# Repository: https://gitlab.com/quantify-os/quantify-scheduler
# Licensed according to the LICENCE file on the main branch
"""Standard acquisition protocols for use with the quantify_scheduler."""
from __future__ import annotations
import warnings
from typing import Any
import numpy as np
from quantify_core.utilities import deprecated
from quantify_scheduler.enums import BinMode, TimeRef, TimeSource
from quantify_scheduler.operations.operation import Operation
from quantify_scheduler.resources import DigitalClockResource
[docs]
class Acquisition(Operation):
"""
An operation representing data acquisition at the quantum-device abstraction layer.
An Acquisition must consist of (at least) an AcquisitionProtocol specifying how the
acquired signal is to be processed, and an AcquisitionChannel and AcquisitionIndex
specifying where the acquired data is to be stored in the RawDataset.
N.B. This class helps differentiate an acquisition operation from the regular
operations. This enables us to use
:func:`~.quantify_scheduler.schedules._visualization.pulse_diagram.plot_acquisition_operations`
to highlight acquisition pulses in the pulse diagrams.
"""
@deprecated("1.0", Acquisition)
[docs]
class AcquisitionOperation(Acquisition):
"""Deprecated alias."""
pass
[docs]
class Trace(Acquisition):
"""
The Trace acquisition protocol measures a signal s(t).
Only processing performed is rescaling and adding
units based on a calibrated scale. Values are returned
as a raw trace (numpy array of float datatype). Length of
this array depends on the sampling rate of the acquisition
device.
.. important::
The exact duration of this operation, and the possible bin modes may depend on
the control hardware. Please consult your hardware vendor's :ref:`Reference
guide` for more information.
Parameters
----------
port
The acquisition port.
clock
The clock used to demodulate the acquisition.
duration
The acquisition duration in seconds.
acq_channel
The data channel in which the acquisition is stored, is by default 0.
Describes the "where" information of the measurement, which typically
corresponds to a qubit idx.
acq_index
The data register in which the acquisition is stored, by default 0.
Describes the "when" information of the measurement, used to label or
tag individual measurements in a large circuit. Typically corresponds
to the setpoints of a schedule (e.g., tau in a T1 experiment).
bin_mode
Describes what is done when data is written to a memory location that already
contains values. Which bin mode can be used for Trace acquisitions may depend on
the hardware. ``BinMode.AVERAGE``, the default, works on most hardware. This bin
mode stores the weighted average value of the new result and the old values.
``BinMode.FIRST`` is used for hardware where only the result of the first
acquisition in a Schedule is stored, e.g. for a Trace acquisition with Qblox QTM
modules.
t0
The acquisition start time in seconds, by default 0.
"""
def __init__(
self,
duration: float,
port: str,
clock: str,
acq_channel: int = 0,
acq_index: int = 0,
bin_mode: BinMode | str = BinMode.AVERAGE,
t0: float = 0,
) -> None:
if not isinstance(duration, float):
duration = float(duration)
if isinstance(bin_mode, str):
bin_mode = BinMode(bin_mode)
super().__init__(name=self.__class__.__name__)
self.data["acquisition_info"] = [
{
"waveforms": [],
"duration": duration,
"t0": t0,
"port": port,
"clock": clock,
"acq_channel": acq_channel,
"acq_index": acq_index,
"bin_mode": bin_mode,
"protocol": "Trace",
"acq_return_type": np.ndarray,
}
]
self._update()
def __str__(self) -> str:
acq_info = self.data["acquisition_info"][0]
return self._get_signature(acq_info)
[docs]
class WeightedIntegratedSeparated(Acquisition):
r"""
Weighted integration acquisition protocol where two sets weights
are applied separately to the real and imaginary parts
of the signal.
Weights are applied as:
.. math::
\widetilde{A} = \int \mathrm{Re}(S(t))\cdot W_A(t) \mathrm{d}t
.. math::
\widetilde{B} = \int \mathrm{Im}(S(t))\cdot W_B(t) \mathrm{d}t
Parameters
----------
waveform_a
The complex waveform used as integration weights :math:`W_A(t)`.
waveform_b
The complex waveform used as integration weights :math:`W_B(t)`.
port
The acquisition port.
clock
The clock used to demodulate the acquisition.
duration
The acquisition duration in seconds.
acq_channel
The data channel in which the acquisition is stored, by default 0.
