[Autoloop: perf-comparison] Iteration 414: add categorical_ops benchmarks

Benchmarks catFromCodes, catSortByFreq, catFreqTable, and catCrossTab vs
pandas equivalents (Categorical.from_codes, value_counts, crosstab).
Dataset: 100k rows, 5 categories (TS) vs 3 categories (crossTab).

Run: https://github.com/githubnext/tsb/actions/runs/29834097853

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: github-actions[bot]

benchmarks/pandas/bench_categorical_ops.py

@@ -0,0 +1,60 @@
+"""Benchmark: categorical operations on 100k-element Series
+
+Covers pd.Categorical.from_codes, value_counts (sort by freq), and pd.crosstab.
+"""
+import json
+import time
+import numpy as np
+import pandas as pd
+
+ROWS = 100_000
+WARMUP = 3
+ITERATIONS = 10
+
+# Build a categorical Series from codes + categories
+CATEGORIES = ["alpha", "beta", "gamma", "delta", "epsilon"]
+codes = np.arange(ROWS) % len(CATEGORIES)
+cat_series = pd.Series(pd.Categorical.from_codes(codes, categories=CATEGORIES))
+
+# Build a second categorical for crosstab
+CATEGORIES2 = ["x", "y", "z"]
+codes2 = np.arange(ROWS) % len(CATEGORIES2)
+cat_series2 = pd.Series(pd.Categorical.from_codes(codes2, categories=CATEGORIES2))
+
+# Warm up
+for _ in range(WARMUP):
+    cat_series.value_counts()
+    cat_series.value_counts(sort=True)
+    pd.crosstab(cat_series, cat_series2)
+
+# Measure value_counts (analogous to catFreqTable)
+start = time.perf_counter()
+for _ in range(ITERATIONS):
+    cat_series.value_counts(sort=False)
+freq_table_ms = (time.perf_counter() - start) * 1000 / ITERATIONS
+
+# Measure value_counts sorted by freq (analogous to catSortByFreq)
+start = time.perf_counter()
+for _ in range(ITERATIONS):
+    cat_series.value_counts(sort=True)
+sort_by_freq_ms = (time.perf_counter() - start) * 1000 / ITERATIONS
+
+# Measure crosstab (analogous to catCrossTab)
+start = time.perf_counter()
+for _ in range(ITERATIONS):
+    pd.crosstab(cat_series, cat_series2)
+cross_tab_ms = (time.perf_counter() - start) * 1000 / ITERATIONS
+
+mean_ms = (freq_table_ms + sort_by_freq_ms + cross_tab_ms) / 3
+
+print(json.dumps({
+    "function": "categorical_ops",
+    "mean_ms": mean_ms,
+    "iterations": ITERATIONS,
+    "total_ms": mean_ms * ITERATIONS,
+    "details": {
+        "freqTableMs": freq_table_ms,
+        "sortByFreqMs": sort_by_freq_ms,
+        "crossTabMs": cross_tab_ms,
+    },
+}))

benchmarks/tsb/bench_categorical_ops.ts

@@ -0,0 +1,61 @@
+/**
+ * Benchmark: categorical operations on 100k-element Series
+ *
+ * Covers catFromCodes, catSortByFreq, catFreqTable, and catCrossTab.
+ */
+import { Series, catFromCodes, catSortByFreq, catFreqTable, catCrossTab } from "../../src/index.js";
+
+const ROWS = 100_000;
+const WARMUP = 3;
+const ITERATIONS = 10;
+
+// Build a categorical Series from codes + categories
+const CATEGORIES = ["alpha", "beta", "gamma", "delta", "epsilon"];
+const codes = Int32Array.from({ length: ROWS }, (_, i) => i % CATEGORIES.length);
+
+const catSeries = catFromCodes(Array.from(codes), CATEGORIES);
+
+// Build a second categorical for crossTab
+const CATEGORIES2 = ["x", "y", "z"];
+const codes2 = Int32Array.from({ length: ROWS }, (_, i) => i % CATEGORIES2.length);
+const catSeries2 = catFromCodes(Array.from(codes2), CATEGORIES2);
+
+// Warm up
+for (let i = 0; i < WARMUP; i++) {
+  catSortByFreq(catSeries);
+  catFreqTable(catSeries);
+  catCrossTab(catSeries, catSeries2);
+}
+
+// Measure catSortByFreq
+let start = performance.now();
+for (let i = 0; i < ITERATIONS; i++) {
+  catSortByFreq(catSeries);
+}
+const sortByFreqMs = (performance.now() - start) / ITERATIONS;
+
+// Measure catFreqTable
+start = performance.now();
+for (let i = 0; i < ITERATIONS; i++) {
+  catFreqTable(catSeries);
+}
+const freqTableMs = (performance.now() - start) / ITERATIONS;
+
+// Measure catCrossTab
+start = performance.now();
+for (let i = 0; i < ITERATIONS; i++) {
+  catCrossTab(catSeries, catSeries2);
+}
+const crossTabMs = (performance.now() - start) / ITERATIONS;
+
+const mean_ms = (sortByFreqMs + freqTableMs + crossTabMs) / 3;
+
+console.log(
+  JSON.stringify({
+    function: "categorical_ops",
+    mean_ms,
+    iterations: ITERATIONS,
+    total_ms: mean_ms * ITERATIONS,
+    details: { sortByFreqMs, freqTableMs, crossTabMs },
+  }),
+);