Modernizing the EGI MFF Reader for MNE Python via mffpy Integration

The EGI-MFF reader in MNE-Python uses an old internal parser that’s pretty hard to maintain and often runs into timing drifts or memory issues. My plan for GSoC is to completely modernize this by switching the backend over to the 'mffpy' library. I’m going to build a wrapper that keeps everything backward-compatible while fixing those annoying 1-sample errors and array mismatches that have been blocking users for years. By the end of the summer, I’ll have delivered a refactored read_raw_egi module that supports O(1) memory mapping, so researchers can stream huge files without crashing. I'm also making sure we get proper support for PNS channel mapping and automated bad-channel detection. Most importantly, I’ll be building a solid regression test suite to make sure the event timing stays microsecond-accurate, which is crucial for the neuroscience work this library supports.