Live conversion from QUA to Qiskit simulations

[nulinspiratie]

While writing a QUA program, it would be highly useful to see its effect on a simulated quantum system in real time.

Example

Implementation details

Frontend details

• The code input and plots should be created using Plotly Dash. Alternative suitable suggestions can also be propose, but bidirectional communication needs to be supported: the backend should be able to request an update on the frontend.
• An example of a Dash code editor is found at https://github.com/dawidkopczyk/dash-editor-components
• This feature can later be integrated into #4

simulation details

Here are useful APIs and programs used to generate the example:

• Power Rabi program: This program is a single-dimensional sweep with one QUA variable a, and is a great program to verify the front-end.
• Schedule Plotting: The schedules_to_plot argument accepts a list of indices of schedules to plot. These plotting mechanisms behind this function should be called/cached and reshaped in the background to display the schedule as a function of the QUA variable value on the slider.

  results = simulate_program(
      qua_program=prog,
      qua_config=transmon_pair_qua_config,
      qua_config_to_backend_map=config_to_transmon_pair_backend_map,
      backend=transmon_pair_backend,
      num_shots=10_000,
      schedules_to_plot=[0] <-------- here!
  )

• Results plotting: Once the results are fetched, they can be plot on the y-axis, with the values that the QUA variable will take on on the x-axis.

  # make sure no single point is different to expected within 0.1 tolerance
  q1_state_probabilities = np.array(results[0])
  q2_state_probabilities = np.array(results[0])
  amps = np.arange(start, stop, step)
  
  for i, result in enumerate(results):
      plt.plot(np.arange(start, stop, step), results[i], '.-', label=f"Simulated Q{i}")
      plt.ylim(-0.05, 1.05)
  plt.legend()
  plt.show()

Open questions

• There are different levels of "real time".
  • At a minimum, there should be a command that can be called that takes a QUA program, compiles and simulates it, and then displays a plot.
  • A step above this is a background process with an API that can receive a serialized QUA program and compile+simulate it
  • Another step above is that this background process watches a file for changes and then plots that.
  • What level of real time are we looking for?
• How to add the config to this live converter?
• Where should it plot the results? In the frontend web app or as a program?

[Piwakk]

Hi @nulinspiratie, just like #4, could I be assigned to the issue? Thanks!

[nulinspiratie]

Apologies, I thought I was supposed to add you to the PR. Fixed!