Optimize scan_memory_for_exact_value

[BonePolk]

I replaces the custom pure-Python KMPSearch implementation with the built-in bytes.find() method, and not it works much faster about 100x!!!

Builtin function is written in C and use more effective algorithm for byte pattern searching.

Testing with

import time
from PyMemoryEditor.util.search.kmp import KMPSearch


def generate_text(size):
    # Return a random text.
    return "".join([chr(random.randint(ord("A"), ord("Z"))) for letter in range(size)])


def test_search_speed():
    target_value_size = 8
    target_value = generate_text(target_value_size).encode()
    memory_region_data_size = target_value_size*1024*1024
    memory_region_data = generate_text(memory_region_data_size).encode()

    start_time = time.perf_counter()
    found_index = memory_region_data.find(target_value, 0)
    while found_index != -1:
        found_index = memory_region_data.find(target_value, found_index+1)
    builtin_time  = time.perf_counter() - start_time
    print(f"builtin find() time: {builtin_time:.6f} sec")

    start_time = time.perf_counter()
    searcher = KMPSearch(target_value, target_value_size)
    for found_index in searcher.search(memory_region_data, memory_region_data_size):
        pass
    kmp_time = time.perf_counter() - start_time
    print(f"KMP search time: {kmp_time:.6f} sec")
    print(f"{kmp_time / builtin_time:.1f}x faster")
    
    
test_search_speed()

Show results:

builtin find() time: 0.014394 sec
KMP search time: 1.522867 sec
105.8x faster

Some OS Information

OS Information: Windows-10-10.0.26100-SP0 - WindowsPE 64bit
Processor Information: AMD64 | Intel64 Family 6 Model 154 Stepping 4, GenuineIntel

[JeanExtreme002]

Hi, @BonePolk !

Sorry for the delay in getting back to this. I’ve reviewed the changes and they do indeed improve the performance of the search. Everything looks good, so I’ll be merging this PR. Thanks for the improvement! 🙌