adding benchmarking, currently only for http funcs

Author: abebus

tests/test_benchmark_http.py

@@ -0,0 +1,58 @@
+from collections import OrderedDict
+
+import pytest
+
+from w3lib.http import headers_dict_to_raw, headers_raw_to_dict
+
+pytestmark = pytest.mark.benchmark
+
+
+def _header_case_long_headers():
+    dct = OrderedDict(
+        [
+            (b"X-Custom-Header", [b"a" * 1_000]),
+            (b"X-Custom-Header-2", [b"b" * 1_000]),
+        ]
+    )
+    raw = (
+        b"X-Custom-Header: " + b"a" * 1_000 + b"\r\n"
+        b"X-Custom-Header-2: " + b"b" * 1_000
+    )
+    return "long_headers", dct, raw
+
+
+def _header_case_many_unique_headers():
+    dct = OrderedDict(
+        [(f"Header-{i}".encode(), [f"value-{i}".encode()]) for i in range(100)]
+    )
+    raw = b"\r\n".join([f"Header-{i}: value-{i}".encode() for i in range(100)])
+    return "many_unique_headers", dct, raw
+
+
+def _header_case_many_repeated_headers():
+    values = [f"id={i}".encode() for i in range(100)]
+    dct = OrderedDict([(b"Set-Cookie", values)])
+    raw = b"\r\n".join([b"Set-Cookie: " + val for val in values])
+    return "many_repeated_headers", dct, raw
+
+
+header_cases = [
+    _header_case_long_headers(),
+    _header_case_many_unique_headers(),
+    _header_case_many_repeated_headers(),
+]
+
+
+@pytest.mark.parametrize(
+    ("_id", "dct", "raw"),
+    header_cases,
+    ids=[case[0] for case in header_cases],
+)
+class TestBenchmarkHttp:
+    def test_bench_dict_to_raw(self, benchmark, _id, dct, raw):  # noqa: PT019
+        result = benchmark(lambda: headers_dict_to_raw(dct))
+        assert result == raw
+
+    def test_bench_raw_to_dict(self, benchmark, _id, dct, raw):  # noqa: PT019
+        result = benchmark(lambda: headers_raw_to_dict(raw))
+        assert result == dct