ProjectQ
ProjectQ is an opensource software framework for quantum computing. It aims at providing tools which facilitate inventing, implementing, testing, debugging, and running quantum algorithms using either classical hardware or actual quantum devices.
The four core principles of this opensource effort are
Open & Free: ProjectQ is released under the Apache 2 license
Simple learning curve: It is implemented in Python and has an intuitive syntax
Easily extensible: Anyone can contribute to the compiler, the embedded domainspecific language, and libraries
Code quality: Code reviews, continuous integration testing (unit and functional tests)
 Please cite
Damian S. Steiger, Thomas Häner, and Matthias Troyer “ProjectQ: An Open Source Software Framework for Quantum Computing” Quantum 2, 49 (2018) (published on arXiv on 23 Dec 2016)
Thomas Häner, Damian S. Steiger, Krysta M. Svore, and Matthias Troyer “A Software Methodology for Compiling Quantum Programs” Quantum Sci. Technol. 3 (2018) 020501 (published on arXiv on 5 Apr 2016)
 Contents
Tutorial: Tutorial containing instructions on how to get started with ProjectQ.
Examples: Example implementations of few quantum algorithms
Code Documentation: The code documentation of ProjectQ.
Tutorial
Getting started
To start using ProjectQ, simply run
python m pip install user projectq
Since version 0.6.0, ProjectQ is available as precompiled binary wheels in addition to the traditional source package. These wheels should work on most platforms, provided that your processor supports AVX2 instructions. Should you encounter any troubles while installation ProjectQ in binary form, you can always try tom compile the project manually as described below. You may want to pass the –nobinary projectq flag to Pip during the installation to make sure that you are downloading the source package.
Alternatively, you can also clone/download this repository (e.g., to your /home directory) and run
cd /home/projectq
python m pip install user .
ProjectQ comes with a highperformance quantum simulator written in C++. Please see the detailed OS specific installation instructions below to make sure that you are installing the fastest version.
Note
The setup will try to build a C++Simulator, which is much faster than the Python implementation. If the C++ compilation were to fail, the setup will install a pure Python implementation of the simulator instead. The Python simulator should work fine for small examples (e.g., running Shor’s algorithm for factoring 15 or 21).
If you want to skip the installation of the C++Simulator altogether, you can define the PROJECTQ_DISABLE_CEXT
environment variable to avoid any compilation steps.
Note
If building the C++Simulator does not work out of the box, consider specifying a different compiler. For example:
env CC=g++10 python m pip install user projectq
Please note that the compiler you specify must support at least C++11!
Note
Please use pip version v6.1.0 or higher as this ensures that dependencies are installed in the correct order.
Note
ProjectQ should be installed on each computer individually as the C++ simulator compilation creates binaries which are optimized for the specific hardware on which it is being installed (potentially using our AVX version and march=native). Therefore, sharing the same ProjectQ installation across different hardware may cause some problems.
Install AWS Braket Backend requirement
AWS Braket Backend requires the use of the official AWS SDK for Python, Boto3. This is an extra requirement only needed if you plan to use the AWS Braket Backend. To install ProjectQ inluding this requirement you can include it in the installation instruction as
python m pip install user projectq[braket]
Install Azure Quantum Backend requirement
Azure Quantum Backend requires the use of the official Azure Quantum SDK for Python. This is an extra requirement only needed if you plan to use the Azure Quantum Backend. To install ProjectQ inluding this requirement you can include it in the installation instruction as
python m pip install user projectq[azurequantum]
Detailed instructions and OSspecific hints
Ubuntu:
After having installed the build tools (for g++):
sudo aptget install buildessentialYou only need to install Python (and the package manager). For version 3, run
sudo aptget install python3 python3pipWhen you then run
sudo python3 m pip install user projectqall dependencies (such as numpy and pybind11) should be installed automatically.
ArchLinux/Manjaro:
Make sure that you have a C/C++ compiler installed:
sudo pacman Syu gccYou only need to install Python (and the package manager). For version 3, run
sudo pacman Syu python pythonpipWhen you then run
sudo python3 m pip install user projectqall dependencies (such as numpy and pybind11) should be installed automatically.
Windows:
It is easiest to install a precompiled version of Python, including numpy and many more useful packages. One way to do so is using, e.g., the Python 3.7 installers from python.org or ANACONDA. Installing ProjectQ right away will succeed for the (slow) Python simulator. For a compiled version of the simulator, install the Visual C++ Build Tools and the Microsoft Windows SDK prior to doing a pip install. The built simulator will not support multithreading due to the limited OpenMP support of the Visual Studio compiler.
If the Python executable is added to your PATH (option normally suggested at the end of the Python installation procedure), you can then open a cmdline window (WIN + R, type “cmd” and click OK) and enter the following in order to install ProjectQ:
python m pip install user projectqShould you want to run multithreaded simulations, you can install a compiler which supports newer OpenMP versions, such as MinGW GCC and then manually build the C++ simulator with OpenMP enabled.
macOS:
Similarly to the other platforms, installing ProjectQ without the C++ simulator is really easy:
python3 m pip install user projectqIn order to install the fast C++ simulator, we require that a C++ compiler is installed on your system. There are essentially three options you can choose from:
Using the compiler provided by Apple through the XCode command line tools.
Using Homebrew
Using MacPorts
For both options 2 and 3, you will be required to first install the XCode command line tools
Apple XCode command line tool
Install the XCode command line tools by opening a terminal window and running the following command:
xcodeselect installNext, you will need to install Python and pip. See options 2 and 3 for information on how to install a newer python version with either Homebrew or MacPorts. Here, we are using the standard python which is preinstalled with macOS. Pip can be installed by:
sudo easy_install pipNow, you can install ProjectQ with the C++ simulator using the standard command:
python3 m pip install user projectqNote that the compiler provided by Apple is currently not able to compile ProjectQ’s multithreaded code.
Homebrew
First install the XCode command line tools. Then install Homebrew with the following command:
/usr/bin/ruby e "$(curl fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"Then proceed to install Python as well as a C/C++ compiler (note: gcc installed via Homebrew may lead to some issues):
brew install python llvmYou should now be able to install ProjectQ with the C++ simulator using the following command:
env P=/usr/local/opt/llvm/bin CC=$P/clang CXX=$P/clang++ python3 m pip install user projectqMacPorts
Visit macports.org and install the latest version that corresponds to your operating system’s version. Afterwards, open a new terminal window.
Then, use macports to install Python 3.7 by entering the following command
sudo port install python37It might show a warning that if you intend to use python from the terminal. In this case, you should also install
sudo port install py37gnureadlineInstall pip by
sudo port install py37pipNext, we can install ProjectQ with the high performance simulator written in C++. First, we will need to install a suitable compiler with support for C++11, OpenMP, and instrinsics. The best option is to install clang 9.0 also using macports (note: gcc installed via macports does not work).
sudo port install clang9.0ProjectQ is now installed by:
env CC=clangmp9.0 env CXX=clang++mp9.0 /opt/local/bin/python3.7 m pip install user projectq
The ProjectQ syntax
Our goal is to have an intuitive syntax in order to enable an easy learning curve. Therefore, ProjectQ features a lean syntax which is close to the mathematical notation used in physics.
For example, consider applying an xrotation by an angle theta to a qubit. In ProjectQ, this looks as follows:
Rx(theta)  qubit
whereas the corresponding notation in physics would be
\(R_x(\theta) \; \text{qubit}\rangle\)
Moreover, the operator separates the classical arguments (on the left) from the quantum arguments (on the right). Next, you will see a basic quantum program using this syntax. Further examples can be found in the docs (Examples in the panel on the left) and in the ProjectQ examples folder on GitHub.
Basic quantum program
To check out the ProjectQ syntax in action and to see whether the installation worked, try to run the following basic example
from projectq import MainEngine # import the main compiler engine
from projectq.ops import (
H,
Measure,
) # import the operations we want to perform (Hadamard and measurement)
eng = MainEngine() # create a default compiler (the backend is a simulator)
qubit = eng.allocate_qubit() # allocate 1 qubit
H  qubit # apply a Hadamard gate
Measure  qubit # measure the qubit
eng.flush() # flush all gates (and execute measurements)
print(f"Measured {int(qubit)}") # output measurement result
Which creates random bits (0 or 1).
Examples
All of these example codes and more can be found on GitHub.
Quantum Random Numbers
The most basic example is a quantum random number generator (QRNG). It can be found in the examplesfolder of ProjectQ. The code looks as follows
# pylint: skipfile
"""Example of a simple quantum random number generator."""
from projectq import MainEngine
from projectq.ops import H, Measure
# create a main compiler engine
eng = MainEngine()
# allocate one qubit
q1 = eng.allocate_qubit()
# put it in superposition
H  q1
# measure
Measure  q1
eng.flush()
# print the result:
print(f"Measured: {int(q1)}")
Running this code three times may yield, e.g.,
$ python examples/quantum_random_numbers.py
Measured: 0
$ python examples/quantum_random_numbers.py
Measured: 0
$ python examples/quantum_random_numbers.py
Measured: 1