Describes the "where" information of the measurement, which typically
corresponds to a qubit idx.
acq_index
The data register in which the acquisition is stored, by default 0.
Describes the "when" information of the measurement, used to label or
tag individual measurements in a large circuit. Typically corresponds
to the setpoints of a schedule (e.g., tau in a T1 experiment).
bin_mode
Describes what is done when data is written to a register that already
contains a value. Options are "append" which appends the result to the
list or "average" which stores the weighted average value of the
new result and the old register value, by default BinMode.APPEND.
phase
The phase of the pulse and acquisition in degrees, by default 0.
t0
The acquisition start time in seconds, by default 0.
Raises
------
NotImplementedError
"""
def __init__(
self,
waveform_a: dict[str, Any],
waveform_b: dict[str, Any],
port: str,
clock: str,
duration: float,
acq_channel: int = 0,
acq_index: int = 0,
bin_mode: BinMode | str = BinMode.APPEND,
phase: float = 0,
t0: float = 0,
) -> None:
if phase != 0:
raise NotImplementedError("Non-zero phase not yet implemented")
super().__init__(name=self.__class__.__name__)
self.data["acquisition_info"] = [
{
"waveforms": [waveform_a, waveform_b],
"t0": t0,
"clock": clock,
"port": port,
"duration": duration,
"phase": phase,
"acq_channel": acq_channel,
"acq_index": acq_index,
"bin_mode": bin_mode,
"protocol": "WeightedIntegratedSeparated",
"acq_return_type": complex,
}
]
self._update()
# certain fields are required in the acquisition data
if "acq_return_type" not in self.data["acquisition_info"][0]:
self.data["acquisition_info"][0]["acq_return_type"] = complex
self.data["acquisition_info"][0]["protocol"] = "WeightedIntegratedSeparated"
def __str__(self) -> str:
acq_info = self.data["acquisition_info"][0]
return self._get_signature(acq_info)
[docs]
class SSBIntegrationComplex(Acquisition):
"""
Single sideband integration acquisition protocol with complex results.
A weighted integrated acquisition on a complex signal using a
square window for the acquisition weights.
The signal is demodulated using the specified clock, and the
square window then effectively specifies an integration window.
Parameters
----------
port
The acquisition port.
clock
The clock used to demodulate the acquisition.
duration
The acquisition duration in seconds.
acq_channel
The data channel in which the acquisition is stored, by default 0.
Describes the "where" information of the measurement, which typically
corresponds to a qubit idx.
acq_index
The data register in which the acquisition is stored, by default 0.
Describes the "when" information of the measurement, used to label or
tag individual measurements in a large circuit. Typically corresponds
to the setpoints of a schedule (e.g., tau in a T1 experiment).
bin_mode
Describes what is done when data is written to a register that already
contains a value. Options are "append" which appends the result to the
list or "average" which stores the weighted average value of the
new result and the old register value, by default BinMode.AVERAGE.
phase
The phase of the pulse and acquisition in degrees, by default 0.
t0
The acquisition start time in seconds, by default 0.
"""
def __init__(
self,
port: str,
clock: str,
duration: float,
acq_channel: int = 0,
acq_index: int = 0,
bin_mode: BinMode | str = BinMode.AVERAGE,
phase: float = 0,
t0: float = 0,
) -> None:
if phase != 0:
raise NotImplementedError("Non-zero phase not yet implemented")
super().__init__(name=self.__class__.__name__)
self.data["acquisition_info"] = [
{
"waveforms": [waveform_i, waveform_q],
"t0": t0,
"clock": clock,
"port": port,
"duration": duration,
"phase": phase,
"acq_channel": acq_channel,
"acq_index": acq_index,
"bin_mode": bin_mode,
"acq_return_type": complex,
"protocol": "SSBIntegrationComplex",
}
]
self._update()
# certain fields are required in the acquisition data
if "acq_return_type" not in self.data["acquisition_info"][0]:
self.data["acquisition_info"][0]["acq_return_type"] = complex
self.data["acquisition_info"][0]["protocol"] = "SSBIntegrationComplex"
def __str__(self) -> str:
acq_info = self.data["acquisition_info"][0]
return self._get_signature(acq_info)
[docs]
class ThresholdedAcquisition(Acquisition):
"""
Acquisition protocol allowing to control rotation and threshold.
