# Source code for pyquil.api._quantum_computer

```
##############################################################################
# Copyright 2018 Rigetti Computing
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
##############################################################################
import re
import socket
import warnings
from math import pi, log
from typing import List, Dict, Tuple, Iterator, Union
import itertools
import subprocess
from contextlib import contextmanager
import networkx as nx
import numpy as np
from rpcq.messages import BinaryExecutableResponse, PyQuilExecutableResponse
from pyquil.api._compiler import QPUCompiler, QVMCompiler
from pyquil.api._config import PyquilConfig
from pyquil.api._devices import get_lattice, list_lattices
from pyquil.api._error_reporting import _record_call
from pyquil.api._qac import AbstractCompiler
from pyquil.api._qam import QAM
from pyquil.api._qpu import QPU
from pyquil.api._qvm import ForestConnection, QVM
from pyquil.device import AbstractDevice, NxDevice, gates_in_isa, ISA, Device
from pyquil.gates import RX, MEASURE
from pyquil.noise import decoherence_noise_with_asymmetric_ro, NoiseModel
from pyquil.pyqvm import PyQVM
from pyquil.quil import Program, validate_supported_quil
Executable = Union[BinaryExecutableResponse, PyQuilExecutableResponse]
[docs]class QuantumComputer:
def __init__(self, *,
name: str,
qam: QAM,
device: AbstractDevice,
compiler: AbstractCompiler,
symmetrize_readout: bool = False) -> None:
"""
A quantum computer for running quantum programs.
A quantum computer has various characteristics like supported gates, qubits, qubit
topologies, gate fidelities, and more. A quantum computer also has the ability to
run quantum programs.
A quantum computer can be a real Rigetti QPU that uses superconducting transmon
qubits to run quantum programs, or it can be an emulator like the Rigetti QVM with
noise models and mimicked topologies.
:param name: A string identifying this particular quantum computer.
:param qam: A quantum abstract machine which handles executing quantum programs. This
dispatches to a QVM or QPU.
:param device: A collection of connected qubits and associated specs and topology.
:param symmetrize_readout: Whether to apply readout error symmetrization. See
:py:func:`run_symmetrized_readout` for a complete description.
"""
self.name = name
self.qam = qam
self.device = device
self.compiler = compiler
self.symmetrize_readout = symmetrize_readout
[docs] def qubits(self) -> List[int]:
"""
Return a sorted list of this QuantumComputer's device's qubits
See :py:func:`AbstractDevice.qubits` for more.
"""
return self.device.qubits()
[docs] def qubit_topology(self) -> nx.graph:
"""
Return a NetworkX graph representation of this QuantumComputer's device's qubit
connectivity.
See :py:func:`AbstractDevice.qubit_topology` for more.
"""
return self.device.qubit_topology()
[docs] def get_isa(self, oneq_type: str = 'Xhalves',
twoq_type: str = 'CZ') -> ISA:
"""
Return a target ISA for this QuantumComputer's device.
See :py:func:`AbstractDevice.get_isa` for more.
:param oneq_type: The family of one-qubit gates to target
:param twoq_type: The family of two-qubit gates to target
"""
return self.device.get_isa(oneq_type=oneq_type, twoq_type=twoq_type)
[docs] @_record_call
def run(self, executable: Executable,
memory_map: Dict[str, List[Union[int, float]]] = None) -> np.ndarray:
"""
Run a quil executable. If the executable contains declared parameters, then a memory
map must be provided, which defines the runtime values of these parameters.
:param executable: The program to run. You are responsible for compiling this first.
:param memory_map: The mapping of declared parameters to their values. The values
are a list of floats or integers.
:return: A numpy array of shape (trials, len(ro-register)) that contains 0s and 1s.
"""
self.qam.load(executable)
if memory_map:
for region_name, values_list in memory_map.items():
for offset, value in enumerate(values_list):
# TODO gh-658: have write_memory take a list rather than value + offset
self.qam.write_memory(region_name=region_name, offset=offset, value=value)
return self.qam.run() \
.wait() \
.read_memory(region_name='ro')
[docs] @_record_call
def run_symmetrized_readout(self, program: Program, trials: int, symm_type: int = 3,
meas_qubits: List[int] = None) -> np.ndarray:
r"""
Run a quil program in such a way that the readout error is made symmetric. Enforcing
symmetric readout error is useful in simplifying the assumptions in some near
term error mitigation strategies, see ``measure_observables`` for more information.
The simplest example is for one qubit. In a noisy device, the probability of accurately
reading the 0 state might be higher than that of the 1 state; due to e.g. amplitude
damping. This makes correcting for readout more difficult. In the simplest case, this
function runs the program normally ``(trials//2)`` times. The other half of the time,
it will insert an ``X`` gate prior to any ``MEASURE`` instruction and then flip the
measured classical bit back. Overall this has the effect of symmetrizing the readout error.
The details. Consider preparing the input bitstring ``|i>`` (in the computational basis) and
measuring in the Z basis. Then the Confusion matrix for the readout error is specified by
the probabilities
p(j|i) := Pr(measured = j | prepared = i ).
