The plot might become pretty crowded if multiple peaks are close to each other though. Anyway:

spec = spec.scale_intensity(max_intensity=1)
for fragment_mz, fragment_in in zip(spec.mz, spec.intensity):
    if fragment_in >= 0.05:
        spec.annotate_mz_fragment(fragment_mz, None, 0.1, 'Da')

If you select the most intense peak within a certain window (e.g. 10 mz, or scale the window with the max(mz) so it's a larger window for higher mz spectra) for annotation, you will be able to avoid most overlap automatically, but still be informative (at least those values are what worked for us in R). Depends how you wish to orient the text. I piggybacked off the satellite peak filtering in RMassBank:
https://github.com/schymane/ReSOLUTION/blob/master/R/MS1_functions.R#L115

I could easily implement Emma's suggestion within the app...

Done - it's a bit messy, but we can tidy up.

• Plots just 2dp of mz
• Doesn't plot things too close in mz to one another
• Width (for exclusion) is calculated based upon min and max mz range to ensure zoomed in regions can still get plenty of labels

Can we have 4 dp please? Or an option to add more than 2? It's essential for interpretation.

Done...will be online next time Ming updates

Awesome! Just saw the thread. I made a few changes to display both the unlabled and labeled SVG in the default plot:

https://metabolomics-usi.ucsd.edu/spectrum/?usi=mzspec:GNPSLIBRARY:CCMSLIB00000840338

Deployed live, cutting release.

Thanks @schymane @sdrogers @bittremieux @alan-jarmusch.

Ming

Looking great! Tiny suggestion, if you tilt the masses on an angle (eg 25-30 degrees?) you'll avoid it crossing the line and it makes it easier for people to read ...
Great work! Need to integrate MassBank at some point ;-)

Awesome, will make issue about suggestion!

Going to also add issue integrating massbank. If you have a web api should be super easy to integrate.

@bittremieux if you are responsible for the italics on m/z... then kudos

I'm actually not a big fan of the rotated peak labels because that way they can overlap with neighboring peaks.

yes I see what you mean, my x axis is usually much more stretched wrt y so it works better as the peaks are more spread out. Maybe add this as an option that people can adjust? Vertical is very difficult to read, and some like horizontal even if this overlaps even more ...