This acquisition protocol is similar to the :class:`~.SSBIntegrationComplex`
acquisition protocol, but the complex result is now rotated and thresholded
to produce a "0" or a "1", as controlled by the parameters for rotation
angle `<qubit>.measure.acq_rotation` and threshold value
`<qubit>.measure.acq_threshold` in the device configuration (see example
below).
The rotation angle and threshold value for each qubit can be set through
the device configuration.
.. admonition:: Note
Thresholded acquisition is currently only supported by the Qblox
backend.
.. admonition:: Examples
.. jupyter-execute::
from quantify_scheduler import Schedule
from quantify_scheduler.device_under_test.transmon_element import BasicTransmonElement
from quantify_scheduler.operations.acquisition_library import ThresholdedAcquisition
# set up qubit
qubit = BasicTransmonElement("q0")
qubit.clock_freqs.readout(8.0e9)
# set rotation and threshold value
rotation, threshold = 20, -0.1
qubit.measure.acq_rotation(rotation)
qubit.measure.acq_threshold(threshold)
# basic schedule
schedule = Schedule("thresholded acquisition")
schedule.add(ThresholdedAcquisition(port="q0:res", clock="q0.ro", duration=1e-6))
Parameters
----------
port : str
The acquisition port.
clock : str
The clock used to demodulate the acquisition.
duration : float
The acquisition duration in seconds.
acq_channel : int
The data channel in which the acquisition is stored, by default 0.
Describes the "where" information of the measurement, which
typically corresponds to a qubit idx.
acq_index : int
The data register in which the acquisition is stored, by default 0.
Describes the "when" information of the measurement, used to label
or tag individual measurements in a large circuit. Typically
corresponds to the setpoints of a schedule (e.g., tau in a T1
experiment).
bin_mode : BinMode or str
Describes what is done when data is written to a register that
already contains a value. Options are "append" which appends the
result to the list or "average" which stores the weighted average
value of the new result and the old register value, by default
BinMode.AVERAGE.
feedback_trigger_label : str
The label corresponding to the feedback trigger, which is mapped by the
compiler to a feedback trigger address on hardware, by default None.
phase : float
The phase of the pulse and acquisition in degrees, by default 0.
t0 : float
The acquisition start time in seconds, by default 0.
"""
def __init__(
self,
port: str,
clock: str,
duration: float,
acq_channel: int = 0,
acq_index: int = 0,
bin_mode: BinMode | str = BinMode.AVERAGE,
feedback_trigger_label: str | None = None,
phase: float = 0,
t0: float = 0,
acq_rotation: float = 0,
acq_threshold: float = 0,
) -> None:
if phase != 0:
raise NotImplementedError("Non-zero phase not yet implemented")
super().__init__(name=self.__class__.__name__)
self.data["acquisition_info"] = [
{
"waveforms": [waveform_i, waveform_q],
"t0": t0,
"clock": clock,
"port": port,
"duration": duration,
"phase": phase,
"acq_channel": acq_channel,
"acq_index": acq_index,
"bin_mode": bin_mode,
"acq_return_type": np.uint32,
"protocol": "ThresholdedAcquisition",
"feedback_trigger_label": feedback_trigger_label,
"acq_threshold": acq_threshold,
"acq_rotation": acq_rotation,
},
]
self._update()
def __str__(self) -> str:
acq_info = self.data["acquisition_info"][0]
return self._get_signature(acq_info)
[docs]
class NumericalSeparatedWeightedIntegration(WeightedIntegratedSeparated):
r"""
Subclass of :class:`~WeightedIntegratedSeparated` with parameterized waveforms as weights.
A WeightedIntegratedSeparated class using parameterized waveforms and
interpolation as the integration weights.
Weights are applied as:
.. math::
\widetilde{A} = \int \mathrm{Re}(S(t)\cdot W_A(t) \mathrm{d}t
.. math::
\widetilde{B} = \int \mathrm{Im}(S(t))\cdot W_B(t) \mathrm{d}t
Parameters
----------
port
The acquisition port.
clock
The clock used to demodulate the acquisition.
weights_a
The list of complex values used as weights :math:`A(t)` on
the incoming complex signal.
weights_b
The list of complex values used as weights :math:`B(t)` on
the incoming complex signal.
weights_sampling_rate
The rate with which the weights have been sampled, in Hz. By default equal
to 1 GHz. Note that during hardware compilation, the weights will be resampled
with the sampling rate supported by the target hardware.
interpolation
The type of interpolation to use, by default "linear". This argument is
passed to :obj:`~scipy.interpolate.interp1d`.
acq_channel
The data channel in which the acquisition is stored, by default 0.
