[Autoloop: perf-comparison] Iteration 417: inferFreq benchmark (6 freq types, 200 dates each, 500 iters)

Run: https://github.com/githubnext/tsb/actions/runs/29971651261

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

Author: github-actions[bot]

benchmarks/pandas/bench_infer_freq.py

@@ -0,0 +1,30 @@
+"""Benchmark: pandas.tseries.frequencies.infer_freq — infer frequency from date arrays."""
+import json
+import time
+import pandas as pd
+from pandas.tseries.frequencies import infer_freq
+
+WARMUP = 5
+ITERATIONS = 500
+
+# Build DatetimeIndex arrays for various frequencies
+date_sets = [
+    pd.date_range("2020-01-01", periods=200, freq="ms"),
+    pd.date_range("2020-01-01", periods=200, freq="s"),
+    pd.date_range("2020-01-01", periods=200, freq="min"),
+    pd.date_range("2020-01-01", periods=200, freq="h"),
+    pd.date_range("2020-01-01", periods=200, freq="D"),
+    pd.date_range("2020-01-01", periods=200, freq="W"),
+]
+
+for _ in range(WARMUP):
+    for ds in date_sets:
+        infer_freq(ds)
+
+start = time.perf_counter()
+for _ in range(ITERATIONS):
+    for ds in date_sets:
+        infer_freq(ds)
+total = (time.perf_counter() - start) * 1000
+
+print(json.dumps({"function": "infer_freq", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/tsb/bench_infer_freq.ts

@@ -0,0 +1,56 @@
+/**
+ * Benchmark: inferFreq — infer frequency from an array of regularly-spaced dates.
+ * Outputs JSON: {"function": "infer_freq", "mean_ms": ..., "iterations": ..., "total_ms": ...}
+ */
+import { inferFreq } from "../../src/index.js";
+
+const WARMUP = 5;
+const ITERATIONS = 500;
+
+// Build date arrays for various frequencies
+function makeDates(start: Date, stepMs: number, count: number): Date[] {
+  const out: Date[] = [];
+  let t = start.getTime();
+  for (let i = 0; i < count; i++) {
+    out.push(new Date(t));
+    t += stepMs;
+  }
+  return out;
+}
+
+const MS_DAY = 86_400_000;
+const MS_HOUR = 3_600_000;
+const MS_MIN = 60_000;
+
+const base = new Date("2020-01-01T00:00:00Z");
+const dateSets = [
+  makeDates(base, 1, 200),            // 1ms
+  makeDates(base, 1000, 200),         // 1s
+  makeDates(base, MS_MIN, 200),       // 1min
+  makeDates(base, MS_HOUR, 200),      // 1h
+  makeDates(base, MS_DAY, 200),       // daily
+  makeDates(base, 7 * MS_DAY, 200),   // weekly
+];
+
+for (let i = 0; i < WARMUP; i++) {
+  for (const ds of dateSets) {
+    inferFreq(ds);
+  }
+}
+
+const start = performance.now();
+for (let i = 0; i < ITERATIONS; i++) {
+  for (const ds of dateSets) {
+    inferFreq(ds);
+  }
+}
+const total = performance.now() - start;
+
+console.log(
+  JSON.stringify({
+    function: "infer_freq",
+    mean_ms: total / ITERATIONS,
+    iterations: ITERATIONS,
+    total_ms: total,
+  }),
+);