Iteration 127: Add 8 benchmark pairs (353 total, re-establishing from real 345 baseline)

Run: https://github.com/githubnext/tsessebe/actions/runs/24511277028

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

Author: github-actions[bot]

benchmarks/pandas/bench_groupby_ngroups.py

@@ -0,0 +1,26 @@
+"""Benchmark: DataFrameGroupBy.ngroups and .groups property access."""
+import json
+import time
+import pandas as pd
+
+ROWS = 10_000
+WARMUP = 5
+ITERATIONS = 50
+
+df = pd.DataFrame({
+    "key": [f"g{i % 100}" for i in range(ROWS)],
+    "val": [i * 1.5 for i in range(ROWS)],
+})
+gbk = df.groupby("key")
+
+for _ in range(WARMUP):
+    gbk.ngroups
+    list(gbk.groups.keys())
+
+t0 = time.perf_counter()
+for _ in range(ITERATIONS):
+    gbk.ngroups
+    list(gbk.groups.keys())
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "groupby_ngroups", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/pandas/bench_index_getloc.py

@@ -0,0 +1,20 @@
+"""Benchmark: Index.get_loc (pandas equivalent)."""
+import json
+import time
+import pandas as pd
+
+SIZE = 10_000
+WARMUP = 5
+ITERATIONS = 30
+
+idx = pd.Index(range(SIZE))
+
+for _ in range(WARMUP):
+    idx.get_loc(5000)
+
+t0 = time.perf_counter()
+for i in range(ITERATIONS):
+    idx.get_loc(i % SIZE)
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "index_getloc", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/pandas/bench_merge_left_on_right_on.py

@@ -0,0 +1,27 @@
+"""Benchmark: merge with left_on/right_on (pandas equivalent)."""
+import json
+import time
+import pandas as pd
+
+ROWS = 20_000
+WARMUP = 3
+ITERATIONS = 10
+
+left = pd.DataFrame({
+    "emp_id": list(range(ROWS)),
+    "salary": [30000 + i * 10 for i in range(ROWS)],
+})
+right = pd.DataFrame({
+    "id": list(range(ROWS // 2)),
+    "dept": [f"dept{i % 10}" for i in range(ROWS // 2)],
+})
+
+for _ in range(WARMUP):
+    pd.merge(left, right, left_on="emp_id", right_on="id")
+
+t0 = time.perf_counter()
+for _ in range(ITERATIONS):
+    pd.merge(left, right, left_on="emp_id", right_on="id")
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "merge_left_on_right_on", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/pandas/bench_multi_index_contains.py

@@ -0,0 +1,22 @@
+"""Benchmark: MultiIndex.__contains__ (pandas equivalent)."""
+import json
+import time
+import pandas as pd
+
+SIZE = 5_000
+WARMUP = 5
+ITERATIONS = 50
+
+arr1 = [f"a{i % 50}" for i in range(SIZE)]
+arr2 = [i % 100 for i in range(SIZE)]
+mi = pd.MultiIndex.from_arrays([arr1, arr2])
+
+for _ in range(WARMUP):
+    ("a0", 0) in mi
+
+t0 = time.perf_counter()
+for i in range(ITERATIONS):
+    (f"a{i % 50}", i % 100) in mi
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "multi_index_contains", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/pandas/bench_multi_index_fromarrays.py

@@ -0,0 +1,21 @@
+"""Benchmark: MultiIndex.from_arrays (pandas equivalent)."""
+import json
+import time
+import pandas as pd
+
+SIZE = 5_000
+WARMUP = 3
+ITERATIONS = 20
+
+arr1 = [f"a{i % 50}" for i in range(SIZE)]
+arr2 = [i % 100 for i in range(SIZE)]
+
+for _ in range(WARMUP):
+    pd.MultiIndex.from_arrays([arr1, arr2])
+
+t0 = time.perf_counter()
+for _ in range(ITERATIONS):
+    pd.MultiIndex.from_arrays([arr1, arr2])
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "multi_index_fromarrays", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/pandas/bench_multi_index_fromproduct.py

