Perf: Optimize string construction in _find_and_replace

[mohasarc]

This PR optimizes _find_and_replace by replacing incremental string concatenation (wm_text += ...) with list accumulation and .join().

The Issue: Using += inside the loop forces quadratic time complexity O(N^2) for string construction because Python creates a new string object for every iteration.

The Solution: I refactored the code to accumulate segments in a list (wm_parts) and join them once at the end. This ensures linear time complexity O(N) for memory allocation without altering the underlying search logic.

This is in line with standard python performance recommendation: https://peps.python.org/pep-0008/#programming-recommendations