# This code is part of Qiskit.
#
# (C) Copyright IBM 2021.
#
# This code is licensed under the Apache License, Version 2.0. You may
# obtain a copy of this license in the LICENSE.txt file in the root directory
# of this source tree or at http://www.apache.org/licenses/LICENSE-2.0.
#
# Any modifications or derivative works of this code must retain this
# copyright notice, and modified files need to carry a notice indicating
# that they have been altered from the originals.
"""
Batch Experiment class.
"""
from typing import List, Optional, Dict
from collections import OrderedDict, defaultdict
import warnings
from qiskit import QuantumCircuit
from qiskit.providers import Job, Backend, Options
from .composite_experiment import CompositeExperiment, BaseExperiment
from .composite_analysis import CompositeAnalysis
[docs]
class BatchExperiment(CompositeExperiment):
"""Combine multiple experiments into a batch experiment.
Batch experiments combine individual experiments on any subset of qubits
into a single composite experiment which appends all the circuits from
each component experiment into a single batch of circuits to be executed
as one experiment job.
Analysis of batch experiments is performed using the
:class:`~qiskit_experiments.framework.CompositeAnalysis` class which handles
sorting the composite experiment circuit data into individual child
:class:`ExperimentData` containers for each component experiment which are
then analyzed using the corresponding analysis class for that component
experiment.
See :class:`~qiskit_experiments.framework.CompositeAnalysis`
documentation for additional information.
"""
def __init__(
self,
experiments: List[BaseExperiment],
backend: Optional[Backend] = None,
flatten_results: bool = None,
analysis: Optional[CompositeAnalysis] = None,
experiment_type: Optional[str] = None,
):
"""Initialize a batch experiment.
Args:
experiments: a list of experiments.
backend: Optional, the backend to run the experiment on.
flatten_results: If True flatten all component experiment results
into a single ExperimentData container, including
nested composite experiments. If False save each
component experiment results as a separate child
ExperimentData container. This kwarg is ignored
if the analysis kwarg is used.
analysis: Optional, the composite analysis class to use. If not
provided this will be initialized automatically from the
supplied experiments.
"""
if flatten_results is None:
# Backward compatibility for 0.6
# This if-clause will be removed in 0.7 and flatten_result=True is set in arguments.
warnings.warn(
"Default value of flatten_results will be turned to True in Qiskit Experiments 0.7. "
"If you want child experiment data for each subset experiment, "
"set 'flatten_results=False' explicitly.",
DeprecationWarning,
)
flatten_results = False
# Generate qubit map
self._qubit_map = OrderedDict()
logical_qubit = 0
for expr in experiments:
for physical_qubit in expr.physical_qubits:
if physical_qubit not in self._qubit_map:
self._qubit_map[physical_qubit] = logical_qubit
logical_qubit += 1
qubits = tuple(self._qubit_map.keys())
super().__init__(
experiments,
qubits,
backend=backend,
analysis=analysis,
flatten_results=flatten_results,
experiment_type=experiment_type,
)
[docs]
def circuits(self):
return self._batch_circuits(to_transpile=False)
def _transpiled_circuits(self):
return self._batch_circuits(to_transpile=True)
def _batch_circuits(self, to_transpile=False):
batch_circuits = []
# Generate data for combination
for index, expr in enumerate(self._experiments):
if self.physical_qubits == expr.physical_qubits or to_transpile:
qubit_mapping = None
else:
qubit_mapping = [self._qubit_map[qubit] for qubit in expr.physical_qubits]
if isinstance(expr, BatchExperiment):
# Batch experiments don't contain their own native circuits.
# If to_transpile is True then the circuits will be transpiled at the non-batch
# experiments.
# Fetch the circuits from the sub-experiments.
expr_circuits = expr._batch_circuits(to_transpile)
elif to_transpile:
expr_circuits = expr._transpiled_circuits()
else:
expr_circuits = expr.circuits()
for circuit in expr_circuits:
# Update metadata
circuit.metadata = {
"experiment_type": self._type,
"composite_metadata": [circuit.metadata],
"composite_index": [index],
}
# Remap qubits if required
if qubit_mapping:
circuit = self._remap_qubits(circuit, qubit_mapping)
batch_circuits.append(circuit)
return batch_circuits
def _remap_qubits(self, circuit, qubit_mapping):
"""Remap qubits if physical qubit layout is different to batch layout"""
num_qubits = self.num_qubits
num_clbits = circuit.num_clbits
new_circuit = QuantumCircuit(num_qubits, num_clbits, name="batch_" + circuit.name)
new_circuit.metadata = circuit.metadata
new_circuit.append(circuit, qubit_mapping, list(range(num_clbits)))
return new_circuit
def _run_jobs_recursive(
self, circuits: List[QuantumCircuit], truncated_metadata: List[Dict], **run_options
) -> List[Job]:
# The truncated metadata is a truncation of the original composite metadata.
# During the recursion, the current experiment (self) will be at the head of the truncated
# metadata.
if self.experiment_options.separate_jobs:
# A dictionary that maps sub-experiments to their circuits
circs_by_subexps = defaultdict(list)
for circ_iter, (circ, tmd) in enumerate(zip(circuits, truncated_metadata)):
# For batch experiments the composite index is always a list of length 1,
# because unlike parallel experiment, each circuit originates from a single
# sub-experiment.
circ_index = tmd["composite_index"][0]
circs_by_subexps[circ_index].append(circ)
# For batch experiments the composite metadata is always a list of length 1,
# because unlike parallel experiment, each circuit originates from a single
# sub-experiment.
truncated_metadata[circ_iter] = tmd["composite_metadata"][0]
jobs = []
for index, exp in enumerate(self.component_experiment()):
# Currently all the sub-experiments must use the same set of run options,
# even if they run in different jobs
if isinstance(exp, BatchExperiment):
new_jobs = exp._run_jobs_recursive(
circs_by_subexps[index], truncated_metadata, **run_options
)
else:
new_jobs = exp._run_jobs(circs_by_subexps[index], **run_options)
jobs.extend(new_jobs)
else:
jobs = super()._run_jobs(circuits, **run_options)
return jobs
def _run_jobs(self, circuits: List[QuantumCircuit], **run_options) -> List[Job]:
truncated_metadata = [circ.metadata for circ in circuits]
jobs = self._run_jobs_recursive(circuits, truncated_metadata, **run_options)
return jobs
@classmethod
def _default_experiment_options(cls) -> Options:
"""Default experiment options.
Experiment Options:
separate_jobs (Boolean): Whether to route different sub-experiments to different jobs.
"""
options = super()._default_experiment_options()
options.separate_jobs = False
return options