In the case of a single qubit i,j \in [0,1] then:
there is no readout error if p(0|0) = p(1|1) = 1.
the readout error is symmetric if p(0|0) = p(1|1) = 1 - epsilon.
the readout error is asymmetric if p(0|0) != p(1|1).
If your quantum computer has this kind of asymmetric readout error then
``qc.run_symmetrized_readout`` will symmetrize the readout error.
The readout error above is only asymmetric on a single bit. In practice the confusion
matrix on n bits need not be symmetric, e.g. for two qubits p(ij|ij) != 1 - epsilon for
all i,j. In these situations a more sophisticated means of symmetrization is needed; and
we use orthogonal arrays (OA) built from Hadamard matrices.
The symmetrization types are specified by an int; the types available are:
-1 -- exhaustive symmetrization uses every possible combination of flips
0 -- trivial that is no symmetrization
1 -- symmetrization using an OA with strength 1
2 -- symmetrization using an OA with strength 2
3 -- symmetrization using an OA with strength 3
In the context of readout symmetrization the strength of the orthogonal array enforces
the symmetry of the marginal confusion matrices.
By default a strength 3 OA is used; this ensures expectations of the form
``<b_k . b_j . b_i>`` for bits any bits i,j,k will have symmetric readout errors. Here
expectation of a random variable x as is denote ``<x> = sum_i Pr(i) x_i``. It turns out that
a strength 3 OA is also a strength 2 and strength 1 OA it also ensures ``<b_j . b_i>`` and
``<b_i>`` have symmetric readout errors for any bits b_j and b_i.
:param program: The program to run symmetrized readout on.
:param trials: The minimum number of times to run the program; it is recommend that this
number should be in the hundreds or thousands. This parameter will be mutated if
necessary.
:param symm_type: the type of symmetrization
:param meas_qubits: An advanced feature. The groups of measurement qubits. Only these
qubits will be symmetrized over, even if the program acts on other qubits.
:return: A numpy array of shape (trials, len(ro-register)) that contains 0s and 1s.
"""
if not isinstance(symm_type, int):
raise ValueError("Symmetrization options are indicated by an int. See "
"the docstrings for more information.")
if meas_qubits is None:
meas_qubits = list(program.get_qubits())
# It is desirable to have hundreds or thousands of trials more than the minimum
trials = _check_min_num_trials_for_symmetrized_readout(len(meas_qubits), trials, symm_type)
sym_programs, flip_arrays = _symmetrization(program, meas_qubits, symm_type)
# Floor division so e.g. 9 // 8 = 1 and 17 // 8 = 2.
num_shots_per_prog = trials // len(sym_programs)
if num_shots_per_prog * len(sym_programs) < trials:
warnings.warn(f"The number of trials was modified from {trials} to "
f"{num_shots_per_prog * len(sym_programs)}. To be consistent with the "
f"number of trials required by the type of readout symmetrization "
f"chosen.")
results = _measure_bitstrings(self, sym_programs, meas_qubits, num_shots_per_prog)
return _consolidate_symmetrization_outputs(results, flip_arrays)
[docs] @_record_call
def run_and_measure(self, program: Program, trials: int) -> Dict[int, np.ndarray]:
"""
Run the provided state preparation program and measure all qubits.
This will measure all the qubits on this QuantumComputer, not just qubits
that are used in the program.
The returned data is a dictionary keyed by qubit index because qubits for a given
QuantumComputer may be non-contiguous and non-zero-indexed. To turn this dictionary
into a 2d numpy array of bitstrings, consider::
bitstrings = qc.run_and_measure(...)
bitstring_array = np.vstack(bitstrings[q] for q in qc.qubits()).T
bitstring_array.shape # (trials, len(qc.qubits()))
.. note::
In contrast to :py:class:`QVMConnection.run_and_measure`, this method simulates
noise correctly for noisy QVMs. However, this method is slower for ``trials > 1``.
For faster noise-free simulation, consider
:py:class:`WavefunctionSimulator.run_and_measure`.
:param program: The state preparation program to run and then measure.
:param trials: The number of times to run the program.
:return: A dictionary keyed by qubit index where the corresponding value is a 1D array of
measured bits.
"""
program = program.copy()
validate_supported_quil(program)
ro = program.declare('ro', 'BIT', len(self.qubits()))
for i, q in enumerate(self.qubits()):
program.inst(MEASURE(q, ro[i]))
program.wrap_in_numshots_loop(trials)
executable = self.compile(program)
bitstring_array = self.run(executable=executable)
bitstring_dict = {}
for i, q in enumerate(self.qubits()):
bitstring_dict[q] = bitstring_array[:, i]
return bitstring_dict
[docs] @_record_call
def compile(self, program: Program,
to_native_gates: bool = True,
optimize: bool = True,
protoquil: bool = None) -> Union[BinaryExecutableResponse, PyQuilExecutableResponse]:
"""
A high-level interface to program compilation.
Compilation currently consists of two stages. Please see the :py:class:`AbstractCompiler`
docs for more information. This function does all stages of compilation.
Right now both ``to_native_gates`` and ``optimize`` must be either both set or both
unset. More modular compilation passes may be available in the future.
