Simon Rogers
Simon Rogers
Would it make sense for matchms to return a generator rather than the matrix itself? It can then be iterated through and handled however by the user? Sent from my...
I suggest making the filter name more explicit as to what mass is being compared to what? Parent mz? The vast majority of spectra in the universe don’t have a...
Aside from buying a load more RAM :-) Given that the vast majority of scores will be zero, and the scores are symmetric, I’d be tempted to write my own...
Be interested to know why numpy needs to reserve that much memory to store references to 36 million objects. Seems an awful lot.
Bit wary of chemical name as an identifier - will very often not be unique. No better suggestion. in mzml, filename and scan number combined is the best identifier
Feels like this is something that should be decided when the spectra are loaded rather than saved though
Looks sensible to me
@florian-huber there is a library around for adding a spectrum hash (SPLASH) - that would be really neat. It’s computed from the spectrum itself. Maybe that’s already what you meant.....
Morning all, Thanks for this - some observations from me... 1. This is just a dot product after the spectra have been a) normalised so that intensities sum to one...
That’s right - there are two general ways for performing these similarity calculations: 1. Bin the spectra into a fixed set of bins. The spectra is now a vector, and...