Describes the "where" information of the measurement, which typically
corresponds to a qubit idx.
acq_index
The data register in which the acquisition is stored, by default 0.
Describes the "when" information of the measurement, used to label or
tag individual measurements in a large circuit. Typically corresponds
to the setpoints of a schedule (e.g., tau in a T1 experiment).
bin_mode
Describes what is done when data is written to a register that already
contains a value. Options are "append" which appends the result to the
list or "average" which stores the weighted average value of the
new result and the old register value, by default BinMode.APPEND.
phase
The phase of the pulse and acquisition in degrees, by default 0.
t0
The acquisition start time in seconds, by default 0.
"""
def __init__(
self,
port: str,
clock: str,
weights_a: list[complex] | np.ndarray,
weights_b: list[complex] | np.ndarray,
weights_sampling_rate: float = 1e9,
interpolation: str = "linear",
acq_channel: int = 0,
acq_index: int = 0,
bin_mode: BinMode | str = BinMode.APPEND,
phase: float = 0,
t0: float = 0,
) -> None:
[docs]
t_samples = np.arange(len(weights_a)) / weights_sampling_rate
[docs]
weights_a = np.array(weights_a)
[docs]
weights_b = np.array(weights_b)
[docs]
duration = len(t_samples) / weights_sampling_rate
super().__init__(
waveform_a=waveforms_a,
waveform_b=waveforms_b,
port=port,
clock=clock,
duration=duration,
acq_channel=acq_channel,
acq_index=acq_index,
bin_mode=bin_mode,
phase=phase,
t0=t0,
)
self.data["name"] = self.__class__.__name__
self.data["acquisition_info"][0][
"protocol"
] = "NumericalSeparatedWeightedIntegration"
self._update()
def __str__(self) -> str:
acq_info = self.data["acquisition_info"][0]
weights_a = np.array2string(
acq_info["waveforms"][0]["samples"], separator=", ", precision=9
)
weights_b = np.array2string(
acq_info["waveforms"][1]["samples"], separator=", ", precision=9
)
t_samples = acq_info["waveforms"][0]["t_samples"]
weights_sampling_rate = 1 / (t_samples[1] - t_samples[0])
port = acq_info["port"]
clock = acq_info["clock"]
interpolation = acq_info["waveforms"][0]["interpolation"]
acq_channel = acq_info["acq_channel"]
acq_index = acq_info["acq_index"]
bin_mode = acq_info["bin_mode"].value
phase = acq_info["phase"]
t0 = acq_info["t0"]
return (
f"{self.__class__.__name__}(weights_a={weights_a}, weights_b={weights_b}, "
f"{weights_sampling_rate=}, {port=}, {clock=}, {interpolation=}, "
f"{acq_channel=}, {acq_index=}, {bin_mode=}, {phase=}, {t0=})"
)
def __repr__(self) -> str:
return str(self)
[docs]
class NumericalWeightedIntegrationComplex(WeightedIntegratedSeparated):
"""Deprecated, renamed to :class:`~NumericalSeparatedWeightedIntegration`."""
def __new__(cls, *args, **kwargs) -> NumericalSeparatedWeightedIntegration:
"""Return :class:`~NumericalSeparatedWeightedIntegration`."""
warnings.warn(
(
f"{NumericalWeightedIntegrationComplex.__name__} is "
f"deprecated and will be removed in quantify-scheduler>=0.20.0. Use "
f"{NumericalSeparatedWeightedIntegration.__name__} instead."
),
FutureWarning,
)
return NumericalSeparatedWeightedIntegration(*args, **kwargs)
[docs]
class NumericalWeightedIntegration(NumericalSeparatedWeightedIntegration):
"""
Subclass of :class:`~NumericalSeparatedWeightedIntegration` returning a complex number.
Parameters
----------
port
The acquisition port.
clock
The clock used to demodulate the acquisition.
weights_a
The list of complex values used as weights :math:`A(t)` on
the incoming complex signal.
weights_b
The list of complex values used as weights :math:`B(t)` on
the incoming complex signal.
weights_sampling_rate
The rate with which the weights have been sampled, in Hz. By default equal
to 1 GHz. Note that during hardware compilation, the weights will be resampled
with the sampling rate supported by the target hardware.
t
The time values of each weight. This parameter is deprecated in favor of
``weights_sampling_rate``. If a value is provided for ``t``, the
``weights_sampling_rate`` parameter will be ignored.
interpolation
The type of interpolation to use, by default "linear". This argument is
passed to :obj:`~scipy.interpolate.interp1d`.
acq_channel
The data channel in which the acquisition is stored, by default 0.
