(wip) Additional window groups splitting strategies.

[jspaezp]

This is a pretty rough implementation of alternative ways to split DIA frames into raw spectra.

If this is within the scope of what you want to support I would be happy to finish it as a full feature (+ add testing, + add documentation/comments, + remove debug prints)

LMK what you think! @sander-willems-bruker

[jspaezp]

This is here just for debugging ...

[jspaezp]

Obviously this would be passed as an argument and not read from the env. variables. IMO

[jspaezp]

This would also be a good time to expose the smoothing-centroiding parameters to the public API ... which would be something that gets passed to timsrust::io::readers::SpectrumReader::new ... which could be something like ...

struct SpectrumReaderParameters {
    spectrum_processing_params: <... smoothing and centroiding stuff ...>,
    frame_splitting_params: Option<FrameSplittingStrategy>
}

[jspaezp]

An additional thing that could go into the strategy would be whether it should be done at the quad+window level or only at the window level (which would enable diagonal pasef methods to be split as well).

[sander-willems-bruker]

Thanks for the PR, looks very nice! Tested it out locally and can confirm your findings with regards to speed and identification performance. If you are already working on making it a clean PR, I'll let you finish that up. If you haven't started yet though, I can also take it from here and work on it locallu before merging into dev.

[sander-willems-bruker]

This would also be a good time to expose the smoothing-centroiding parameters to the public API ... which would be something that gets passed to timsrust::io::readers::SpectrumReader::new ... which could be something like ...

struct SpectrumReaderParameters {
    spectrum_processing_params: <... smoothing and centroiding stuff ...>,
    frame_splitting_params: Option<FrameSplittingStrategy>
}

This indeed is definately something that needs to happen indeed. For now, leave it to me, I have some local developments in sage that will also utilize this. DiagonalPASEF is still onb my todo list, don't have a clear idea how I would integrate that as clean as possible although your implementation serves as a very decent start for brainstorming

[jspaezp]

Right on! I will propagate some of the changes and implement the strategy pattern for the config! (I first wanted to know that this was something you would be interested in before going ahead and "back-propagating" the configuration)

[jspaezp]

@sander-willems-bruker Just FYI your last push to develop introduced A LOT of merge conflicts. I would have appreciated a heads up.

reference: https://github.com/MannLabs/timsrust/compare/cc83557b5dd6ed567abd2d72780726edd2c55b1a..21efdf9e5279a200195c4bdb61fdd7738e219b66

[sander-willems-bruker]

Yes, my apologies. I should have made a separate branch before merging to develop. Will try to make that my new routine for future work. We'll continue on the other PR;)