Split large rec files between multiple SpikegadgetsRawIO iterators

[samuelbray32]

Excessively large iterators can cause memory issues due to lading large chunks of data into memory during get_analogsignal_multiplexed or get_digitalsignal.

This merge splits the memmap of these large rec files into multiple SpikegadgetsRawIO iterators. The ordering of these is tracked by the RecDataChunkIterator so almost any use case doesn't need to worry about this.

If trying too get info for a specific epoch, note that this means there is not a 1:1 relation between iterators and epochs. Use the file_info pandas dataframe to determine which epoch each belongs to (see add_epochs for example)

[codecov]

Codecov Report

Attention: 100 lines in your changes are missing coverage. Please review.

Comparison is base (cf7561e) 94.39% compared to head (d859f90) 90.10%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #50      +/-   ##
==========================================
- Coverage   94.39%   90.10%   -4.29%     
==========================================
  Files          25       25              
  Lines        2157     2275     +118     
==========================================
+ Hits         2036     2050      +14     
- Misses        121      225     +104     

Files	Coverage Δ
src/spikegadgets_to_nwb/convert_dios.py	97.43% <100.00%> (+0.21%)	⬆️
...pikegadgets_to_nwb/tests/test_convert_intervals.py	100.00% <100.00%> (ø)
src/spikegadgets_to_nwb/convert_ephys.py	84.84% <40.00%> (-11.41%)	⬇️
src/spikegadgets_to_nwb/spike_gadgets_raw_io.py	72.32% <9.27%> (-19.57%)	⬇️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.