Strange behavior in the time series of the sequence adder

[Louis-PaulHenry]

I've stumbled upon this issue while trying to get the 'spline extrapolated' pulse used by Qutip: when adding a pulse to the sequence, the resulting list of times gather_data(seq)[ch]['time'] is not always strictly increasing, jumping back one time step at the junction between the two pulses (This might have to do with the leading 0 in the list).

[HGSilveri]

Last time I checked, we weren't using gather-data for what is given to the Qutip solver. Are you doing this yourself?

[Louis-PaulHenry]

I'm using gather_data myself to check. The issue was raised by pytest when trying to draw the sequence, and using the times from gather_data

[HGSilveri]

So if the error comes from drawing, were does Qutip come in?

Edit: You want to draw the spline from Qutip, is that it?

[Louis-PaulHenry]

The Qutip does not really come in. It's just that Qutip is using spline interpolation and I'm trying to plot this interpolation (not necessarily via Qutip, currently by extracting the data 'by hand')

edit : gather_data is just a shortcut to get the discretized pulse that Qutip is using

[HGSilveri]

But you want to plot the interpolation that Qutip is actually doing in Simulation, or just an interpolation of the data used for plotting?

[HGSilveri]

edit : gather_data is just a shortcut to get the discretized pulse that Qutip is using

That's my point, the data Qutip is using does not come from gather_data

[Louis-PaulHenry]

I want to plot the interpolation that Qutip is actually doing, but without having to wait for the simulation to be actually run. I'm not sure I can get Qutip to give me beforehands the extrapolated times. I can easily get the hamiltonian at intermediate times, but not the pulse, no?

edit : What I'm trying to get is something similat to sim.get_hamtiltonian(t) but that would spit out omega(t) and delta(t) instead of the full hamiltonian

[HGSilveri]

I don't know how to do that with Qutip, or if Qutip is even useful. I would look into the interpolator that Qutip is using and use that instead. I can fix this error you're getting, but bear in mind that it only comes from an improper use of gather_data, which is not meant to provide data for simulation or interpolation.

I'll push a commit that should solve this issue.

[Louis-PaulHenry]

That's what I tried to do. I just need to find another way than gather_data to extract the discretization of the puse.

[Louis-PaulHenry]

get_hamiltonian works and is consistent with a spline of the output of simul._times for sampling_rate=1

[HGSilveri]

Cool, awesome! If this is solved, feel free to close the issue.

[Louis-PaulHenry]

I still need to do it for generic Hamiltonians/pulses, but the time ordering issue is solved, so I'll close this issue