Additionally, a call to compile also calls the ``reset`` method if one is running
on the QPU. This is a bit of a sneaky hack to guard against stale compiler connections,
but shouldn't result in any material hit to performance (especially when taking advantage
of parametric compilation for hybrid applications).
:param program: A Program
:param to_native_gates: Whether to compile non-native gates to native gates.
:param optimize: Whether to optimize the program to reduce the number of operations.
:param protoquil: Whether to restrict the input program to and the compiled program
to protoquil (executable on QPU). A value of ``None`` means defer to server.
:return: An executable binary suitable for passing to :py:func:`QuantumComputer.run`.
"""
if isinstance(self.qam, QPU):
self.reset()
flags = [to_native_gates, optimize]
assert all(flags) or all(not f for f in flags), "Must turn quilc all on or all off"
quilc = all(flags)
if quilc:
nq_program = self.compiler.quil_to_native_quil(program, protoquil=protoquil)
else:
nq_program = program
binary = self.compiler.native_quil_to_executable(nq_program)
return binary
@_record_call
def reset(self):
"""
Reset the QuantumComputer's QAM to its initial state, and refresh all the connection
objects in the event that the ~/.forest_config file has changed during the existence
of this QuantumComputer object.
"""
self.qam.reset()
self.compiler.reset()
def __str__(self) -> str:
return self.name
def __repr__(self):
return f'QuantumComputer[name="{self.name}"]'
[docs]@_record_call
def list_quantum_computers(connection: ForestConnection = None,
qpus: bool = True,
qvms: bool = True) -> List[str]:
"""
List the names of available quantum computers
:param connection: An optional :py:class:ForestConnection` object. If not specified,
the default values for URL endpoints will be used, and your API key
will be read from ~/.pyquil_config. If you deign to change any
of these parameters, pass your own :py:class:`ForestConnection` object.
:param qpus: Whether to include QPU's in the list.
:param qvms: Whether to include QVM's in the list.
"""
if connection is None:
connection = ForestConnection()
qc_names: List[str] = []
if qpus:
qc_names += list(list_lattices(connection=connection).keys())
if qvms:
qc_names += ['9q-square-qvm', '9q-square-noisy-qvm']
return qc_names
def _parse_name(name: str, as_qvm: bool, noisy: bool) -> Tuple[str, str, bool]:
"""
Try to figure out whether we're getting a (noisy) qvm, and the associated qpu name.
See :py:func:`get_qc` for examples of valid names + flags.
"""
parts = name.split('-')
if len(parts) >= 2 and parts[-2] == 'noisy' and parts[-1] in ['qvm', 'pyqvm']:
if as_qvm is not None and (not as_qvm):
raise ValueError("The provided qc name indicates you are getting a noisy QVM, "
"but you have specified `as_qvm=False`")
if noisy is not None and (not noisy):
raise ValueError("The provided qc name indicates you are getting a noisy QVM, "
"but you have specified `noisy=False`")
qvm_type = parts[-1]
noisy = True
prefix = '-'.join(parts[:-2])
return prefix, qvm_type, noisy
if len(parts) >= 1 and parts[-1] in ['qvm', 'pyqvm']:
if as_qvm is not None and (not as_qvm):
raise ValueError("The provided qc name indicates you are getting a QVM, "
"but you have specified `as_qvm=False`")
qvm_type = parts[-1]
if noisy is None:
noisy = False
prefix = '-'.join(parts[:-1])
return prefix, qvm_type, noisy
if as_qvm is not None and as_qvm:
qvm_type = 'qvm'
else:
qvm_type = None
if noisy is None:
noisy = False
return name, qvm_type, noisy
def _canonicalize_name(prefix, qvm_type, noisy):
"""Take the output of _parse_name to create a canonical name.
"""
if noisy:
noise_suffix = '-noisy'
else:
noise_suffix = ''
if qvm_type is None:
qvm_suffix = ''
elif qvm_type == 'qvm':
qvm_suffix = '-qvm'
elif qvm_type == 'pyqvm':
qvm_suffix = '-pyqvm'
else:
raise ValueError(f"Unknown qvm_type {qvm_type}")
name = f'{prefix}{noise_suffix}{qvm_suffix}'
return name
def _get_qvm_or_pyqvm(qvm_type, connection, noise_model=None, device=None,
requires_executable=False):
if qvm_type == 'qvm':
return QVM(connection=connection, noise_model=noise_model,
requires_executable=requires_executable)
elif qvm_type == 'pyqvm':
return PyQVM(n_qubits=device.qubit_topology().number_of_nodes())
raise ValueError("Unknown qvm type {}".format(qvm_type))
def _get_qvm_qc(name: str, qvm_type: str, device: AbstractDevice, noise_model: NoiseModel = None,
requires_executable: bool = False,
connection: ForestConnection = None) -> QuantumComputer:
"""Construct a QuantumComputer backed by a QVM.
This is a minimal wrapper over the QuantumComputer, QVM, and QVMCompiler constructors.
:param name: A string identifying this particular quantum computer.
:param qvm_type: The type of QVM. Either qvm or pyqvm.
:param device: A device following the AbstractDevice interface.
:param noise_model: An optional noise model
:param requires_executable: Whether this QVM will refuse to run a :py:class:`Program` and
only accept the result of :py:func:`compiler.native_quil_to_executable`. Setting this
to True better emulates the behavior of a QPU.