These values are obtained by simulating this quantum algorithm classically. By changing three lines of code, we can run an actual quantum random number generator using the IBM Quantum Experience backend:
$ python examples/quantum_random_numbers_ibm.py
Measured: 1
$ python examples/quantum_random_numbers_ibm.py
Measured: 0
All you need to do is:
Create an account for IBM’s Quantum Experience
And perform these minor changes:
 /home/docs/checkouts/readthedocs.org/user_builds/projectq/checkouts/v0.8.0/examples/quantum_random_numbers.py +++ /home/docs/checkouts/readthedocs.org/user_builds/projectq/checkouts/v0.8.0/examples/quantum_random_numbers_ibm.py @@ 1,12 +1,14 @@ # pylint: skipfile """Example of a simple quantum random number generator.""" +"""Example of a simple quantum random number generator using IBM's API.""" +import projectq.setups.ibm from projectq import MainEngine +from projectq.backends import IBMBackend from projectq.ops import H, Measure # create a main compiler engine eng = MainEngine() +eng = MainEngine(IBMBackend(), engine_list=projectq.setups.ibm.get_engine_list()) # allocate one qubit q1 = eng.allocate_qubit()
Quantum Teleportation
Alice has a qubit in some interesting state \(\psi\rangle\), which she would like to show to Bob. This does not really make sense, since Bob would not be able to look at the qubit without collapsing the superposition; but let’s just assume Alice wants to send her state to Bob for some reason. What she can do is use quantum teleportation to achieve this task. Yet, this only works if Alice and Bob share a Bellpair (which luckily happens to be the case). A Bellpair is a pair of qubits in the state
They can create a Bellpair using a very simple circuit which first applies a Hadamard gate to the first qubit, and then flips the second qubit conditional on the first qubit being in \(1\rangle\). The circuit diagram can be generated by calling the function
from projectq.meta import Control, Dagger
eng (MainEngine): MainEngine from which to allocate the qubits.
Returns:
bell_pair (tuple<Qubits>): The Bellpair.
"""
b1 = eng.allocate_qubit()
b2 = eng.allocate_qubit()
with a main compiler engine which has a CircuitDrawer backend, i.e.,
# pylint: skipfile
"""Example implementation of a quantum circuit generating a Bell pair state."""
import matplotlib.pyplot as plt
from teleport import create_bell_pair
from projectq import MainEngine
from projectq.backends import CircuitDrawer
from projectq.libs.hist import histogram
from projectq.setups.default import get_engine_list
# create a main compiler engine
drawing_engine = CircuitDrawer()
eng = MainEngine(engine_list=get_engine_list() + [drawing_engine])
qb0, qb1 = create_bell_pair(eng)
eng.flush()
print(drawing_engine.get_latex())
histogram(eng.backend, [qb0, qb1])
plt.show()
The resulting LaTeX code can be compiled to produce the circuit diagram:
$ python examples/bellpair_circuit.py > bellpair_circuit.tex
$ pdflatex bellpair_circuit.tex
The output looks as follows:
Now, this Bellpair can be used to achieve the quantum teleportation: Alice entangles her qubit with her share of the Bellpair. Then, she measures both qubits; one in the Zbasis (Measure) and one in the Hadamard basis (Hadamard, then Measure). She then sends her measurement results to Bob who, depending on these outcomes, applies a PauliX or Z gate.
The complete example looks as follows:
1# pylint: skipfile
2
3"""Example of a quantum teleportation circuit."""
4
5from projectq import MainEngine
6from projectq.meta import Control, Dagger
7 eng (MainEngine): MainEngine from which to allocate the qubits.
8
9 Returns:
10 bell_pair (tuple<Qubits>): The Bellpair.
11 """
12 b1 = eng.allocate_qubit()
13 b2 = eng.allocate_qubit()
14
15 H  b1
16 CNOT  (b1, b2)
17 verbose (bool): If True, info messages will be printed.
18
19 """
20 # make a Bellpair
21 b1, b2 = create_bell_pair(eng)
22
23 # Alice creates a nice state to send
24 psi = eng.allocate_qubit()
25 if verbose:
26 print("Alice is creating her state from scratch, i.e., 0>.")
27 state_creation_function(eng, psi)
28
29 # entangle it with Alice's b1
30 CNOT  (psi, b1)
31 if verbose:
32 print("Alice entangled her qubit with her share of the Bellpair.")
33
34 # measure two values (once in Hadamard basis) and send the bits to Bob
35 H  psi
36 Measure  psi
37 Measure  b1
38 msg_to_bob = [int(psi), int(b1)]
39 if verbose:
40 print(f"Alice is sending the message {msg_to_bob} to Bob.")
41
42 # Bob may have to apply up to two operation depending on the message sent
43 # by Alice:
44 with Control(eng, b1):
45 X  b2
46 with Control(eng, psi):
47 Z  b2
48
49 # try to uncompute the psi state
50 if verbose:
51 print("Bob is trying to uncompute the state.")
52 with Dagger(eng):
53 state_creation_function(eng, b2)
54
55 # check whether the uncompute was successful. The simulator only allows to
56 # delete qubits which are in a computational basis state.
57 del b2
58 eng.flush()
59
60 if verbose:
61 print("Bob successfully arrived at 0>")
62
63
64if __name__ == "__main__":
65 # create a main compiler engine with a simulator backend:
66 eng = MainEngine()
67
68 # define our statecreation routine, which transforms a 0> to the state
69 # we would like to send. Bob can then try to uncompute it and, if he
70 # arrives back at 0>, we know that the teleportation worked.
71 def create_state(eng, qb):
72 """Create a quantum state."""
73 H  qb
and the corresponding circuit can be generated using
$ python examples/teleport_circuit.py > teleport_circuit.tex
$ pdflatex teleport_circuit.tex
which produces (after renaming of the qubits inside the texfile):
Shor’s algorithm for factoring
As a third example, consider Shor’s algorithm for factoring, which for a given (large) number \(N\) determines the two prime factor \(p_1\) and \(p_2\) such that \(p_1\cdot p_2 = N\) in polynomial time! This is a superpolynomial speedup over the best known classical algorithm (which is the number field sieve) and enables the breaking of modern encryption schemes such as RSA on a future quantum computer.
 A tiny bit of number theory
There is a small amount of number theory involved, which reduces the problem of factoring to periodfinding of the function
\[f(x) = a^x\operatorname{mod} N\]for some a (relative prime to N, otherwise we get a factor right away anyway by calling gcd(a,N)). The period r for a function f(x) is the number for which \(f(x) = f(x+r)\forall x\) holds. In this case, this means that \(a^x = a^{x+r}\;\; (\operatorname{mod} N)\;\forall x\). Therefore, \(a^r = 1 + qN\) for some integer q and hence, \(a^r  1 = (a^{r/2}  1)(a^{r/2}+1) = qN\). This suggests that using the gcd on N and \(a^{r/2} \pm 1\) we may find a factor of N!
 Factoring on a quantum computer: An example
At the heart of Shor’s algorithm lies modular exponentiation of a classically known constant (denoted by a in the code) by a quantum superposition of numbers \(x\), i.e.,
\[x\rangle0\rangle \mapsto x\ranglea^x\operatorname{mod} N\rangle\]Using \(N=15\) and \(a=2\), and applying this operation to the uniform superposition over all \(x\) leads to the superposition (modulo renormalization)
\[0\rangle1\rangle + 1\rangle2\rangle + 2\rangle4\rangle + 3\rangle8\rangle + 4\rangle1\rangle + 5\rangle2\rangle + 6\rangle4\rangle + \cdots\]In Shor’s algorithm, the second register will not be touched again before the end of the quantum program, which means it might as well be measured now. Let’s assume we measure 2; this collapses the state above to
\[1\rangle2\rangle + 5\rangle2\rangle + 9\rangle2\rangle + \cdots\]The period of a modulo N can now be read off. On a quantum computer, this information can be accessed by applying an inverse quantum Fourier transform to the xregister, followed by a measurement of x.
 Implementation
There is an implementation of Shor’s algorithm in the examples folder. It uses the implementation by Beauregard, arxiv:0205095 to factor an nbit number using 2n+3 qubits. In this implementation, the modular exponentiation is carried out using modular multiplication and shift. Furthermore it uses the semiclassical quantum Fourier transform [see arxiv:9511007]: Pulling the final measurement of the xregister through the final inverse quantum Fourier transform allows to run the 2n modular multiplications serially, which keeps one from having to store the 2n qubits of x.
Let’s run it using the ProjectQ simulator:
$ python3 examples/shor.py projectq  Implementation of Shor's algorithm. Number to factor: 15 Factoring N = 15: 00000001 Factors found :) : 3 * 5 = 15
Simulating Shor’s algorithm at the level of singlequbit gates and CNOTs already takes quite a bit of time for larger numbers than 15. To turn on our emulation feature, which does not decompose the modular arithmetic to lowlevel gates, but carries it out directly instead, we can change the line
86 return r 87 88 89# Filter function, which defines the gate set for the first optimization 90# (don't decompose QFTs and iQFTs to make cancellation easier) 91def high_level_gates(eng, cmd): 92 """Filter highlevel gates.""" 93 g = cmd.gate 94 if g == QFT or get_inverse(g) == QFT or g == Swap: 95 return True 96 if isinstance(g, BasicMathGate): 97 return False 98 if isinstance(g, AddConstant): 99 return True
in examples/shor.py to return True. This allows to factor, e.g. \(N=4,028,033\) in under 3 minutes on a regular laptop!
The most important part of the code is
50 measurements = [0] * (2 * n) # will hold the 2n measurement results 51 52 ctrl_qubit = eng.allocate_qubit() 53 54 for k in range(2 * n): 55 current_a = pow(a, 1 << (2 * n  1  k), N) 56 # one iteration of 1qubit QPE 57 H  ctrl_qubit 58 with Control(eng, ctrl_qubit): 59 MultiplyByConstantModN(current_a, N)  x 60 61 # perform inverse QFT > Rotations conditioned on previous outcomes 62 for i in range(k): 63 if measurements[i]: 64 R(math.pi / (1 << (k  i)))  ctrl_qubit 65 H  ctrl_qubit 66 67 # and measure 68 Measure  ctrl_qubit 69 eng.flush()
which executes the 2n modular multiplications conditioned on a control qubit ctrl_qubit in a uniform superposition of 0 and 1. The control qubit is then measured after performing the semiclassical inverse quantum Fourier transform and the measurement outcome is saved in the list measurements, followed by a reset of the control qubit to state 0.
Code Documentation
Welcome to the package documentation of ProjectQ. You may now browse through the entire documentation and discover the capabilities of the ProjectQ framework.
For a detailed documentation of a subpackage or module, click on its name below:
backends
ProjectQ module for supporting the AQT platform 

