import time
import regex
def bench(name, iterations, func):
for _ in range(100):
func()
start = time.perf_counter()
found = 0
for _ in range(iterations):
if func():
found += 1
elapsed = time.perf_counter() - start
per_iter_us = (elapsed * 1_000_000) / iterations
print(f"{name:55} {per_iter_us:>8.2f} µs (found: {found}/{iterations})")
return per_iter_us
def main():
print("=== mrab-regex Comprehensive Benchmark ===\n")
print("--- 1. Text Length Scaling (pattern at start) ---")
base_text = "The quick brown fox jumps over the lazy dog. "
for length in [50, 100, 500, 1000, 5000, 10000]:
text = (base_text * ((length // len(base_text)) + 1))[:length]
re_obj = regex.compile(r"(?:quick){e<=1}", flags=regex.BESTMATCH)
bench(f"'quick' e<=1 in {length} chars", 10_000, lambda: bool(re_obj.search(text)))
print("\n--- 2. Match Position Impact (1000 char text) ---")
text_1000 = "X" * 1000
for pos in [0, 10, 50, 100, 500, 900]:
text = text_1000[:pos] + "quick" + text_1000[pos+5:]
re_obj = regex.compile(r"(?:quick){e<=1}", flags=regex.BESTMATCH)
bench(f"'quick' at position {pos}", 10_000, lambda: bool(re_obj.search(text)))
print("\n--- 3. Edit Distance Scaling ---")
text = "The quikc brown fox jumps over the lazy dog."
for e in range(5):
re_obj = regex.compile(rf"(?:quick){{e<={e}}}", flags=regex.BESTMATCH)
bench(f"'quick' e<={e}", 10_000, lambda: bool(re_obj.search(text)))
print("\n--- 4. Pattern Length Scaling (e<=2) ---")
long_text = "Lorem ipsum dolor sit amet, consectetur adipiscing elit. " * 20
patterns = [
("Lorem", 5),
("consectetur", 11),
("adipiscing elit", 15),
("Lorem ipsum dolor", 17),
("consectetur adipiscing", 22),
]
for pat, length in patterns:
re_obj = regex.compile(rf"(?:{regex.escape(pat)}){{e<=2}}", flags=regex.BESTMATCH)
bench(f"'{pat[:15]}' ({length} chars) e<=2", 10_000, lambda: bool(re_obj.search(long_text)))
print("\n--- 5. DNA/Bioinformatics ---")
dna = ''.join(['ACGT'[i % 4] for i in range(10000)])
for motif_len in [4, 8, 12, 16, 20]:
motif = ''.join(['ACGT'[i % 4] for i in range(motif_len)])
re_obj = regex.compile(rf"(?:{motif}){{e<=2}}", flags=regex.BESTMATCH)
bench(f"DNA motif {motif_len} bp, e<=2, 10kb", 1_000, lambda: bool(re_obj.search(dna)))
print("\n--- 5b. DNA Size Scaling ---")
motif = "ACGTACGT"
for size in [100, 1000, 10000, 100000]:
dna = ''.join(['ACGT'[i % 4] for i in range(size)])
re_obj = regex.compile(rf"(?:{motif}){{e<=2}}", flags=regex.BESTMATCH)
iters = 100 if size > 10000 else 1_000
bench(f"ACGTACGT e<=2 in {size} bp", iters, lambda: bool(re_obj.search(dna)))
print("\n--- 6. No Match (Full Scan) ---")
for size in [100, 500, 1000, 5000]:
text = "X" * size
re_obj = regex.compile(r"(?:quick){e<=1}", flags=regex.BESTMATCH)
iters = 1_000 if size > 1000 else 10_000
bench(f"No match in {size} chars", iters, lambda: bool(re_obj.search(text)))
print("\n--- 7. Alternation Patterns ---")
text = "The quick brown fox jumps over the lazy dog."
alt_patterns = [
(r"(?:quick|slow){e<=1}", "2 alts, short"),
(r"(?:quick|brown|lazy){e<=1}", "3 alts, short"),
(r"(?:the|quick|brown|fox|jumps){e<=1}", "5 alts, short"),
(r"(?:quick|brown|fox|jumps|over|lazy|dog|the|a|an){e<=1}", "10 alts"),
]
for pattern, desc in alt_patterns:
re_obj = regex.compile(pattern, flags=regex.BESTMATCH)
bench(f"{desc}", 10_000, lambda: bool(re_obj.search(text)))
print("\n--- 8. Case Insensitive ---")
text = "THE QUICK BROWN FOX JUMPS OVER THE LAZY DOG."
re_ci = regex.compile(r"(?:quick){e<=1}", flags=regex.BESTMATCH | regex.IGNORECASE)
re_cs = regex.compile(r"(?:QUICK){e<=1}", flags=regex.BESTMATCH)
bench("Case insensitive 'quick' e<=1", 10_000, lambda: bool(re_ci.search(text)))
bench("Case sensitive 'QUICK' e<=1", 10_000, lambda: bool(re_cs.search(text)))
print("\n--- 9. Real-World Patterns ---")
text_email = "Contact us at support@example.com for more information."
re_email = regex.compile(r"(?:support){e<=2}", flags=regex.BESTMATCH)
bench("Email prefix 'support' e<=2", 10_000, lambda: bool(re_email.search(text_email)))
names = "John Smith, Jane Doe, Robert Johnson, Michael Williams, David Brown"
re_name = regex.compile(r"(?:Johnson){e<=2}", flags=regex.BESTMATCH)
bench("Name 'Johnson' e<=2", 10_000, lambda: bool(re_name.search(names)))
address = "123 Main Street, Springfield, IL 62701"
re_addr = regex.compile(r"(?:Springfield){e<=2}", flags=regex.BESTMATCH)
bench("City 'Springfield' e<=2", 10_000, lambda: bool(re_addr.search(address)))
products = "SKU: ABC-12345-XYZ, Price: $99.99, Stock: 150 units"
re_sku = regex.compile(r"(?:ABC-12345){e<=1}", flags=regex.BESTMATCH)
bench("SKU 'ABC-12345' e<=1", 10_000, lambda: bool(re_sku.search(products)))
print("\n--- 10. Unicode/International ---")
text_unicode = "Привет мир! Hello world! 你好世界！"
re_ru = regex.compile(r"(?:Привет){e<=1}", flags=regex.BESTMATCH)
re_zh = regex.compile(r"(?:你好){e<=1}", flags=regex.BESTMATCH)
bench("Russian 'Привет' e<=1", 10_000, lambda: bool(re_ru.search(text_unicode)))
bench("Chinese '你好' e<=1", 10_000, lambda: bool(re_zh.search(text_unicode)))
print("\n--- 11. Exact vs Fuzzy Match ---")
text = "The quick brown fox jumps over the lazy dog."
for e in [0, 1, 2, 3]:
re_obj = regex.compile(rf"(?:quick){{e<={e}}}", flags=regex.BESTMATCH)
bench(f"Exact text, e<={e}", 10_000, lambda: bool(re_obj.search(text)))
print("\n--- 12. Multiple Matches ---")
text_repeat = "cat bat rat cat bat rat cat bat rat " * 10
re_multi = regex.compile(r"(?:cat|bat|rat){e<=1}", flags=regex.BESTMATCH)
bench("Count all matches (cat|bat|rat)", 1_000, lambda: len(re_multi.findall(text_repeat)) > 0)
print("\n=== Benchmark Complete ===")
if __name__ == "__main__":
main()