Quantum Instructions in Parallel (QuIP)¶
Gates¶

class
trueq.
Gate
(u)¶ Represents a unitary quantum gate.
import trueq as tq x = tq.Gate([[0, 1], [1, 0]]) cz = tq.Gate([[1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0], [0, 0, 0, 1]])
Note
Gates are not checked to be unitary on instantiation; call
Gate.is_unitary
manually to check.Many standard gates can be instantiatied through convenience attributes:
tq.Gate.id tq.Gate.x tq.Gate.y tq.Gate.z tq.Gate.h tq.Gate.s tq.Gate.t tq.Gate.cx tq.Gate.cy tq.Gate.cz tq.Gate.swap tq.Gate.iswap
The static method
from_generators()
can be used to construct gates of the form \(exp(1j * \sum_i (\pi\theta_i/180) P_i/ 2)\) where \(P_i\) are the input Pauli strings and \(\theta_i\) are the input rotation angles in degrees. This can be input either as an alternating sequence of Pauli strings and angles, or as a dictionary.g1 = tq.Gate.from_generators("X", 90) g2 = tq.Gate.from_generators("XX", 90, "YY", 180) g3 = tq.Gate.from_generators({"XX": 90, "YY": 180})
Finally, the constructor also allows gates to be entered as coefficients on the pauli basis (rather than rotation angles) as follows.
g1 = tq.Gate({"I": np.sqrt(1/2), "X": 1j * np.sqrt(1/2)}) g2 = tq.Gate({"II": np.sqrt(1/2), "XY": 1j * np.sqrt(1/2)})
 Parameters
u (
numpy.ndarray
like dict
) – Either a unitary matrix in the computational basis, or a dictionary specifying coefficients of the pauli basis; see examples above.

static
from_generators
(*args)¶ Constructs a
Gate
instance from rotations about Paulis. Specifically, the gate is constructed as \(exp(1j * \sum_i (\pi\theta_i/180) P_i/ 2)\) where \(P_i\) are the input Pauli strings and \(\theta_i\) are the input rotation angles in degrees. This can be input either as an alternating sequence of Pauli strings and angles, or as adict
.g1 = tq.Gate.from_generators("X", 90) g2 = tq.Gate.from_generators("XX", 90, "YY", 180) g3 = tq.Gate.from_generators({"XX": 90, "YY": 180})
 Parameters
args – A dictionary mapping Pauli strings to angles in degrees, or an alternating sequence thereof.
 Return type

property
is_unitary
¶ Whether or not this gate is unitary.
 Type
bool

property
is_identity
¶ Whether or not this gate is the identity gate.
 Type
bool

property
mat
¶ The matrix representation of this gate in the canonical basis.
 Type
numpy.ndarray

property
width
¶ The number of columns of the matrix representation of this gate.
 Type
int

property
expansion
¶ Assuming this gate acts on qubits, a sparse representation of this gate in terms of the
trueq.math.pauli_basis
; a dictionary mapping Pauli strings to (complex) coefficients. Type
dict

static
from_alias
(alias)¶ Returns a
Gate
from common names, such as CX, swap, H, S etc.ALIASES
is a dictionary which stores available gates. Raises
KeyError – If the gate is not found.

property
generators
¶ Assumes this gate acts on qubits and returns rotations about Pauli axes performed by this
Gate
. Specifically, outputs a description of this gate in the form: \(exp(1j * \sum_i (\pi\theta_i/180) P_i/ 2)\) where \(P_i\) are the Pauli strings and \(\theta_i\) are the rotation angles in degrees.g1 = tq.Gate.from_generators("X", 90).generators {"X": 90} g2 = tq.Gate.from_generators("X", 90, "Z", 20).generators {"X": 90, "Z": 20}
 Type
dict

static
random
(dim)¶ Generate a Haar random unitary gate of dimension
dim
. Parameters
dim (
int
) – The dimension of the gate. Return type

plot
(abs_max=None, ax=None)¶ Plots the matrix representation of this gate.
 Parameters
abs_max (
None
float
) – The value to scale absolute values of the matrix by; the value at which plotted colors become fully opaque. By default, this is the largest absolute magnitude of the input matrix.ax (
matplotlib.Axis
) – An existing axis to plot on. If not given, a new figure will be created.
Native Gates¶