ProjectQ module for supporting the AWS Braket platform 

ProjectQ module for supporting the Azure Quantum platform. 

ProjectQ module for exporting/printing quantum circuits 

Exception classes for projectq.backends. 

ProjectQ module for supporting the IBM QE platform 

ProjectQ module for supporting the IonQ platform 

Contains a compiler engine which prints commands to stdout prior to sending them on to the next engines. 

Contain a compiler engine to calculate resource count used by a quantum circuit. 

ProjectQ module dedicated to simulation 

Contain a backend that saves the unitary of a quantum circuit. 

Module containing some utility functions. 


Backend for building circuits and submitting them to the AQT API. 

Dummy class 

Dummy class 
CircuitDrawer is a compiler engine which generates TikZ code for drawing quantum circuits. 

CircuitDrawerMatplotlib is a compiler engine which using Matplotlib library for drawing quantum circuits. 

A simple introspective simulator that only permits classical operations. 

Compiler engine that prints command to the standard output. 

Exception raised if a selected device cannot handle the circuit or is not supported by ProjectQ. 

Raised when a device is required but is currently offline. 

Raised when a device does not have enough qubits for a desired job. 


Define the compiler engine class that handles interactions with the IBM API. 
Backend for building circuits and submitting them to the IonQ API. 

ResourceCounter is a compiler engine which counts the number of gates and max. 