:param connection: An optional :py:class:`ForestConnection` object. If not specified,
the default values for URL endpoints will be used.
:return: A QuantumComputer backed by a QVM with the above options.
"""
if connection is None:
connection = ForestConnection()
return QuantumComputer(name=name,
qam=_get_qvm_or_pyqvm(
qvm_type=qvm_type,
connection=connection,
noise_model=noise_model,
device=device,
requires_executable=requires_executable),
device=device,
compiler=QVMCompiler(
device=device,
endpoint=connection.compiler_endpoint))
def _get_qvm_with_topology(name: str, topology: nx.Graph,
noisy: bool = False,
requires_executable: bool = True,
connection: ForestConnection = None,
qvm_type: str = 'qvm') -> QuantumComputer:
"""Construct a QVM with the provided topology.
:param name: A name for your quantum computer. This field does not affect behavior of the
constructed QuantumComputer.
:param topology: A graph representing the desired qubit connectivity.
:param noisy: Whether to include a generic noise model. If you want more control over
the noise model, please construct your own :py:class:`NoiseModel` and use
:py:func:`_get_qvm_qc` instead of this function.
:param requires_executable: Whether this QVM will refuse to run a :py:class:`Program` and
only accept the result of :py:func:`compiler.native_quil_to_executable`. Setting this
to True better emulates the behavior of a QPU.
:param connection: An optional :py:class:`ForestConnection` object. If not specified,
the default values for URL endpoints will be used.
:param qvm_type: The type of QVM. Either 'qvm' or 'pyqvm'.
:return: A pre-configured QuantumComputer
"""
# Note to developers: consider making this function public and advertising it.
device = NxDevice(topology=topology)
if noisy:
noise_model = decoherence_noise_with_asymmetric_ro(gates=gates_in_isa(device.get_isa()))
else:
noise_model = None
return _get_qvm_qc(name=name, qvm_type=qvm_type, connection=connection, device=device,
noise_model=noise_model, requires_executable=requires_executable)
def _get_9q_square_qvm(name: str, noisy: bool,
connection: ForestConnection = None,
qvm_type: str = 'qvm') -> QuantumComputer:
"""
A nine-qubit 3x3 square lattice.
This uses a "generic" lattice not tied to any specific device. 9 qubits is large enough
to do vaguely interesting algorithms and small enough to simulate quickly.
:param name: The name of this QVM
:param connection: The connection to use to talk to external services
:param noisy: Whether to construct a noisy quantum computer
:param qvm_type: The type of QVM. Either 'qvm' or 'pyqvm'.
:return: A pre-configured QuantumComputer
"""
topology = nx.convert_node_labels_to_integers(nx.grid_2d_graph(3, 3))
return _get_qvm_with_topology(name=name, connection=connection,
topology=topology,
noisy=noisy,
requires_executable=True,
qvm_type=qvm_type)
def _get_unrestricted_qvm(name: str, noisy: bool,
n_qubits: int = 34,
connection: ForestConnection = None,
qvm_type: str = 'qvm') -> QuantumComputer:
"""
A qvm with a fully-connected topology.
This is obviously the least realistic QVM, but who am I to tell users what they want.
:param name: The name of this QVM
:param noisy: Whether to construct a noisy quantum computer
:param n_qubits: 34 qubits ought to be enough for anybody.
:param connection: The connection to use to talk to external services
:param qvm_type: The type of QVM. Either 'qvm' or 'pyqvm'.
:return: A pre-configured QuantumComputer
"""
topology = nx.complete_graph(n_qubits)
return _get_qvm_with_topology(name=name, connection=connection,
topology=topology,
noisy=noisy,
requires_executable=False,
qvm_type=qvm_type)
def _get_qvm_based_on_real_device(name: str, device: Device,
noisy: bool, connection: ForestConnection = None,
qvm_type: str = 'qvm'):
"""
A qvm with a based on a real device.
This is the most realistic QVM.
:param name: The full name of this QVM
:param device: The device from :py:func:`get_lattice`.
:param noisy: Whether to construct a noisy quantum computer by using the device's
associated noise model.
:param connection: An optional :py:class:`ForestConnection` object. If not specified,
the default values for URL endpoints will be used.
:return: A pre-configured QuantumComputer based on the named device.
"""
if noisy:
noise_model = device.noise_model
else:
noise_model = None
return _get_qvm_qc(name=name, connection=connection, device=device,
noise_model=noise_model, requires_executable=True,
qvm_type=qvm_type)
[docs]@_record_call
def get_qc(name: str, *, as_qvm: bool = None, noisy: bool = None,
connection: ForestConnection = None) -> QuantumComputer:
"""
Get a quantum computer.
A quantum computer is an object of type :py:class:`QuantumComputer` and can be backed
either by a QVM simulator ("Quantum/Quil Virtual Machine") or a physical Rigetti QPU ("Quantum
Processing Unit") made of superconducting qubits.