class
trueq.
NativeGate
(name, u, params=None, classname=None)¶ A
Gate
like object that also contains a name and parameters used on a physical system.This class is a thin layer on top of
Gate
to additionally store gate names and parameters used to construct the gate by the physical system. From the perspective of aNativeGate
, this information is static throughout its lifetime and treated as metadata.Native gates are typically (and most easily) constructed by
trueq.config.GateFactory
’s, which are usually owned bytrueq.Config
objects.>>> conf = Config(file) >>> gate = conf.x(phi=90) >>> gate NativeGate.conf_name.x(phi=90) >>> gate.parameters {"phi": 90} >>> gate.name x
Native gates are not intended to be mutable, and the parameters and names should not be changed after instantiation.
 Parameters
name (
str
) – Name of the gate operation.u (
dict
) – SeeGate
.params (
dict
) – Any parameters that the gate needs in order to be constructed.classname (
str
) – This is used for recordkeeping in the trueq.config.GateFactory.

static
from_generators
(name, *args, params=None, classname=None)¶ Constructs a
NativeGate
instance from rotations about Paulis. Also adds a name and any parameters used to construct the native gate. Specifically, the output gate in this case is given by \(exp(1j * \sum_i (\pi\theta_i/180) P_i/ 2)\) where \(P_i\) are the input Pauli strings and \(\theta_i\) are the input rotation angles in degrees. This can be input either as an alternating sequence of Pauli strings and angles, or as adict
.g1 = tq.NativeGate.from_generators("name", {}, "X", 90) g2 = tq.NativeGate.from_generators("name", {}, "XX", 90, "YY", 180) g3 = tq.NativeGate.from_generators("name", {}, {"XX": 90, "YY": 180})
 Parameters
name (
str
) – Name of the gate operation.params (
dict
) – Any parameters that the gate needs in order to be constructed.args – A dictionary mapping Pauli strings to angles in degrees, or an alternating sequence thereof.
classname (
str
) – This is used for recordkeeping in the trueq.config.GateFactory.
 Return type

property
parameters
¶ The stored parameters.
 Type
dict

property
name
¶ The name of the
NativeGate
. Type
str
Cycles¶

class
trueq.
Cycle
(content=None, immutable=None)¶ Represents a “clock cycle” of a
Circuit
; a set ofGate
s that happen in parallel to a disjoint set of systems. The storage format is adict
that mapstuple
s of system labels toGate
objects. Any given system can only be addressed once in aCycle
.This class is iterable, yielding tuples of the form
(labels, gate)
.Note
We are agnostic to the physical duration of a cycle, and different cycles in a circuit do not need to have the same duration.
 Parameters
content (
dict
) – Adict
whose keys aretuple
s and whose values areOperators
s; the methodadd()
is called on each pair.immutable (
bool
) – If true, the contents of thisCycle
cannot be mutated by tools such asrandomly_compile()
.

add
(labels, operator)¶ Adds a
Operator
to thisCycle
. In particular, this can be aGate
,NativeGate
,Prep
, orMeas
.As a convenience, if the operator acts on one qudit (such as single qubit gates,
Meas
, orPrep
), then this operation is added to each label, even when multiple labels are provided. Parameters
labels – Which systems this gate targets. Note that the order of labels is relevant for multisystem gates. For example, it controls the direction of a CX.
operator – A
Operator
object or a gate alias.

property
dim
¶ The dimension of each subsystem, e.g.
2
if this cycle acts on qubits. However, if no gates have been added to this cycle, then the dimension is unknown and this value is equal toNone
. Type
int
None

property
immutable
¶ Whether or not this cycle is immutable.
 Type
bool

property
n_operators
¶ The total number of operators in this
Cycle
, equal to the sum ofn_gates
,n_meas
, andn_prep
. Type
int

property
gates_single
¶ The singlequdit gates of this cycle; a
dict
ofGate
s in thisCycle
that act on one subsystem. Type
dict

property
gates_multi
¶ The multiqudit gates of this cycle; a
dict
ofGate
s in thisCycle
that act on at least two subsystems. Type
dict

property
gate_labels
¶ A sorted tuple of all labels in the cycle that have gates (i.e. labels that point to a
Gate
operation). Type
list

property
prep_labels
¶ A sorted tuple of all labels in the cycle that are a preparation (i.e. labels that point to a
Prep
operation). Type
list