Simulator is a compiler engine which simulates a quantum computer using C++based kernels. 
Simulator engine aimed at calculating the unitary transformation that represents the current quantum circuit. 
Submodules
_aqt
ProjectQ module for supporting the AQT platform
_awsbraket
ProjectQ module for supporting the AWS Braket platform
_azure
ProjectQ module for supporting the Azure Quantum platform.
_circuits
ProjectQ module for exporting/printing quantum circuits
_exceptions
Exception classes for projectq.backends.
 exception projectq.backends._exceptions.DeviceNotHandledError[source]
Exception raised if a selected device cannot handle the circuit or is not supported by ProjectQ.
 exception projectq.backends._exceptions.DeviceOfflineError[source]
Raised when a device is required but is currently offline.
 exception projectq.backends._exceptions.DeviceTooSmall[source]
Raised when a device does not have enough qubits for a desired job.
 exception projectq.backends._exceptions.InvalidCommandError[source]
Raised if the backend encounters an invalid command.
 exception projectq.backends._exceptions.JobSubmissionError[source]
Raised when the job creation API contains an error of some kind.
_ibm
ProjectQ module for supporting the IBM QE platform
_ionq
ProjectQ module for supporting the IonQ platform
 class projectq.backends._ionq.IonQBackend(use_hardware=False, num_runs=100, verbose=False, token=None, device='ionq_simulator', num_retries=3000, interval=1, retrieve_execution=None)[source]
Backend for building circuits and submitting them to the IonQ API.
 get_probabilities(qureg)[source]
Given the provided qubit register, determine the probability of each possible outcome.
Note
This method should only be called after a circuit has been run and its results are available.
 Parameters
qureg (Qureg) – A ProjectQ Qureg object.
 Returns
A dict mapping of states > probability.
 Return type
dict
 get_probability(state, qureg)[source]
Shortcut to get a specific state’s probability.
 Parameters
state (str) – A state in bitstring format.
qureg (Qureg) – A ProjectQ Qureg object.
 Returns
The probability for the provided state.
 Return type
float
_printer
Contains a compiler engine which prints commands to stdout prior to sending them on to the next engines.
 class projectq.backends._printer.CommandPrinter(accept_input=True, default_measure=False, in_place=False)[source]
Compiler engine that prints command to the standard output.
CommandPrinter is a compiler engine which prints commands to stdout prior to sending them on to the next compiler engine.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: Returns True if the CommandPrinter is the last engine (since it can print any command).
 Parameters
cmd (Command) – Command of which to check availability (all Commands can be printed).
 Returns
True, unless the next engine cannot handle the Command (if there is a next engine).
 Return type
availability (bool)
_resource
Contain a compiler engine to calculate resource count used by a quantum circuit.
A resrouce counter compiler engine counts the number of calls for each type of gate used in a circuit, in addition to the max. number of active qubits.
 class projectq.backends._resource.ResourceCounter[source]
ResourceCounter is a compiler engine which counts the number of gates and max. number of active qubits.
 gate_counts
Dictionary of gate counts. The keys are tuples of the form (cmd.gate, ctrl_cnt), where ctrl_cnt is the number of control qubits.
 Type
dict
 gate_class_counts
Dictionary of gate class counts. The keys are tuples of the form (cmd.gate.__class__, ctrl_cnt), where ctrl_cnt is the number of control qubits.
 Type
dict
 max_width
Maximal width (=max. number of active qubits at any given point).
 Type
int
 Properties:
depth_of_dag (int): It is the longest path in the directed acyclic graph (DAG) of the program.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: Returns True if the ResourceCounter is the last engine (since it can count any command).
 Parameters
cmd (Command) – Command for which to check availability (all Commands can be counted).
 Returns
True, unless the next engine cannot handle the Command (if there is a next engine).
 Return type
availability (bool)
_sim
ProjectQ module dedicated to simulation
_unitary
Contain a backend that saves the unitary of a quantum circuit.
 class projectq.backends._unitary.UnitarySimulator[source]
Simulator engine aimed at calculating the unitary transformation that represents the current quantum circuit.
 unitary[source]
Current unitary representing the quantum circuit being processed so far.
 Type
np.ndarray
Note
The current implementation of this backend resets the unitary after the first gate that is neither a qubit deallocation nor a measurement occurs after one of those two aforementioned gates.
The old unitary call be accessed at anytime after such a situation occurs via the history property.
eng = MainEngine(backend=UnitarySimulator(), engine_list=[]) qureg = eng.allocate_qureg(3) All(X)  qureg eng.flush() All(Measure)  qureg eng.deallocate_qubit(qureg[1]) X  qureg[0] # WARNING: appending gate after measurements or deallocations resets the unitary
 property history[source]
Access all previous unitary matrices.
The current unitary matrix is appended to this list once a gate is received after either a measurement or a qubit deallocation has occurred.
 Returns
A list where the elements are all previous unitary matrices representing the circuit, separated by measurement/deallocate gates.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: The unitary simulator can deal with all arbitrarilycontrolled gates which provide a gatematrix (via gate.matrix).
 Parameters
cmd (Command) – Command for which to check availability (single qubit gate, arbitrary controls)
 Returns
True if it can be simulated and False otherwise.
 measure_qubits(ids)[source]
Measure the qubits with IDs ids and return a list of measurement outcomes (True/False).
 Parameters
ids (list<int>) – List of qubit IDs to measure.
 Returns
List of measurement results (containing either True or False).
 receive(command_list)[source]
Receive a list of commands.
 Receive a list of commands from the previous engine and handle them:
update the unitary of the quantum circuit
update the internal quantum state if a measurement or a qubit deallocation occurs
prior to sending them on to the next engine.
 Parameters
command_list (list<Command>) – List of commands to execute on the simulator.
_utils
Module containing some utility functions.
Module contents
Contains backends for ProjectQ.
This includes:
a debugging tool to print all received commands (CommandPrinter)
a circuit drawing engine (which can be used anywhere within the compilation chain)
a simulator with emulation capabilities
a resource counter (counts gates and keeps track of the maximal width of the circuit)
an interface to the IBM Quantum Experience chip (and simulator).
an interface to the AQT trapped ion system (and simulator).
an interface to the AWS Braket service decives (and simulators)
an interface to the Azure Quantum service devices (and simulators)
an interface to the IonQ trapped ionq hardware (and simulator).
 class projectq.backends.AQTBackend(use_hardware=False, num_runs=100, verbose=False, token='', device='simulator', num_retries=3000, interval=1, retrieve_execution=None)[source]
Backend for building circuits and submitting them to the AQT API.
The AQT Backend class, which stores the circuit, transforms it to the appropriate data format, and sends the circuit through the AQT API.
 __init__(use_hardware=False, num_runs=100, verbose=False, token='', device='simulator', num_retries=3000, interval=1, retrieve_execution=None)[source]
Initialize the Backend object.
 Parameters
use_hardware (bool) – If True, the code is run on the AQT quantum chip (instead of using the AQT simulator)
num_runs (int) – Number of runs to collect statistics. (default is 100, max is usually around 200)
verbose (bool) – If True, statistics are printed, in addition to the measurement result being registered (at the end of the circuit).
token (str) – AQT user API token.
device (str) – name of the AQT device to use. simulator By default