You can choose the quantum computer to target through a combination of its name and optional
flags. There are multiple ways to get the same quantum computer. The following are equivalent::
>>> qc = get_qc("Aspen-1-16Q-A-noisy-qvm")
>>> qc = get_qc("Aspen-1-16Q-A", as_qvm=True, noisy=True)
and will construct a simulator of an Aspen-1 lattice with a noise model based on device
characteristics. We also provide a means for constructing generic quantum simulators that
are not related to a given piece of Rigetti hardware::
>>> qc = get_qc("9q-square-qvm")
>>> qc = get_qc("9q-square", as_qvm=True)
Finally, you can get request a QVM with "no" topology of a given number of qubits
(technically, it's a fully connected graph among the given number of qubits) with::
>>> qc = get_qc("5q-qvm") # or "6q-qvm", or "34q-qvm", ...
These less-realistic, fully-connected QVMs will also be more lenient on what types of programs
they will ``run``. Specifically, you do not need to do any compilation. For the other, realistic
QVMs you must use :py:func:`qc.compile` or :py:func:`qc.compiler.native_quil_to_executable`
prior to :py:func:`qc.run`.
The Rigetti QVM must be downloaded from https://www.rigetti.com/forest and run as a server
alongside your python program. To use pyQuil's built-in QVM, replace all ``"-qvm"`` suffixes
with ``"-pyqvm"``::
>>> qc = get_qc("5q-pyqvm")
Redundant flags are acceptable, but conflicting flags will raise an exception::
>>> qc = get_qc("9q-square-qvm") # qc is fully specified by its name
>>> qc = get_qc("9q-square-qvm", as_qvm=True) # redundant, but ok
>>> qc = get_qc("9q-square-qvm", as_qvm=False) # Error!
Use :py:func:`list_quantum_computers` to retrieve a list of known qc names.
This method is provided as a convenience to quickly construct and use QVM's and QPU's.
Power users may wish to have more control over the specification of a quantum computer
(e.g. custom noise models, bespoke topologies, etc.). This is possible by constructing
a :py:class:`QuantumComputer` object by hand. Please refer to the documentation on
:py:class:`QuantumComputer` for more information.
:param name: The name of the desired quantum computer. This should correspond to a name
returned by :py:func:`list_quantum_computers`. Names ending in "-qvm" will return
a QVM. Names ending in "-pyqvm" will return a :py:class:`PyQVM`. Names ending in
"-noisy-qvm" will return a QVM with a noise model. Otherwise, we will return a QPU with
the given name.
:param as_qvm: An optional flag to force construction of a QVM (instead of a QPU). If
specified and set to ``True``, a QVM-backed quantum computer will be returned regardless
of the name's suffix
:param noisy: An optional flag to force inclusion of a noise model. If
specified and set to ``True``, a quantum computer with a noise model will be returned
regardless of the name's suffix. The noise model for QVMs based on a real QPU
is an empirically parameterized model based on real device noise characteristics.
The generic QVM noise model is simple T1 and T2 noise plus readout error. See
:py:func:`~pyquil.noise.decoherence_noise_with_asymmetric_ro`.
:param connection: An optional :py:class:`ForestConnection` object. If not specified,
the default values for URL endpoints will be used. If you deign to change any
of these parameters, pass your own :py:class:`ForestConnection` object.
:return: A pre-configured QuantumComputer
"""
# 1. Parse name, check for redundant options, canonicalize names.
prefix, qvm_type, noisy = _parse_name(name, as_qvm, noisy)
del as_qvm # do not use after _parse_name
name = _canonicalize_name(prefix, qvm_type, noisy)
# 2. Check for unrestricted {n}q-qvm
ma = re.fullmatch(r'(\d+)q', prefix)
if ma is not None:
n_qubits = int(ma.group(1))
if qvm_type is None:
raise ValueError("Please name a valid device or run as a QVM")
return _get_unrestricted_qvm(name=name, connection=connection,
noisy=noisy, n_qubits=n_qubits, qvm_type=qvm_type)
# 3. Check for "9q-square" qvm
if prefix == '9q-generic' or prefix == '9q-square':
if prefix == '9q-generic':
warnings.warn("Please prefer '9q-square' instead of '9q-generic'", DeprecationWarning)
if qvm_type is None:
raise ValueError("The device '9q-square' is only available as a QVM")
return _get_9q_square_qvm(name=name, connection=connection, noisy=noisy, qvm_type=qvm_type)
# 4. Not a special case, query the web for information about this device.
device = get_lattice(prefix)
if qvm_type is not None:
# 4.1 QVM based on a real device.
return _get_qvm_based_on_real_device(name=name, device=device,
noisy=noisy, connection=connection, qvm_type=qvm_type)
else:
# 4.2 A real device
pyquil_config = PyquilConfig()
if noisy is not None and noisy:
warnings.warn("You have specified `noisy=True`, but you're getting a QPU. This flag "
"is meant for controlling noise models on QVMs.")
return QuantumComputer(name=name,
qam=QPU(
endpoint=pyquil_config.qpu_url,
user=pyquil_config.user_id),
device=device,
compiler=QPUCompiler(
quilc_endpoint=pyquil_config.quilc_url,
qpu_compiler_endpoint=pyquil_config.qpu_compiler_url,
device=device,
name=prefix))
@contextmanager
def local_qvm() -> Iterator[Tuple[subprocess.Popen, subprocess.Popen]]:
"""A context manager for the Rigetti local QVM and QUIL compiler.