property
meas_labels
¶ A sorted tuple of all labels in the cycle that are a preparation (i.e. labels that point to a
Meas
operation). Type
list

property
block_labels
¶ A sorted tuple of all labels in the cycle that are blocked (i.e. labels that point to a
Block
operation). Type
list

tensor
(labels=None)¶ Returns a
Tensor
representation of theGate
s in thisCycle
.The matrix representation (which will be exponentially sized in the number of systems!) of this cycle can be obtained with:
cycle.tensor().mat
 Parameters
labels (
list
) – An optional list of labels to force into the tensor in case they are not already present as labels in this cycle. Return type

is_equivalent
(other, ignore_imm=True, ignore_id=True)¶ Checks if this cycle is equivalent to another cycle.
 Parameters
 Return type
bool

to_dict
()¶ Returns a dictionary representation of a
Cycle
object. This dictionary representation contains only base Python classes. Return type
dict

pop
(item)¶ Deletes the operator from this cycle that acts on the given label tuple and returns it.
cycle = tq.Cycle({(0,): tq.Gate.x, (1,): tq.Meas()}) cycle.pop((0,)) # tq.Gate.x
 Parameters
item (
tuple
) – A tuple of qudit labels. Raises
KeyError – If no operation acts on the given label tuple.
ValueError – If the cycle is immutable.
Circuits¶

class
trueq.
Circuit
(cycles=None, key=None, results=None)¶ Represents a quantum
Circuit
to be performed on a device, along with relevant additional information about the circuit.Note
All TrueQᵀᴹ circuit generation and analysis methods assume that circuits are run by initializing systems independently in the ground state, applying the parallelized gates (referred to as cycles) obtained by iterating over the circuit, measuring the systems in the computational basis, and populating
results
with the results.Specifically, all local gates required to rotate preparations and measurements are already included in
cycles
.A circuit is stored as a list of
Cycle
s. It also has private attributes for ease of use with randomized compiling and standard quantum algorithms. Parameters
cycles (
list
) –Cycle
s of the circuit. This is a list of the form, for example,[{(1,): tq.Gate.x, (0, 2): tq.Gate.cz}, ...]
.key (
Key
) – A hashable object used to group similar circuits for analysis. For example,Key(protocol=SRB,n_random_cycles=10)
.results (
Results
dict
) – Results to add to the circuit on construction, see alsoCircuit.results
.

measure_all
()¶ Add a cycle of measurements at the end of the current circuit.
This adds
trueq.Meas
onto each label currently in the circuit.

property
dim
¶ The dimension of each subsystem, e.g.
2
if this circuit acts on qubits. This property is inherited from its cycles. If no cycles specify a dimension, then the dimension is unknown and this value is equal toNone
. Type
int
None

property
key
¶ Metadata related to this circuit. This is used in analysis methods of protocols, and for organizing circuits within
CircuitCollection
s. Type
Key

property
labels
¶ The sorted union of all system labels acted on by gates in this
Circuit
. Type
tuple

property
results
¶ Results from running this
Circuit
on a quantum device; a dictionary whose keys are bitstrings, and whose values are total observations of corresponding bitstrings.Bitstrings are python strings containing \(0\) and \(1\), e.g.,
'0110'
, see Recording Experimental Results.The initial value of this property is
{}
, which implies thathas_results
isFalse
.Note that these values will internally be normalized to their sum, and not to \(1\) —for example, with respect to all of our analysis routines, the following results are identical:
{'0': 123, '1': 4} {'0': 1230, '1': 40} {'0': 12300, '1': 400}
 Type
dict

expectation_values
(labels=None)¶ Finds the expectation value(s) of the
Circuit
, marginalized over specified labels.The labels can be a list of labels, in which case each label is separately marginalized over. For example,
labels = [0, 1, 2]
shows the expectation value for qubits 0, 1, and 2 separately. The labels can also be a list of lists of labels; each list will be separately marginalized over. For example,labels = [[0], [1], [2]]
is equivalent tolabels = [0, 1, 2]
, butlabels = [[0, 1], [2]]
will report the expectation value of the subregister[0, 1]
and the expectation value of the subregister 2.If labels are not provided, they are taken from
key.twirl
if it exists. Parameters
labels (
list
) – A list of labels (or list of lists of labels) specifying which which group(s) of subregisters to marginalize over. Return type
A
KeyMap
whose keys are grouped based on labels, and whose values are lists of expectation values for those labels.