num_retries (int) – Number of times to retry to obtain results from the AQT API. (default is 3000)
interval (float, int) – Number of seconds between successive attempts to obtain results from the AQT API. (default is 1)
retrieve_execution (int) – Job ID to retrieve instead of re running the circuit (e.g., if previous run timed out).
 get_probabilities(qureg)[source]
Return the probability of the outcome bit_string when measuring the quantum register qureg.
Return the list of basis states with corresponding probabilities. If input qureg is a subset of the register used for the experiment, then returns the projected probabilities over the other states. The measured bits are ordered according to the supplied quantum register, i.e., the leftmost bit in the statestring corresponds to the first qubit in the supplied quantum register.
Warning
Only call this function after the circuit has been executed!
 Parameters
qureg (list<Qubit>) – Quantum register determining the order of the qubits.
 Returns
Dictionary mapping nbit strings to probabilities.
 Return type
probability_dict (dict)
 Raises
RuntimeError – If no data is available (i.e., if the circuit has not been executed). Or if a qubit was supplied which was not present in the circuit (might have gotten optimized away).
 class projectq.backends.CircuitDrawer(accept_input=False, default_measure=0)[source]
CircuitDrawer is a compiler engine which generates TikZ code for drawing quantum circuits.
The circuit can be modified by editing the settings.json file which is generated upon first execution. This includes adjusting the gate width, height, shadowing, line thickness, and many more options.
After initializing the CircuitDrawer, it can also be given the mapping from qubit IDs to wire location (via the
set_qubit_locations()
function):circuit_backend = CircuitDrawer() circuit_backend.set_qubit_locations({0: 1, 1: 0}) # swap lines 0 and 1 eng = MainEngine(circuit_backend) ... # run quantum algorithm on this main engine print(circuit_backend.get_latex()) # prints LaTeX code
To see the qubit IDs in the generated circuit, simply set the draw_id option in the settings.json file under “gates”:”AllocateQubitGate” to True:
{ "gates": { "AllocateQubitGate": { "draw_id": True, "height": 0.15, "width": 0.2, "pre_offset": 0.1, "offset": 0.1, }, # ... } }
The settings.json file has the following structure:
{ "control": {"shadow": false, "size": 0.1}, # settings for control "circle" "gate_shadow": true, # enable/disable shadows for all gates "gates": { "GateClassString": {GATE_PROPERTIES}, "GateClassString2": { # ... }, }, "lines": { # settings for qubit lines "double_classical": true, # draw doublelines for # classical bits "double_lines_sep": 0.04, # gap between the two lines # for double lines "init_quantum": true, # start out with quantum bits "style": "very thin", # line style }, }
All gates (except for the ones requiring special treatment) support the following properties:
{ "GateClassString": { "height": GATE_HEIGHT, "width": GATE_WIDTH, "pre_offset": OFFSET_BEFORE_PLACEMENT, "offset": OFFSET_AFTER_PLACEMENT, } }
 __init__(accept_input=False, default_measure=0)[source]
Initialize a circuit drawing engine.
The TikZ code generator uses a settings file (settings.json), which can be altered by the user. It contains gate widths, heights, offsets, etc.
 Parameters
accept_input (bool) – If accept_input is true, the printer queries the user to input measurement results if the CircuitDrawer is the last engine. Otherwise, all measurements yield the result default_measure (0 or 1).
default_measure (bool) – Default value to use as measurement results if accept_input is False and there is no underlying backend to register real measurement results.
 get_latex(ordered=False, draw_gates_in_parallel=True)[source]
Return the latex document string representing the circuit.
Simply write this string into a texfile or, alternatively, pipe the output directly to, e.g., pdflatex:
python3 my_circuit.py  pdflatex
where my_circuit.py calls this function and prints it to the terminal.
 Parameters
ordered (bool) – flag if the gates should be drawn in the order they were added to the circuit
draw_gates_in_parallel (bool) – flag if parallel gates should be drawn parallel (True), or not (False)
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: Returns True if the CircuitDrawer is the last engine (since it can print any command).
 Parameters
cmd (Command) – Command for which to check availability (all Commands can be printed).
 Returns
True, unless the next engine cannot handle the Command (if there is a next engine).
 Return type
availability (bool)
 receive(command_list)[source]
Receive a list of commands.
Receive a list of commands from the previous engine, print the commands, and then send them on to the next engine.
 Parameters
command_list (list<Command>) – List of Commands to print (and potentially send on to the next engine).
 set_qubit_locations(id_to_loc)[source]
Set the qubit lines to use for the qubits explicitly.
To figure out the qubit IDs, simply use the setting draw_id in the settings file. It is located in “gates”:”AllocateQubitGate”. If draw_id is True, the qubit IDs are drawn in red.
 Parameters
id_to_loc (dict) – Dictionary mapping qubit ids to qubit line numbers.
 Raises
RuntimeError – If the mapping has already begun (this function needs be called before any gates have been received).
 class projectq.backends.CircuitDrawerMatplotlib(accept_input=False, default_measure=0)[source]
CircuitDrawerMatplotlib is a compiler engine which using Matplotlib library for drawing quantum circuits.
 __init__(accept_input=False, default_measure=0)[source]
Initialize a circuit drawing engine(mpl).
 Parameters
accept_input (bool) – If accept_input is true, the printer queries the user to input measurement results if the CircuitDrawerMPL is the last engine. Otherwise, all measurements yield the result default_measure (0 or 1).
default_measure (bool) – Default value to use as measurement results if accept_input is False and there is no underlying backend to register real measurement results.
 draw(qubit_labels=None, drawing_order=None, **kwargs)[source]
Generate and returns the plot of the quantum circuit stored so far.
 Parameters
qubit_labels (dict) – label for each wire in the output figure. Keys: qubit IDs, Values: string to print out as label for that particular qubit wire.
drawing_order (dict) – position of each qubit in the output graphic. Keys: qubit IDs, Values: position of qubit on the qubit line in the graphic.
**kwargs (dict) – additional parameters are used to update the default plot parameters
 Returns
A tuple containing the matplotlib figure and axes objects
Note
Additional keyword arguments can be passed to this function in order to further customize the figure output by matplotlib (default value in parentheses):
fontsize (14): Font size in pt
column_spacing (.5): Vertical spacing between two neighbouring gates (roughly in inches)
control_radius (.015): Radius of the circle for controls
labels_margin (1): Margin between labels and begin of wire (roughly in inches)
linewidth (1): Width of line
not_radius (.03): Radius of the circle for X/NOT gates
gate_offset (.05): Inner margins for gates with a text representation
mgate_width (.1): Width of the measurement gate
swap_delta (.02): Halfsize of the SWAP gate
x_offset (.05): Absolute Xoffset for drawing within the axes
wire_height (1): Vertical spacing between two qubit wires (roughly in inches)
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: Returns True if the CircuitDrawerMatplotlib is the last engine (since it can print any command).
 Parameters
cmd (Command) – Command for which to check availability (all Commands can be printed).
 Returns
True, unless the next engine cannot handle the Command (if there is a next engine).
 Return type
availability (bool)
 class projectq.backends.ClassicalSimulator[source]
A simple introspective simulator that only permits classical operations.
Allows allocation, deallocation, measuring (noop), flushing (noop), controls, NOTs, and any BasicMathGate. Supports reading/writing directly from/to bits and registers of bits.
