+
+aggregatorRegistry.register("ord2biv", {
+  create: (fieldX: string, fieldY: string) => new Ord2BivariateAggregator(fieldX, fieldY),
+  argCounts: [2],
+  shortUsage: "compute second-order bivariate statistics for two fields",
+  longUsage:
+    "Usage: ord2biv,<field1>,<field2>\n" +
+    "   Compute covariance, correlation, and linear regression parameters\n" +
+    "   between two fields.",
+  aliases: ["ord2bivariate"],
+});

src/aggregators/Ord2Univariate.ts

@@ -0,0 +1,82 @@
+import type { Aggregator } from "../Aggregator.ts";
+import type { Record } from "../Record.ts";
+import type { JsonValue } from "../types/json.ts";
+import { findKey } from "../KeySpec.ts";
+import { aggregatorRegistry } from "../Aggregator.ts";
+
+// [count, sumX, sumX2, sumX3, sumX4]
+type Ord2UniState = [number, number, number, number, number];
+
+/**
+ * Second-order univariate statistics aggregator.
+ * Computes count, mean, variance, standard deviation, skewness, and kurtosis
+ * for a single field using a single pass.
+ *
+ * Analogous to App::RecordStream::Aggregator::Ord2Univariate in Perl.
+ */
+export class Ord2UnivariateAggregator implements Aggregator<Ord2UniState | null> {
+  field: string;
+
+  constructor(field: string) {
+    this.field = field;
+  }
+
+  initial(): Ord2UniState | null {
+    return null;
+  }
+
+  combine(state: Ord2UniState | null, record: Record): Ord2UniState | null {
+    const value = findKey(record.dataRef(), this.field, true);
+    if (value === undefined || value === null) return state;
+    const x = Number(value);
+    const mapped: Ord2UniState = [1, x, x * x, x * x * x, x * x * x * x];
+    if (state === null) return mapped;
+    return [
+      state[0] + mapped[0],
+      state[1] + mapped[1],
+      state[2] + mapped[2],
+      state[3] + mapped[3],
+      state[4] + mapped[4],
+    ];
+  }
+
+  squish(state: Ord2UniState | null): JsonValue {
+    if (state === null) return null;
+    const [n, sumX, sumX2, sumX3, sumX4] = state;
+
+    const mean = sumX / n;
+    const variance = sumX2 / n - mean * mean;
+    const stddev = Math.sqrt(variance);
+
+    const result: { [key: string]: JsonValue } = {
+      count: n,
+      mean,
+      variance,
+      stddev,
+    };
+
+    // Skewness and kurtosis require variance > 0
+    if (variance > 0) {
+      // E[(X - mean)^3] = E[X^3] - 3*mean*E[X^2] + 2*mean^3
+      const m3 = sumX3 / n - 3 * mean * (sumX2 / n) + 2 * mean * mean * mean;
+      result["skewness"] = m3 / (stddev * stddev * stddev);
+
+      // E[(X - mean)^4] = E[X^4] - 4*mean*E[X^3] + 6*mean^2*E[X^2] - 3*mean^4
+      const m4 = sumX4 / n - 4 * mean * (sumX3 / n) + 6 * mean * mean * (sumX2 / n) - 3 * mean * mean * mean * mean;
+      result["kurtosis"] = m4 / (variance * variance);
+    }
+
+    return result;
+  }
+}
+
+aggregatorRegistry.register("ord2uni", {
+  create: (field: string) => new Ord2UnivariateAggregator(field),
+  argCounts: [1],
+  shortUsage: "compute second-order univariate statistics for a field",
+  longUsage:
+    "Usage: ord2uni,<field>\n" +
+    "   Compute count, mean, variance, standard deviation, skewness,\n" +
+    "   and kurtosis for the specified field.",
+  aliases: ["ord2univariate"],
+});

src/aggregators/registry.ts

@@ -33,5 +33,7 @@ import "./FirstRecord.ts";
 import "./LastRecord.ts";
 import "./RecordForMaximum.ts";
 import "./RecordForMinimum.ts";
+import "./Ord2Univariate.ts";
+import "./Ord2Bivariate.ts";
 
 export { aggregatorRegistry, makeAggregators } from "../Aggregator.ts";

src/clumpers/Options.ts

@@ -132,30 +132,38 @@ export class ClumperOptions {
       keyParts.push(String(val ?? ""));
     }
 