add_cycle
(cycles)¶ Adds a
Cycle
to thisCircuit
and updates its labels with those found in theCycle
. Parameters
cycles –
Cycle
: A Cycle or iterable of Cycles.

property
meas_locs
¶ Returns a list of the locations of all measurements in the Circuit.
The ordering of this list is the same as the expected bitstring order in the circuit’s
Result
s.The order of measurements is decided first by the index of the cycle, e.g.: measurements placed in the first cycle of a circuit are returned before measurements in later cycles. Then the order inside of a cycle is decided by the label on which the measurement is placed, e.g.: measurements placed on label
(0,)
comes before label(5,)
The returned list contains tuples, where the first entry of tuple is the cycle index, and the second entry of the tuple is the label the measurement is on.
Examples:
circ = Circuit([{(0,): Meas()}, {(5,): Meas(), (1,): Meas()}]) circ.meas_locs # [(0, 0), (1, 1), (1, 5)] # Cycle 0, Qubit 0, Cycle 1, Qubit 1 and 5
 Return type
list
oftuple

property
n_meas
¶ The total number of measurement operations in the circuit.
 Type
int

to_dict
(include_cycles=True)¶ Returns a dictionary representation of a
Circuit
object. This dictionary representation contains only base Python classes. Parameters
include_cycles (
bool
) – Whether to include the cycles in the dictionary representation. Return type
dict

to_cirq
(metadata=None, config='ZXZ', device=None, join_meas=True)¶ Converts a TrueQᵀᴹ
Circuit
into an Cirq Circuit object.import cirq import trueq as tq mat = cirq.ops.FSimGate(theta=5, phi=1)._unitary_() fsim_factory = tq.config.GateFactory("fsim", matrix=mat) x90_factory = tq.config.GateFactory("x", hamiltonian=[["X", 90]]) z_factory = tq.config.GateFactory("z", hamiltonian=[["Z", "phi"]]) config = tq.Config.from_params( name="example", n_sys=2, mode="ZXZXZ", factories=[z_factory, x90_factory, fsim_factory], ) circ = tq.Circuit([{(0, 1): tq.Gate.cx}]) tq.interface.cirq.from_trueq_circ(circ, config=config)
Note
If nothing else is provided, this automatically decomposes single qubit operations into \(R_Z(\theta)R_X(\phi)R_Z(\gamma)\) single qubit gates.
 Parameters
circuit (
Circuit
) – A TrueQᵀᴹ circuit to be converted into an Cirq Circuitmetadata (
trueq.interface.metadata.CirqMetadata
) – Metadata required to accurately reproduce the original cirq circuit. If this is not provided, measurements are added at the end of the circuit.config (
Config
orstr
) – If aConfig
which contains gates which are directly decomposable into Cirq native gates is provided, then conversion will take place into gates specified by that config. Additionallyconfig
also accepts text strings describing decomposition methods, strings must be of paulis which are defined intrueq.config.decomposition.QubitMode
, thenGate
s are decomposed into all allowed Cirq gates, with single qubit operations being decomposed into the specified pauli string.device (
None
orcirq.Device
) – Acirq.Device
, if no device is provided, an unconstrained device withcirq.GridQubit
is assumed.join_meas (
bool
) – This determines ifMeas
are joined into singlecirq.meas()
objects, or single meas object in parallel in the given moment/cycle.
 Returns
A Cirq circuit representation of the
Circuit

to_pyquil
(metadata=None, config='ZXZ')¶ Converts from a
Circuit
to apyquil.Program
.This accepts a metadata object which contains information not present in a trueq circuit object.
Note
Measurements are always assumed to be at the end of a circuit.
If metadata is not provided, measurements are placed at the end of the circuit on all qubits, reading out to a classical register called ‘ro’
Aribtrary single qubit gates are decomposed using the ZXZ decomposition by default.
If an SU(4) gate is provided which is not performable by Pyquil, it is decomposed using the
trueq.compilation.Compiler()
. Parameters
circ (
Circuit
) – The TrueQᵀᴹ circuit to be converted to pyquil.metadata (
PyquilMetadata
) – Metadata which may not be present in the TrueQᵀᴹ representation, but is present in pyquil.config (
Config
orstr
) – If aConfig
which contains gates which are directly decomposable into Pyquil native gates is provided, then conversion will take place into gates specified by that config. Additionallyconfig
also accepts text strings describing decomposition methods, strings must be of paulis which are defined intrueq.config.decomposition.QubitMode
, thenGate
s are decomposed into all allowed Pyquil gates, with single qubit operations being decomposed into the specified pauli string.

