perf : optimized creating hashset from EventLog in convert_log_to_dataframe

[AThomas314]

Performance Improvements

Optimized the initial dataframe creation phase from event_log

Metric	OLD	NEW	Improvement
Total Allocations	11.77M blocks	3.23M blocks	~72.6% reduction
Total Bytes Allocated	3.00 GB	2.33 GB	~22.6% reduction
Peak Usage (t-gmax)	504 MB	430 MB	~14.8% reduction
Final State (t-end)	205 KB	205 KB	No change
Reduced the time required to create the hashset from ~200ms to ~20ms