 read_bit(qubit)[source]
Read a bit.
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 Parameters
qubit (projectq.types.Qubit) – The bit to read.
 Returns
0 if the target bit is off, 1 if it’s on.
 Return type
int
 read_register(qureg)[source]
Read a group of bits as a littleendian integer.
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 Parameters
qureg (projectq.types.Qureg) – The group of bits to read, in littleendian order.
 Returns
Littleendian register value.
 Return type
int
 receive(command_list)[source]
Receive a list of commands.
This implementation simply forwards all commands to the next engine.
 write_bit(qubit, value)[source]
Resets/sets a bit to the given value.
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 Parameters
qubit (projectq.types.Qubit) – The bit to write.
value (boolint) – Writes 1 if this value is truthy, else 0.
 write_register(qureg, value)[source]
Set a group of bits to store a littleendian integer value.
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 Parameters
qureg (projectq.types.Qureg) – The bits to write, in littleendian order.
value (int) – The integer value to store. Must fit in the register.
 class projectq.backends.CommandPrinter(accept_input=True, default_measure=False, in_place=False)[source]
Compiler engine that prints command to the standard output.
CommandPrinter is a compiler engine which prints commands to stdout prior to sending them on to the next compiler engine.
 __init__(accept_input=True, default_measure=False, in_place=False)[source]
Initialize a CommandPrinter.
 Parameters
accept_input (bool) – If accept_input is true, the printer queries the user to input measurement results if the CommandPrinter is the last engine. Otherwise, all measurements yield default_measure.
default_measure (bool) – Default measurement result (if accept_input is False).
in_place (bool) – If in_place is true, all output is written on the same line of the terminal.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: Returns True if the CommandPrinter is the last engine (since it can print any command).
 Parameters
cmd (Command) – Command of which to check availability (all Commands can be printed).
 Returns
True, unless the next engine cannot handle the Command (if there is a next engine).
 Return type
availability (bool)
 exception projectq.backends.DeviceNotHandledError[source]
Exception raised if a selected device cannot handle the circuit or is not supported by ProjectQ.
 exception projectq.backends.DeviceOfflineError[source]
Raised when a device is required but is currently offline.
 exception projectq.backends.DeviceTooSmall[source]
Raised when a device does not have enough qubits for a desired job.
 class projectq.backends.IBMBackend(use_hardware=False, num_runs=1024, verbose=False, token='', device='ibmq_essex', num_retries=3000, interval=1, retrieve_execution=None)[source]
Define the compiler engine class that handles interactions with the IBM API.
The IBM Backend class, which stores the circuit, transforms it to JSON, and sends the circuit through the IBM API.
 __init__(use_hardware=False, num_runs=1024, verbose=False, token='', device='ibmq_essex', num_retries=3000, interval=1, retrieve_execution=None)[source]
Initialize the Backend object.
 Parameters
use_hardware (bool) – If True, the code is run on the IBM quantum chip (instead of using the IBM simulator)
num_runs (int) – Number of runs to collect statistics. (default is 1024)
verbose (bool) – If True, statistics are printed, in addition to the measurement result being registered (at the end of the circuit).
token (str) – IBM quantum experience user password.
device (str) – name of the IBM device to use. ibmq_essex By default
num_retries (int) – Number of times to retry to obtain results from the IBM API. (default is 3000)
interval (float, int) – Number of seconds between successive attempts to obtain results from the IBM API. (default is 1)
retrieve_execution (int) – Job ID to retrieve instead of re running the circuit (e.g., if previous run timed out).
 get_probabilities(qureg)[source]
Return the probability of the outcome bit_string when measuring the quantum register qureg.
Return the list of basis states with corresponding probabilities. If input qureg is a subset of the register used for the experiment, then returns the projected probabilities over the other states.
The measured bits are ordered according to the supplied quantum register, i.e., the leftmost bit in the statestring corresponds to the first qubit in the supplied quantum register.
Warning
Only call this function after the circuit has been executed!
 Parameters
qureg (list<Qubit>) – Quantum register determining the order of the qubits.
 Returns
Dictionary mapping nbit strings to probabilities.
 Return type
probability_dict (dict)
 Raises
RuntimeError – If no data is available (i.e., if the circuit has not been executed). Or if a qubit was supplied which was not present in the circuit (might have gotten optimized away).
 get_qasm()[source]
Return the QASM representation of the circuit sent to the backend.
Should be called AFTER calling the ibm device.
 class projectq.backends.IonQBackend(use_hardware=False, num_runs=100, verbose=False, token=None, device='ionq_simulator', num_retries=3000, interval=1, retrieve_execution=None)[source]
Backend for building circuits and submitting them to the IonQ API.
 __init__(use_hardware=False, num_runs=100, verbose=False, token=None, device='ionq_simulator', num_retries=3000, interval=1, retrieve_execution=None)[source]
Initialize an IonQBackend object.
 Parameters
use_hardware (bool, optional) – Whether or not to use real IonQ hardware or just a simulator. If False, the ionq_simulator is used regardless of the value of
device
. Defaults to False.num_runs (int, optional) – Number of times to run circuits. Defaults to 100.
verbose (bool, optional) – If True, print statistics after job results have been collected. Defaults to False.
token (str, optional) – An IonQ API token. Defaults to None.
device (str, optional) – Device to run jobs on. Supported devices are
'ionq_qpu'
or'ionq_simulator'
. Defaults to'ionq_simulator'
.num_retries (int, optional) – Number of times to retry fetching a job after it has been submitted. Defaults to 3000.
interval (int, optional) – Number of seconds to wait inbetween result fetch retries. Defaults to 1.
retrieve_execution (str, optional) – An IonQ API Job ID. If provided, a job with this ID will be fetched. Defaults to None.
 get_probabilities(qureg)[source]
Given the provided qubit register, determine the probability of each possible outcome.
Note
This method should only be called after a circuit has been run and its results are available.
 Parameters
qureg (Qureg) – A ProjectQ Qureg object.
 Returns
A dict mapping of states > probability.
 Return type
dict
 get_probability(state, qureg)[source]
Shortcut to get a specific state’s probability.
 Parameters
state (str) – A state in bitstring format.
qureg (Qureg) – A ProjectQ Qureg object.
 Returns
The probability for the provided state.
 Return type
float
 class projectq.backends.ResourceCounter[source]
ResourceCounter is a compiler engine which counts the number of gates and max. number of active qubits.
 gate_counts
Dictionary of gate counts. The keys are tuples of the form (cmd.gate, ctrl_cnt), where ctrl_cnt is the number of control qubits.
 Type
dict
 gate_class_counts
Dictionary of gate class counts. The keys are tuples of the form (cmd.gate.__class__, ctrl_cnt), where ctrl_cnt is the number of control qubits.
 Type
dict
 max_width
Maximal width (=max. number of active qubits at any given point).
 Type
int
 Properties:
depth_of_dag (int): It is the longest path in the directed acyclic graph (DAG) of the program.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: Returns True if the ResourceCounter is the last engine (since it can count any command).
 Parameters
cmd (Command) – Command for which to check availability (all Commands can be counted).
 Returns
True, unless the next engine cannot handle the Command (if there is a next engine).
 Return type
availability (bool)