to_qiskit
(metadata=None, config=None)¶ Converts a TrueQᵀᴹ circuit into an Qiskit Circuit.
Note
If nothing else is provided, this adds measurements on all qubits at the end of the circuit.
Single qubit gates are reduced to the simplest they can up to numerical rounding, IE: some U3 gates become U2 or U1.
Blocking operations are added at the end of every
Cycle
 Parameters
circuit (
Circuit
) – A TrueQᵀᴹ circuit to be converted into an Qiskit QuantumCircuitmetadata (
QiskitMetadata
) – Metadata required to accurately reproduce the original qiskit circuit. If this is not provided, only 1 quantum register is created, and on classical register, then measurements are added at the end of the circuit mapping each quantum bit to each classical bit.config (
Config
) – A given config object which contains gates which can be performed by qiskit, but also has hardware restrictions. For example, a Config object can represent a CNOT gate that is restricted to specific sets of qubits, and a CNOT gate is a known native gate in Qiskit. Providing this config means that not only will the circuit be converted into Qiskit representation, but also that it will satisfy hardware restrictions. If no config is provided, all possible qiskit gates are used in the compilation.
 Returns
A Qiskit circuit representation of the
Circuit
 Return type
qiskit.QuantumCircuit
Circuit Collections¶

class
trueq.
CircuitCollection
(circuits=None)¶ Represents a collection of instances of
Circuit
. Parameters
circuits (
list
) – Alist
ofCircuit
s to initially populate this collection with, orNone
.

append
(circuit)¶ Appends one or more
Circuit
s to the collection. Parameters
circuit (
Circuit
Iterable
) – An instance ofCircuit
, or any iterable ofCircuit
objects (including anotherCircuitCollection
).

property
n_results
¶ The number of experimental
Circuit
s that have been populated with results, an integer between \(0\) andn_circuits
. Type
int

property
has_no_results
¶ Whether none of the experimental
Circuit
’s have been populated with data. Type
bool

property
has_some_results
¶ Whether at least one of the experimental
Circuit
s have been populated with data. Type
bool

property
has_all_results
¶ Whether all of the experimental
Circuit
s have been populated with data. Type
bool

save
(filename, overwrite=False, include_cycles=True)¶ Saves this collection to disk in a binary format, including all results. This collection can be reinstantiated at a later time using
trueq.utils.load()
.import trueq as tq circuits = tq.make_srb([0], [4, 20]) tq.Simulator().add_depolarizing(0.01).run(circuits) # save without cycle/gate information for a smaller filesize circuits.save("myfile1.bin", include_cycles=False) # save with all cycle/gate information circuits.save("myfile2.bin")
 Parameters
filename (
str
) – A filename to save the circuit collection as.overwrite (
bool
) – Any existing file will be overwritten if this isTrue
, else saving will raiseFileExistsError
include_cycles (
bool
) – Whether the cycle information in each circuit should be retained. If false, then all circuits will have their cycles stripped in the saved version to save space. Results and other information will be saved so that analysis and plotting will work on a reinstantiated copy.

expectation_values
(labels=None, **filter)¶ Finds the expectation values of all applicable
Circuit
s, marginalized over specified labels.The labels can be a list of labels, in which case each label is separately marginalized over. For example,
labels = [0, 1, 2]
shows the expectation value for qubits 0, 1, and 2 separately. The labels can also be a list of lists of labels; each list will be separately marginalized over. For example,labels = [[0], [1], [2]]
is equivalent tolabels = [0, 1, 2]
, butlabels = [[0, 1], [2]]
will report the expectation value of the subregister[0, 1]
and the expectation value of the subregister 2.If labels are not provided, they are taken from looking at all existing
key.twirl
values in the relevantKey
of thisCircuitCollection
. If two different twirl values are found, constructing labels is ambiguous and this functions raises. Parameters
labels (
list
) – A list of labels (or list of lists of labels) specifying which subset of systems (or list of subsets) to marginalize over when computing expectaion values.filter – A filter to apply to the datastructure, for example,
circuits.fit(protocol="SRB")
.
 Returns
A
KeyMap
whose keys specify grouped collections of circuits, and whose values are lists of expectation values for circuits in those collections. Return type
KeyMap

