perf: performance benchmarking with Criterion and dhat

[AThomas314]

This PR introduces a performance benchmarking and profiling suite for the rust4pm library. By integrating Criterion for statistical time-series benchmarking and dhat for heap allocation profiling, this PR establishes the necessary infrastructure to track performance and prevent regressions as we continue to scale the engine.
Changes

Added Criterion as a dev-dependency for automated benchmarking of I/O and data conversion routines.

Added dhat as a dev-dependency for tracking memory allocations, critical for future memory-constrained optimization work.

Created four benchmarks:
    load_events: Measures time taken for disk-to-log ingestion.
    load_events_mem: Profiles heap usage during log ingestion.
    load_dataframe: Measures time taken to convert EventLog to polars::DataFrame.
    load_dataframe_mem: Profiles memory usage during DataFrame conversion.

How to Run
To execute the benchmarks, run:
cargo bench --features dataframes

Notes
This is the first phase PR, future PRs will include benchmarking of other functions

[aarkue]

Hi @AThomas314,
Thanks for your contributions! 😄
I won't have time to review your PRs next week, but hopefully I can get to them soon afterwards.

Generally, most additions seem to make sense, but I will let you know my exact thoughts in a more detailed review per PR.

One small observation already: The filepaths used in the benchmarks could likely benefit from the get_test_data_path function that already exists in lib.rs to be more robust.

Cheers,
Aaron

[AThomas314]

Hey @aarkue,
Thanks for your feedback!
I've addressed it in the next commit within this PR.
Cheers,
Ashish