Describes the "where" information of the measurement, which typically
corresponds to a qubit idx.
acq_index
The data register in which the acquisition is stored, by default 0.
Describes the "when" information of the measurement, used to label or
tag individual measurements in a large circuit. Typically corresponds
to the setpoints of a schedule (e.g., tau in a T1 experiment).
bin_mode
Describes what is done when data is written to a register that already
contains a value. Options are "append" which appends the result to the
list or "average" which stores the weighted average value of the
new result and the old register value, by default BinMode.APPEND.
phase
The phase of the pulse and acquisition in degrees, by default 0.
t0
The acquisition start time in seconds, by default 0.
"""
def __init__(
self,
port: str,
clock: str,
weights_a: list[complex] | np.ndarray,
weights_b: list[complex] | np.ndarray,
weights_sampling_rate: float = 1e9,
interpolation: str = "linear",
acq_channel: int = 0,
acq_index: int = 0,
bin_mode: BinMode | str = BinMode.APPEND,
phase: float = 0,
t0: float = 0,
) -> None:
super().__init__(
port=port,
clock=clock,
weights_a=weights_a,
weights_b=weights_b,
weights_sampling_rate=weights_sampling_rate,
interpolation=interpolation,
acq_channel=acq_channel,
acq_index=acq_index,
bin_mode=bin_mode,
phase=phase,
t0=t0,
)
self.data["acquisition_info"][0]["protocol"] = "NumericalWeightedIntegration"
self._update()
[docs]
class TriggerCount(Acquisition):
"""
Trigger counting acquisition protocol returning an integer.
The trigger acquisition mode is used to measure how
many times the trigger level is surpassed. The level is set
in the hardware configuration.
.. important::
The exact duration of this operation, and the possible bin modes may depend on
the control hardware. Please consult your hardware vendor's :ref:`Reference
guide` for more information.
Parameters
----------
port
The acquisition port.
clock
The clock used to demodulate the acquisition.
duration
The duration of the operation in seconds.
acq_channel
The data channel in which the acquisition is stored, by default 0.
Describes the "where" information of the measurement, which typically
corresponds to a qubit idx.
acq_index
The data register in which the acquisition is stored, by default 0.
Describes the "when" information of the measurement, used to label or
tag individual measurements in a large circuit. Typically corresponds
to the setpoints of a schedule (e.g., tau in a T1 experiment).
bin_mode
Describes what is done when data is written to a register that already
contains a value. Options are "append" which appends the result to the
list or "distribution" which stores the count value of the
new result and the old register value, by default BinMode.APPEND.
t0
The acquisition start time in seconds, by default 0.
"""
def __init__(
self,
port: str,
clock: str,
duration: float,
acq_channel: int = 0,
acq_index: int = 0,
bin_mode: BinMode | str = BinMode.APPEND,
t0: float = 0,
) -> None:
if bin_mode == BinMode.AVERAGE:
warnings.warn(
(
f"{bin_mode} is deprecated for the TriggerCount acquisition protocol, "
f"and will be removed in quantify-scheduler>=0.24.0. "
f"Use {BinMode.DISTRIBUTION} instead, which has the same effect."
),
FutureWarning,
)
bin_mode = BinMode.DISTRIBUTION
if bin_mode == BinMode.DISTRIBUTION and acq_index != 0:
# In average mode the count distribution is measured,
# and currently we do not support multiple indices for this,
# or starting the counting from a predefined count number.
raise NotImplementedError(
"Using nonzero acq_index is not yet implemented for AVERAGE bin mode for "
"the trigger count protocol"
)
super().__init__(name=self.__class__.__name__)
self.data["acquisition_info"] = [
{
"waveforms": [],
"t0": t0,
"clock": clock,
"port": port,
"duration": duration,
"acq_channel": acq_channel,
"acq_index": acq_index,
"bin_mode": bin_mode,
"acq_return_type": int,
"protocol": "TriggerCount",
}
]
self._update()
def __str__(self) -> str:
acq_info = self.data["acquisition_info"][0]
return self._get_signature(acq_info)