fit
(labels=None, **filter)¶ Fits all applicable data to exponential curves and returns a
ParameterCollection
detailing summary statistics.The labels can be a list of labels, in which case each label is separately marginalized over. For example,
labels = [0, 1, 2]
shows the expectation value for qubits 0, 1, and 2 separately. The labels can also be a list of lists of labels; each list will be separately marginalized over. For example,labels = [[0], [1], [2]]
is equivalent tolabels = [0, 1, 2]
, butlabels = [[0, 1], [2]]
will report the expectation value of the subregister[0, 1]
and the expectation value of the subregister 2.If labels are not provided, they are taken from looking at all existing ‘twirl’ values in the relevant
Key
of thisCircuitCollection
. If two different twirl values are found, constructing labels is ambiguous and this functions raises. Parameters
labels (
list
) – A list of labels (or list of lists of labels) specifying which qubit subset (or list of qubit subsets) to marginalize over when producing fits.return_expectations (
bool
) – Whether to return the expectation values as well as theParameterCollection
; the return is(fits, decay_data)
in this case.filter – A filter to apply to the datastructure, for example,
circuits.fit(protocol="SRB")
.
 Return type

plot
= PlottingSuite(<8 plotters>)¶ An object that stores all plotting functions deemed relevant to this
CircuitCollection
. If one of these functions is called, the data from this circuit collection is analyzed and used.Note
If there is a plotting function that you expected to be present but is not, double check the circuit’s
Key
s to see which types of circuits are present, and also verify that circuits relevant to the plot you want haveresults
entered. Type
PlottingSuite

batch
(max_size, extra_circuits=None, sequencer=None)¶ Divides the circuits in this collection into a series of batches that are limited in size. This is useful if, for example, your hardware has a finite amount of memory to store circuits with.
import trueq as tq import trueq.assessments as tqa # make a bunch of circuits circuits = tqa.make_crosstalk_diagnostics([0, 1, 2], [4, 70]) # split collection into batches such that there are no more than 80 # circuits in each batch # assume we have a hardware device called my_device with a method run() # which returns the results for each member of the batch for batch in circuits.batch(80): batch.results = my_device.run(batch) # we can also set batch size by total number of cycles or gates import trueq.sequencer as tqs circuits.batch(tqs.NCycles(1000)) circuits.batch(tqs.NGates(1e6)) # we can riffle the circuits by sequence length for batch in circuits.batch(80, sequencer=tqs.RIFFLER): batch.results = my_device.run(batch)
The batcher also supports automatically adding a fixed set of circuits to each batch. For example, we can add readout calibration circuits to each batch:
circuits = tqa.make_crosstalk_diagnostics([0, 1, 2], [4, 70]) ro_circuits = tq.make_readout_calibration([0, 1, 2]) for batch in circuits.batch(80, extra_circuits=ro_circuits): batch.results = my_device.run(batch)
 Parameters
max_size (
int
list
CircuitRuler
) – The maximum number of circuits per batch, or anyCircuitRuler
such as :NCycles
, py:class:~trueq.sequencer.Gates, :or py:class:~trueq.sequencer.TotalTime. You can also specify a list ofCircuitRuler
s; every batch will not surpass any of the rulers’ limits.extra_circuits (
CircuitCollection
) – A list of circuits to be placed at the start of every batch. Copies are made of these circuits in every batch.sequencer (
CircuitSequencer
) – Specifies how the circuits should be ordered when batching. This is useful, for example, if you want to riffle long circuits with short sequences within a batch, so that all of your long circuits don’t end up in the same batch, seetrueq.sequencer.RIFFLER
. If no sequencer is given, the existing circuit collection order is used.