 class projectq.backends.Simulator(gate_fusion=False, rnd_seed=None)[source]
Simulator is a compiler engine which simulates a quantum computer using C++based kernels.
OpenMP is enabled and the number of threads can be controlled using the OMP_NUM_THREADS environment variable, i.e.
export OMP_NUM_THREADS=4 # use 4 threads export OMP_PROC_BIND=spread # bind threads to processors by spreading
 __init__(gate_fusion=False, rnd_seed=None)[source]
Construct the C++/Pythonsimulator object and initialize it with a random seed.
 Parameters
gate_fusion (bool) – If True, gates are cached and only executed once a certain gatesize has been reached (only has an effect for the c++ simulator).
rnd_seed (int) – Random seed (uses random.randint(0, 4294967295) by default).
Example of gate_fusion: Instead of applying a Hadamard gate to 5 qubits, the simulator calculates the kronecker product of the 1qubit gate matrices and then applies one 5qubit gate. This increases operational intensity and keeps the simulator from having to iterate through the state vector multiple times. Depending on the system (and, especially, number of threads), this may or may not be beneficial.
Note
If the C++ Simulator extension was not built or cannot be found, the Simulator defaults to a Python implementation of the kernels. While this is much slower, it is still good enough to run basic quantum algorithms.
If you need to run large simulations, check out the tutorial in the docs which gives futher hints on how to build the C++ extension.
 apply_qubit_operator(qubit_operator, qureg)[source]
Apply a (possibly nonunitary) qubit_operator to the current wave function represented by a quantum register.
 Parameters
qubit_operator (projectq.ops.QubitOperator) – Operator to apply.
qureg (list[Qubit],Qureg) – Quantum bits to which to apply the operator.
 Raises
Exception – If qubit_operator acts on more qubits than present in the qureg argument.
Warning
This function allows applying nonunitary gates and it will not renormalize the wave function! It is for numerical experiments only and should not be used for other purposes.
Note
Make sure all previous commands (especially allocations) have passed through the compilation chain (call main_engine.flush() to make sure).
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 cheat()[source]
Access the ordering of the qubits and the state vector directly.
This is a cheat function which enables, e.g., more efficient evaluation of expectation values and debugging.
 Returns
A tuple where the first entry is a dictionary mapping qubit indices to bitlocations and the second entry is the corresponding state vector.
Note
Make sure all previous commands have passed through the compilation chain (call main_engine.flush() to make sure).
Note
If there is a mapper present in the compiler, this function DOES NOT automatically convert from logical qubits to mapped qubits.
 collapse_wavefunction(qureg, values)[source]
Collapse a quantum register onto a classical basis state.
 Parameters
 Raises
RuntimeError – If an outcome has probability (approximately) 0 or if unknown qubits are provided (see note).
Note
Make sure all previous commands have passed through the compilation chain (call main_engine.flush() to make sure).
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 get_amplitude(bit_string, qureg)[source]
Return the probability amplitude of the supplied bit_string.
The ordering is given by the quantum register qureg, which must contain all allocated qubits.
 Parameters
 Returns
Probability amplitude of the provided bit string.
Note
Make sure all previous commands (especially allocations) have passed through the compilation chain (call main_engine.flush() to make sure).
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 get_expectation_value(qubit_operator, qureg)[source]
Return the expectation value of a qubit operator.
Get the expectation value of qubit_operator w.r.t. the current wave function represented by the supplied quantum register.
 Parameters
qubit_operator (projectq.ops.QubitOperator) – Operator to measure.
 Returns
Expectation value
Note
Make sure all previous commands (especially allocations) have passed through the compilation chain (call main_engine.flush() to make sure).
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 Raises
Exception – If qubit_operator acts on more qubits than present in the qureg argument.
 get_probability(bit_string, qureg)[source]
Return the probability of the outcome bit_string when measuring the quantum register qureg.
 Parameters
 Returns
Probability of measuring the provided bit string.
Note
Make sure all previous commands (especially allocations) have passed through the compilation chain (call main_engine.flush() to make sure).
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: The simulator can deal with all arbitrarilycontrolled gates which provide a gatematrix (via gate.matrix) and acts on 5 or less qubits (not counting the control qubits).
 Parameters
cmd (Command) – Command for which to check availability (single qubit gate, arbitrary controls)
 Returns
True if it can be simulated and False otherwise.
 receive(command_list)[source]
Receive a list of commands.
Receive a list of commands from the previous engine and handle them (simulate them classically) prior to sending them on to the next engine.
 Parameters
command_list (list<Command>) – List of commands to execute on the simulator.
 set_wavefunction(wavefunction, qureg)[source]
Set the wavefunction and the qubit ordering of the simulator.
The simulator will adopt the ordering of qureg (instead of reordering the wavefunction).
 Parameters
Note
Make sure all previous commands (especially allocations) have passed through the compilation chain (call main_engine.flush() to make sure).
Note
If there is a mapper present in the compiler, this function automatically converts from logical qubits to mapped qubits for the qureg argument.
 class projectq.backends.UnitarySimulator[source]
Simulator engine aimed at calculating the unitary transformation that represents the current quantum circuit.
 unitary[source]
Current unitary representing the quantum circuit being processed so far.
 Type
np.ndarray
Note
The current implementation of this backend resets the unitary after the first gate that is neither a qubit deallocation nor a measurement occurs after one of those two aforementioned gates.
The old unitary call be accessed at anytime after such a situation occurs via the history property.
eng = MainEngine(backend=UnitarySimulator(), engine_list=[]) qureg = eng.allocate_qureg(3) All(X)  qureg eng.flush() All(Measure)  qureg eng.deallocate_qubit(qureg[1]) X  qureg[0] # WARNING: appending gate after measurements or deallocations resets the unitary
 property history[source]
Access all previous unitary matrices.
The current unitary matrix is appended to this list once a gate is received after either a measurement or a qubit deallocation has occurred.
 Returns
A list where the elements are all previous unitary matrices representing the circuit, separated by measurement/deallocate gates.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Specialized implementation of is_available: The unitary simulator can deal with all arbitrarilycontrolled gates which provide a gatematrix (via gate.matrix).
 Parameters
cmd (Command) – Command for which to check availability (single qubit gate, arbitrary controls)
 Returns
True if it can be simulated and False otherwise.
 measure_qubits(ids)[source]
Measure the qubits with IDs ids and return a list of measurement outcomes (True/False).
 Parameters
ids (list<int>) – List of qubit IDs to measure.
 Returns
List of measurement results (containing either True or False).
 receive(command_list)[source]
Receive a list of commands.
 Receive a list of commands from the previous engine and handle them:
update the unitary of the quantum circuit
update the internal quantum state if a measurement or a qubit deallocation occurs
prior to sending them on to the next engine.
 Parameters
command_list (list<Command>) – List of commands to execute on the simulator.
cengines
The ProjectQ compiler engines package.
The parent class from which all mappers should be derived. 