.. deprecated:: 2.11
Use py:func:`local_forest_runtime` instead.
"""
warnings.warn(DeprecationWarning("Use of pyquil.api.local_qvm has been deprecated.\n"
"Please use pyquil.api.local_forest_runtime instead."))
with local_forest_runtime() as (qvm, quilc):
yield (qvm, quilc)
def _port_used(host: str, port: int):
"""Check if a (TCP) port is listening.
:param host: Host address to check.
:param port: TCP port to check.
:returns: ``True`` if a process is listening on the specified host/port, ``False`` otherwise
"""
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
try:
s.connect((host, port))
return True
except ConnectionRefusedError:
return False
finally:
s.close()
[docs]@contextmanager
def local_forest_runtime(
*,
host: str = '127.0.0.1',
qvm_port: int = 5000,
quilc_port: int = 5555,
use_protoquil: bool = False) -> Iterator[Tuple[subprocess.Popen, subprocess.Popen]]:
"""A context manager for local QVM and QUIL compiler.
You must first have installed the `qvm` and `quilc` executables from
the forest SDK. [https://www.rigetti.com/forest]
This context manager will ensure that the designated ports are not used, start up `qvm` and
`quilc` proccesses if possible and terminate them when the context is exited.
If one of the ports is in use, a ``RuntimeWarning`` will be issued and the `qvm`/`quilc` process
won't be started.
.. note::
Only processes started by this context manager will be terminated on exit, no external process will
be touched.
>>> from pyquil import get_qc, Program
>>> from pyquil.gates import CNOT, Z
>>> from pyquil.api import local_forest_runtime
>>>
>>> qvm = get_qc('9q-square-qvm')
>>> prog = Program(Z(0), CNOT(0, 1))
>>>
>>> with local_forest_runtime():
>>> results = qvm.run_and_measure(prog, trials=10)
:param host: Host on which `qvm` and `quilc` should listen on.
:param qvm_port: Port which should be used by `qvm`.
:param quilc_port: Port which should be used by `quilc`.
:param use_protoquil: Restrict input/output to protoquil.
.. warning::
If ``use_protoquil`` is set to ``True`` language features you need
may be disabled. Please use it with caution.
:raises: FileNotFoundError: If either executable is not installed.
:returns: The returned tuple contains two ``subprocess.Popen`` objects
for the `qvm` and the `quilc` processes. If one of the designated
ports is in use, the process won't be started and the respective
value in the tuple will be ``None``.
"""
qvm = None
quilc = None
# If the host we should listen to is 0.0.0.0, we replace it
# with 127.0.0.1 to use a valid IP when checking if the port is in use.
if _port_used(host if host != '0.0.0.0' else '127.0.0.1', qvm_port):
warning_msg = ("Unable to start qvm server, since the specified "
"port {} is in use.").format(qvm_port)
warnings.warn(RuntimeWarning(warning_msg))
else:
qvm_cmd = ['qvm', '-S', '--host', host, '-p', str(qvm_port)]
qvm = subprocess.Popen(qvm_cmd,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
if _port_used(host if host != '0.0.0.0' else '127.0.0.1', quilc_port):
warning_msg = ("Unable to start quilc server, since the specified "
"port {} is in use.").format(quilc_port)
warnings.warn(RuntimeWarning(warning_msg))
else:
quilc_cmd = ['quilc', '--host', host, '-p', str(quilc_port), '-R']
if use_protoquil:
quilc_cmd += ['-P']
quilc = subprocess.Popen(quilc_cmd,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
# Return context
try:
yield (qvm, quilc)
finally:
# Exit. Release resource
if qvm:
qvm.terminate()
if quilc:
quilc.terminate()
def _flip_array_to_prog(flip_array: Tuple[bool], qubits: List[int]) -> Program:
"""
Generate a pre-measurement program that flips the qubit state according to the flip_array of
bools.
This is used, for example, in symmetrization to produce programs which flip a select subset
of qubits immediately before measurement.
:param flip_array: tuple of booleans specifying whether the qubit in the corresponding index
should be flipped or not.
:param qubits: list specifying the qubits in order corresponding to the flip_array
:return: Program which flips each qubit (i.e. instructs RX(pi, q)) according to the flip_array.
"""
assert len(flip_array) == len(qubits), "Mismatch of qubits and operations"
prog = Program()
for qubit, flip_output in zip(qubits, flip_array):
if flip_output == 0:
continue
elif flip_output == 1:
prog += Program(RX(pi, qubit))
else:
raise ValueError("flip_bools should only consist of 0s and/or 1s")
return prog
def _symmetrization(program: Program, meas_qubits: List[int], symm_type: int = 3) \
-> Tuple[List[Program], List[np.ndarray]]:
"""
For the input program generate new programs which flip the measured qubits with an X gate in
certain combinations in order to symmetrize readout.
An expanded list of programs is returned along with a list of bools which indicates which
qubits are flipped in each program.