@@ -0,0 +1,20 @@
+"""Benchmark: MultiIndex.from_product (pandas equivalent)."""
+import json
+import time
+import pandas as pd
+
+WARMUP = 3
+ITERATIONS = 30
+
+level1 = [f"a{i}" for i in range(50)]
+level2 = list(range(100))
+
+for _ in range(WARMUP):
+    pd.MultiIndex.from_product([level1, level2])
+
+t0 = time.perf_counter()
+for _ in range(ITERATIONS):
+    pd.MultiIndex.from_product([level1, level2])
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "multi_index_fromproduct", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/pandas/bench_series_groupby_apply.py

@@ -0,0 +1,23 @@
+"""Benchmark: SeriesGroupBy.apply (pandas equivalent)."""
+import json
+import time
+import pandas as pd
+import numpy as np
+
+ROWS = 10_000
+WARMUP = 3
+ITERATIONS = 20
+
+data = [i * 0.5 for i in range(ROWS)]
+by = [i % 100 for i in range(ROWS)]
+s = pd.Series(data)
+
+for _ in range(WARMUP):
+    s.groupby(by).apply(lambda g: g)
+
+t0 = time.perf_counter()
+for _ in range(ITERATIONS):
+    s.groupby(by).apply(lambda g: g - g.mean())
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "series_groupby_apply", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/pandas/bench_series_groupby_filter.py

@@ -0,0 +1,22 @@
+"""Benchmark: SeriesGroupBy.filter (pandas equivalent)."""
+import json
+import time
+import pandas as pd
+
+ROWS = 10_000
+WARMUP = 3
+ITERATIONS = 20
+
+data = [i * 1.0 for i in range(ROWS)]
+by = [i % 100 for i in range(ROWS)]
+s = pd.Series(data)
+
+for _ in range(WARMUP):
+    s.groupby(by).filter(lambda g: g.sum() > 1000)
+
+t0 = time.perf_counter()
+for _ in range(ITERATIONS):
+    s.groupby(by).filter(lambda g: g.sum() > 1000)
+total = (time.perf_counter() - t0) * 1000
+
+print(json.dumps({"function": "series_groupby_filter", "mean_ms": total / ITERATIONS, "iterations": ITERATIONS, "total_ms": total}))

benchmarks/tsb/bench_groupby_ngroups.ts

@@ -0,0 +1,30 @@
+/**
+ * Benchmark: DataFrameGroupBy.ngroups and .groupKeys property access.
+ */
+import { DataFrame } from "../../src/index.js";
+
+const ROWS = 10_000;
+const WARMUP = 5;
+const ITERATIONS = 50;
+
+const df = DataFrame.fromColumns({
+  key: Array.from({ length: ROWS }, (_, i) => `g${i % 100}`),
+  val: Array.from({ length: ROWS }, (_, i) => i * 1.5),
+});
+const gbk = df.groupby("key");
+
+for (let i = 0; i < WARMUP; i++) {
+  gbk.ngroups;
+  gbk.groupKeys;
+}
+
+const t0 = performance.now();
+for (let i = 0; i < ITERATIONS; i++) {
+  gbk.ngroups;
+  gbk.groupKeys;
+}
+const total = performance.now() - t0;
+
+console.log(
+  JSON.stringify({ function: "groupby_ngroups", mean_ms: total / ITERATIONS, iterations: ITERATIONS, total_ms: total }),
+);

benchmarks/tsb/bench_index_getloc.ts

@@ -0,0 +1,26 @@
+/**
+ * Benchmark: Index.getLoc — locate positions of a label in an index.
+ */
+import { Index } from "../../src/index.js";
+
+const SIZE = 10_000;
+const WARMUP = 5;
+const ITERATIONS = 30;
+
+// Index with unique labels
+const labels = Array.from({ length: SIZE }, (_, i) => i);
+const idx = new Index(labels);
+
+for (let i = 0; i < WARMUP; i++) {
+  idx.getLoc(5000);
+}
+
+const t0 = performance.now();
+for (let i = 0; i < ITERATIONS; i++) {
+  idx.getLoc(i % SIZE);
+}
+const total = performance.now() - t0;
+
+console.log(
+  JSON.stringify({ function: "index_getloc", mean_ms: total / ITERATIONS, iterations: ITERATIONS, total_ms: total }),
+);