Waves: add accessor for waves

[gauteh]

For plotting elevation energy spectra.

• This may have to be a dataset accessor because otherwise we do not have access to the time coordinate.
• In the future this may go into a separate package. But as long as we use the same accessor name the API should be the same for users.

    csv_in = test_data / 'csv/omb3.csv'
    ds = read_omb_csv(csv_in)
    print(ds)
    print(ds.elevation_energy_spectrum)
    print(ds.frequencies_waves_imu)

    plt.figure()
    ds.isel(trajectory=0).elevation_energy_spectrum.wave.plot(
        ds.isel(trajectory=0).time_waves_imu.squeeze())

    if plot:
        plt.show()

    plt.close('all')

Solves #134

[jerabaul29]

@gauteh I am not used too much to pushing commits on someone else's branch; if I should do this another way, just let me know :) .

[gauteh]

This is the way to do it!

[jerabaul29]

note:

• this is probably soon ok to merge as a "feature addition"; for example, this should already more or less fix the need of @hevgyrt :) . I can add a few more tweaks so this looks a bit more like the original spectrum plots.

• @gauteh I guess that you would like https://github.com/OpenDrift/trajan/blob/main/trajan/plot/spectra.py to be removed, right? Otherwise, there is duplicated functionality between the "accessor" based and the "free function" way of doing this plot? :)

[gauteh]

Yep, you remove the old one. I think it is ready to merge. In the future we may have to move it up to dataset level, alternatively use to_1d after selecting trajectory and then plotting (then the times variable won't be necessary).

[jerabaul29]

(a couple more commits landing soon :) )

[jerabaul29]

@gauteh this seems to work now, except for the docs building.

About the doc building: it looks like the error is due to:

trajan/docs/source/api.rst

Line 146 in 735769d

plot.spectra.plot_trajan_spectra

which is true has disappeared as I cleaned up old code.

The strange thing is that it looks like / I would have thought that this file should be re-generated and updated automatically by the CI/CD, right? I.e. I can edit this file by hand if you want, but this is probably not the intended behavior, maybe something that has a problem with how the doc building is run? Is this fixed on main and maybe a lingering bug from this branch, that will disappear at merge? Or a bug on main too that needs to be fixed (here or separately)? If so, should we merge? :)

[jerabaul29]

But I think that otherwise this should be more or less ready to merge? :)

[jerabaul29]

@knutfrode I edited the doc api.rst by hand as we discussed, the CICD is green again :) .

[knutfrode]

Good, then I merge now.