The symmetrization types are specified by an int; the types available are:
-1 -- exhaustive symmetrization uses every possible combination of flips
0 -- trivial that is no symmetrization
1 -- symmetrization using an OA with strength 1
2 -- symmetrization using an OA with strength 2
3 -- symmetrization using an OA with strength 3
In the context of readout symmetrization the strength of the orthogonal array enforces the
symmetry of the marginal confusion matrices.
By default a strength 3 OA is used; this ensures expectations of the form <b_k * b_j * b_i>
for bits any bits i,j,k will have symmetric readout errors. Here expectation of a random
variable x as is denote <x> = sum_i Pr(i) x_i. It turns out that a strength 3 OA is also a
strength 2 and strength 1 OA it also ensures <b_j * b_i> and <b_i> have symmetric readout
errors for any bits b_j and b_i.
:param programs: a program which will be symmetrized.
:param meas_qubits: the groups of measurement qubits. Only these qubits will be symmetrized
over, even if the program acts on other qubits.
:param sym_type: an int determining the type of symmetrization performed.
:return: a list of symmetrized programs, the corresponding array of bools indicating which
qubits were flipped.
"""
if symm_type < -1 or symm_type > 3:
raise ValueError("symm_type must be one of the following ints [-1, 0, 1, 2, 3].")
elif symm_type == -1:
# exhaustive = all possible binary strings
flip_matrix = np.asarray(list(itertools.product([0, 1], repeat=len(meas_qubits))))
elif symm_type >= 0:
flip_matrix = _construct_orthogonal_array(len(meas_qubits), symm_type)
# The next part is not rigorous in the sense that we simply truncate to the desired
# number of qubits. The problem is that orthogonal arrays of a certain strength for an
# arbitrary number of qubits are not known to exist.
flip_matrix = flip_matrix[:, :len(meas_qubits)]
symm_programs = []
flip_arrays = []
for flip_array in flip_matrix:
total_prog_symm = program.copy()
prog_symm = _flip_array_to_prog(flip_array, meas_qubits)
total_prog_symm += prog_symm
symm_programs.append(total_prog_symm)
flip_arrays.append(flip_array)
return symm_programs, flip_arrays
def _consolidate_symmetrization_outputs(outputs: List[np.ndarray],
flip_arrays: List[Tuple[bool]]) -> np.ndarray:
"""
Given bitarray results from a series of symmetrization programs, appropriately flip output
bits and consolidate results into new bitarrays.
:param outputs: a list of the raw bitarrays resulting from running a list of symmetrized
programs; for example, the results returned from _measure_bitstrings
:param flip_arrays: a list of boolean arrays in one-to-one correspondence with the list of
outputs indicating which qubits where flipped before each bitarray was measured.
:return: an np.ndarray consisting of the consolidated bitarray outputs which can be treated as
the symmetrized outputs of the original programs passed into a symmetrization method. See
estimate_observables for example usage.
"""
assert len(outputs) == len(flip_arrays)
output = []
for bitarray, flip_array in zip(outputs, flip_arrays):
if len(flip_array) == 0:
output.append(bitarray)
else:
output.append(bitarray ^ flip_array)
return np.vstack(output)
def _measure_bitstrings(qc, programs: List[Program], meas_qubits: List[int],
num_shots: int = 600) -> List[np.ndarray]:
"""
Wrapper for appending measure instructions onto each program, running the program,
and accumulating the resulting bitarrays.
:param qc: a quantum computer object on which to run each program
:param programs: a list of programs to run
:param meas_qubits: groups of qubits to measure for each program
:param num_shots: the number of shots to run for each program
:return: a len(programs) long list of num_shots by num_meas_qubits bit arrays of results for
each program.
"""
results = []
for program in programs:
# copy the program so the original is not mutated
prog = program.copy()
ro = prog.declare('ro', 'BIT', len(meas_qubits))
for idx, q in enumerate(meas_qubits):
prog += MEASURE(q, ro[idx])
prog.wrap_in_numshots_loop(num_shots)
prog = qc.compiler.quil_to_native_quil(prog)
exe = qc.compiler.native_quil_to_executable(prog)
shots = qc.run(exe)
results.append(shots)
return results
def _construct_orthogonal_array(num_qubits: int, strength: int = 3) -> np.ndarray:
"""
Given a strength and number of qubits this function returns an Orthogonal Array (OA)
on 'n' or more qubits. Sometimes the size of the returned array is larger than num_qubits;
typically the next power of two relative to num_qubits. This is corrected later in the code
flow.
:param num_qubits: the minimum number of qubits the OA should act on.
:param strength: the statistical "strength" of the OA
:return: a numpy array where the rows represent the different experiments
"""
if strength < 0 or strength > 3:
raise ValueError("'strength' must be one of the following ints [0, 1, 2, 3].")
if strength == 0:
# trivial flip matrix = an array of zeros
flip_matrix = np.zeros((1, num_qubits)).astype(int)
elif strength == 1:
# orthogonal array with strength equal to 1. See Example 1.4 of [OATA], referenced in the
# `construct_strength_two_orthogonal_array` docstrings, for more details.
zero_array = np.zeros((1, num_qubits))
one_array = np.ones((1, num_qubits))
flip_matrix = np.concatenate((zero_array, one_array), axis=0).astype(int)
elif strength == 2:
flip_matrix = _construct_strength_two_orthogonal_array(num_qubits)
elif strength == 3:
flip_matrix = _construct_strength_three_orthogonal_array(num_qubits)
return flip_matrix
def _next_power_of_2(x):
return 1 if x == 0 else 2 ** (x - 1).bit_length()
# The code below is directly copied from scipy see https://bit.ly/2RjAHJz, the docstrings have
# been modified.
def hadamard(n, dtype=int):
"""
Construct a Hadamard matrix.