keys
(**filter)¶ Returns a
KeySet
of unique keys in thisCircuitCollection
that match the given name/values present in the filter. Parameters
filter – Key/value pairs to be used as a filter.
 Return type
set

similar_keys
(*names, invert=False, **filter)¶ Returns a
generator
over disjointKeySet
s, where all of the keys in each set have equal values for all names appearing in the given names, and have any specific values specified in filter.circuits = tq.make_srb([0], [4, 10, 50], 30) for keys in circuits.similar_keys("protocol"): print(keys) KeySet( Key(n_random_cycles=4, protocol='SRB'), Key(n_random_cycles=10, protocol='SRB'), Key('n_random_cycles=50, protocol='SRB') ) for keys in circuits.similar_keys("protocol", "n_random_cycles"): print(keys) KeySet(Key(n_random_cycles=10, protocol='SRB')) KeySet(Key(n_random_cycles=4, protocol='SRB')) KeySet(Key(n_random_cycles=50, protocol='SRB'))
 Parameters
names (
list
) – An iterable of key names.invert (
bool
) – Whether the list of names to use should be those presented, or all names execpt those presented.filter – Key/value pairs to be used as a filter.

update_keys
(**kwargs)¶ Updates every circuit’s key in this collection with new keywords/values. If a given key does not have a given keyword, it is added. If it already exists, it is overwritten.
import trueq as tq circuits = tq.make_srb(0, [4, 10]) # give each circuit a new keyword 'banana' with value 10 circuits.update_keys(banana=10) # change 'n_random_cycles' to 5 for those circuits where it is 4 circuits.subset(n_random_cycles=4).update_keys(n_random_cycles=5)
 Parameters
kwargs – Keywords ands values to update with.

subset
(has_results=None, new_collection=True, **filter)¶ Selects a subset of this collection’s
Circuit
s by matching against a filter. This returns a newCircuitCollection
, with references to the original circuits, i.e. no circuit copies are made. Optionally, by using thenew_collection
flag, the output type can be a generator over circuits that match the filter rather than a new collection object.# make a new CircuitCollection with only SRB circuits srb_circuits = circuits.subset(protocol="SRB") assert srb_circuits.keys().protocol == {"SRB"} # loop over circuits with no results # for efficiency, use new_collection=False to avoid new instantiation for circuit in circuits.subset(has_results=False, new_collection=False): circuit.results = {"0": 1}
 Parameters
has_results (
None
bool
) – Whether to filter on circuits which have or do not have results entered. IfTrue
, only circuits with results are included, ifFalse
, only circuits without results are included, and ifNone
(default) all circuits are included.new_collection (
bool
) – Whether to create a newCircuitCollection
and return it. Otherwise, a generator over circuits will be returned, which saves memory.filter – Key/value pairs to be used as a filter.
 Return type
generator
CircuitCollection

to_dict_list
(include_cycles=True)¶ Returns a list of dictionary representations of the
Circuit
objects in this collection. Parameters
include_cycles (
bool
) – Whether to include the cycles in the dictionary representations. Return type
list

static
from_dict_list
(lst)¶ Returns a
CircuitCollection
constructed from a list of dictionary representations oftrueq.Circuit
s. Return type

batch_size
(ruler)¶ The size of this collection according to the given method of counting circuit size. For example,
NCycles
,NGates
, orTotalTime
. Parameters
ruler (
CircuitRuler
) – A callable that returns the size of a given circuit.

property
results
¶ A list of the
Results
for every circuit in this collection, where the order of this list matches the order of this collection. This property is also settable, seeCircuit.results
for allowed formats for setting.import trueq as tq circuits = tq.make_srb([0], [4, 100]) circuits.results # [Results(), Results(), ....] # set all results to the same value circuits.results = [{"0": 10, "1": 100}] * len(circuits)
 Type
list

sum_results
(decompile_paulis=True, **filter)¶ Sums together all results in the collection that match the given filter.
import trueq as tq circuits = tq.make_srb([0], [4, 100]) circuits.append(tq.make_xrb([0], [4, 100])) tq.Simulator().add_overrotation(0.05).run(circuits, n_shots=10) circuits.sum_results() # Results({'1': 1187, '0': 1213}) circuits.sum_results(protocol="XRB").n_shots # 1800 circuits.sum_results(n_random_cycles=4).n_shots # 1200
 Parameters
decompile_paulis (
bool
) – Whether to adjust ditstrings of each circuit’s results by the'compiled_pauli'
key value of the corresponding circuit when summing. In most cases, and especially for randomized compiling, this value should beTrue
.filter – Key/value pairs to be used as a filter.
 Return type
Results