-    const groupKey = keyParts.join("\x1E");
-
     if (!this.groups) {
       this.groups = new Map();
     }
 
-    let cookie = this.groups.get(groupKey);
-    if (cookie === undefined) {
-      // Handle LRU eviction if keySize is set and NOT in perfect mode
-      if (!this.keyPerfect && this.keySize !== null && this.groups.size >= this.keySize) {
-        const oldestKey = this.groupOrder.shift()!;
-        const oldCookie = this.groups.get(oldestKey);
-        if (oldCookie !== undefined) {
-          this.callback.clumperCallbackEnd(oldCookie);
-          this.groups.delete(oldestKey);
+    // In cube mode, generate all 2^N combinations of actual values and "ALL"
+    const combos = this.keyCube
+      ? this.cubeKeyValues(keySpecs, keyValues)
+      : [{ keyValues, keyParts }];
+
+    for (const combo of combos) {
+      const groupKey = combo.keyParts.join("\x1E");
+
+      let cookie = this.groups.get(groupKey);
+      if (cookie === undefined) {
+        // Handle LRU eviction if keySize is set and NOT in perfect mode
+        if (!this.keyPerfect && this.keySize !== null && this.groups.size >= this.keySize) {
+          const oldestKey = this.groupOrder.shift()!;
+          const oldCookie = this.groups.get(oldestKey);
+          if (oldCookie !== undefined) {
+            this.callback.clumperCallbackEnd(oldCookie);
+            this.groups.delete(oldestKey);
+          }
         }
+
+        cookie = this.callback.clumperCallbackBegin(combo.keyValues);
+        this.groups.set(groupKey, cookie);
+        this.groupOrder.push(groupKey);
       }
 
-      cookie = this.callback.clumperCallbackBegin(keyValues);
-      this.groups.set(groupKey, cookie);
-      this.groupOrder.push(groupKey);
+      this.callback.clumperCallbackPushRecord(cookie, record);
     }
 
-    this.callback.clumperCallbackPushRecord(cookie, record);
     return true;
   }
 
@@ -183,6 +191,40 @@ export class ClumperOptions {
     }
   }
 
+  /**
+   * Generate all 2^N combinations of actual key values and "ALL" for cube mode.
+   */
+  cubeKeyValues(
+    keySpecs: string[],
+    keyValues: { [key: string]: JsonValue }
+  ): Array<{ keyValues: { [key: string]: JsonValue }; keyParts: string[] }> {
+    const n = keySpecs.length;
+    const combos: Array<{ keyValues: { [key: string]: JsonValue }; keyParts: string[] }> = [];
+
+    // Iterate all 2^N bitmasks
+    for (let mask = 0; mask < (1 << n); mask++) {
+      const comboValues: { [key: string]: JsonValue } = {};
+      const comboParts: string[] = [];
+
+      for (let i = 0; i < n; i++) {
+        const spec = keySpecs[i]!;
+        if (mask & (1 << i)) {
+          // Replace this key with "ALL"
+          comboValues[spec] = "ALL";
+          comboParts.push("ALL");
+        } else {
+          // Use actual value
+          comboValues[spec] = keyValues[spec] ?? null;
+          comboParts.push(String(keyValues[spec] ?? ""));
+        }
+      }
+
+      combos.push({ keyValues: comboValues, keyParts: comboParts });
+    }
+
+    return combos;
+  }
+
   getKeySize(): number | null {
     return this.keySize;
   }

src/deaggregators/registry.ts

@@ -0,0 +1,11 @@
+/**
+ * Deaggregator registry - central place for registering and looking up
+ * deaggregators by name.
+ *
+ * Importing this module ensures all deaggregator implementations are registered.
+ */
+
+// Import all deaggregator implementations to trigger their self-registration
+import "./Split.ts";
+import "./Unarray.ts";
+import "./Unhash.ts";