Module containing the basic definition of a compiler engine. 

A CommandModifier engine that can be used to apply a userdefined transformation to all incoming commands. 

Contains a compiler engine to map to the 5qubit IBM chip. 

Mapper for a quantum circuit to a linear chain of qubits. 

The main engine of every compiler engine pipeline, called MainEngine. 

A compiler engine to add mapping information. 

A local optimizer engine. 

A compiler engine which flips the directionality of CNOTs according to the given connectivity graph. 

The TagRemover compiler engine. 

TestEngine and DummyEngine. 

Mapper for a quantum circuit to a 2D square grid. 


A compiler engine to automatically replace certain commands. 
Basic compiler engine: All compiler engines are derived from this class. 

Parent class for all Mappers. 


Compiler engine applying a userdefined transformation to all incoming commands. 
Command list comparison compiler engine for testing purposes. 


A rule for breaking down specific gates into sequences of simpler gates. 
A collection of indexed decomposition rules. 


DummyEngine used for testing. 

A ForwarderEngine is a trivial engine which forwards all commands to the next engine. 

Mapper to a 2D grid graph. 

Mapper for the 5qubit IBM backend. 

A compiler engine that implements a userdefined is_available() method. 
Exception thrown when the last engine tries to access the next one. 


Map a quantum circuit to a linear chain of nearest neighbour interactions. 

Circuit optimization compiler engine. 

The MainEngine class provides all functionality of the main compiler engine. 

Manual Mapper which adds QubitPlacementTags to Allocate gate commands according to a userspecified mapping. 
Exception raised when trying to access the measurement value of a qubit that has not yet been measured. 

Return the circuit depth to execute these swaps. 

Flip CNOTs and translates Swaps to CNOTs where necessary. 


Compiler engine that remove temporary command tags. 
Exception raised when a nonsupported compiler engine is encountered. 
Submodules
_basicmapper
The parent class from which all mappers should be derived.
There is only one engine currently allowed to be derived from BasicMapperEngine. This allows the simulator to automatically translate logical qubit ids to mapped ids.
_basics
Module containing the basic definition of a compiler engine.
 class projectq.cengines._basics.BasicEngine[source]
Basic compiler engine: All compiler engines are derived from this class.
It provides basic functionality such as qubit allocation/deallocation and functions that provide information about the engine’s position (e.g., next engine).
This information is provided by the MainEngine, which initializes all further engines.
 next_engine
Next compiler engine (or the backend).
 Type
 main_engine
Reference to the main compiler engine.
 Type
 is_last_engine
True for the last engine, which is the backend.
 Type
bool
 allocate_qubit(dirty=False)[source]
Return a new qubit as a list containing 1 qubit object (quantum register of size 1).
Allocates a new qubit by getting a (new) qubit id from the MainEngine, creating the qubit object, and then sending an AllocateQubit command down the pipeline. If dirty=True, the fresh qubit can be replaced by a preallocated one (in an unknown, dirty, initial state). Dirty qubits must be returned to their initial states before they are deallocated / freed.
All allocated qubits are added to the MainEngine’s set of active qubits as weak references. This allows proper cleanup at the end of the Python program (using atexit), deallocating all qubits which are still alive. