Define qubit results

[stavros11]

          Aren't you overwriting `results[acquisition.qubit]`?

acquisition is always the same for all result (outer loop), and then acquisition.qubit as well. Isn't it?

Originally posted by @alecandido in https://github.com/qiboteam/qibolab/pull/738#discussion_r1494463241

Currently platform.execute_pulse_sequence (and related execution methods) returns a dictionary that contains pulse serials and qubit names. If the sequence contains more than one readout per qubit, it is not well defined whether the qubit entries should contain all acquired measurements or only the ones from the latest readout pulse.

[alecandido]

The current proposal would be to have a flat dictionary, dict[PulseId, npt.NDArray] for the new unified execute method (defined in #861), with just the ids of the pulses, even when multiple sequences are passed.

E.g., if two sequences are passed, the resulting dictionary will have as keys the IDs of both sequences, the same that would be obtained by running them individually and merging the resulting dictionary.