Constructs an n-by-n Hadamard matrix, using Sylvester's
construction. `n` must be a power of 2.
Parameters
----------
n : int
The order of the matrix. `n` must be a power of 2.
dtype : numpy dtype
The data type of the array to be constructed.
Returns
-------
H : (n, n) ndarray
The Hadamard matrix.
Notes
-----
.. versionadded:: 0.8.0
Examples
--------
>>> hadamard(2, dtype=complex)
array([[ 1.+0.j, 1.+0.j],
[ 1.+0.j, -1.-0.j]])
>>> hadamard(4)
array([[ 1, 1, 1, 1],
[ 1, -1, 1, -1],
[ 1, 1, -1, -1],
[ 1, -1, -1, 1]])
"""
if n < 1:
lg2 = 0
else:
lg2 = int(log(n, 2))
if 2 ** lg2 != n:
raise ValueError("n must be an positive integer, and n must be "
"a power of 2")
H = np.array([[1]], dtype=dtype)
# Sylvester's construction
for i in range(0, lg2):
H = np.vstack((np.hstack((H, H)), np.hstack((H, -H))))
return H
def _construct_strength_three_orthogonal_array(num_qubits: int) -> np.ndarray:
r"""
Given a number of qubits this function returns an Orthogonal Array (OA)
on 'n' qubits where n is the next power of two relative to num_qubits.
Specifically it returns the OA(2n, n, 2, 3).
The parameters of the OA(N, k, s, t) are interpreted as
N: Number of rows, level combinations or runs
k: Number of columns, constraints or factors
s: Number of symbols or levels
t: Strength
See [OATA] for more details.
[OATA] Orthogonal Arrays: theory and applications
Hedayat, Sloane, Stufken
Springer Science & Business Media, 2012.
https://dx.doi.org/10.1007/978-1-4612-1478-6
:param num_qubits: minimum number of qubits the OA should run on.
:return: A numpy array representing the OA with shape N by k
"""
num_qubits_power_of_2 = _next_power_of_2(num_qubits)
H = hadamard(num_qubits_power_of_2)
Hfold = np.concatenate((H, -H), axis=0)
orthogonal_array = ((Hfold + 1) / 2).astype(int)
return orthogonal_array
def _construct_strength_two_orthogonal_array(num_qubits: int) -> np.ndarray:
r"""
Given a number of qubits this function returns an Orthogonal Array (OA) on 'n-1' qubits
where n-1 is the next integer lambda so that 4*lambda -1 is larger than num_qubits.
Specifically it returns the OA(n, n − 1, 2, 2).
The parameters of the OA(N, k, s, t) are interpreted as
N: Number of rows, level combinations or runs
k: Number of columns, constraints or factors
s: Number of symbols or levels
t: Strength
See [OATA] for more details.
[OATA] Orthogonal Arrays: theory and applications
Hedayat, Sloane, Stufken
Springer Science & Business Media, 2012.
https://dx.doi.org/10.1007/978-1-4612-1478-6
:param num_qubits: minimum number of qubits the OA should run on.
:return: A numpy array representing the OA with shape N by k
"""
# next line will break post denali at 275 qubits
# valid_num_qubits = 4 * lambda - 1
valid_numbers = [4 * lam - 1 for lam in range(1, 70)]
# 4 * lambda
four_lam = min(x for x in valid_numbers if x >= num_qubits) + 1
H = hadamard(_next_power_of_2(four_lam))
# The minus sign in front of H fixes the 0 <-> 1 inversion relative to the reference [OATA]
orthogonal_array = ((-H[1:, :].T + 1) / 2).astype(int)
return orthogonal_array
def _check_min_num_trials_for_symmetrized_readout(num_qubits: int, trials: int, symm_type: int) \
-> int:
"""
This function sets the minimum number of trials; it is desirable to have hundreds or
thousands of trials more than the minimum.
:param num_qubits: number of qubits to symmetrize
:param trials: number of trials
:param symm_type: symmetrization type see
:return: possibly modified number of trials
"""
if symm_type < -1 or symm_type > 3:
raise ValueError("symm_type must be one of the following ints [-1, 0, 1, 2, 3].")
if symm_type == -1:
min_num_trials = 2 ** num_qubits
elif symm_type == 2:
def _f(x):
return 4 * x - 1
min_num_trials = min(_f(x) for x in range(1, 1024) if _f(x) >= num_qubits) + 1
elif symm_type == 3:
min_num_trials = _next_power_of_2(2 * num_qubits)
else:
# symm_type == 0 or symm_type == 1 require one and two trials respectively; ensured by:
min_num_trials = 2
if trials < min_num_trials:
trials = min_num_trials
warnings.warn(f"Number of trials was too low, it is now {trials}.")
return trials
```