Speeding Up numba_enhanced_matching_of_library_to_query_pooled_spectra

[CCranney]

Just had a few thoughts on how the numba_enhanced_matching_of_library_to_query_pooled_spectra function might be sped up. Writing here for future reference.

I have it such that a query peak is chosen, then subsequent library peaks are looped over and evaluted for matching. At the time, this seemed best. Now I'm wondering if the following workflow might be better:

1. Pick a query peak.
2. Translate the base PPM tolerance (default 30) into a lower and upper m/z range.
3. Find the indices of all library peaks within this tolerance.
4. Add them as matches.

The intent being to remove the for loop so this scales better with spectra where many library peaks may match to a query peak.

[CCranney]

This is in the matchingFunctions.py file.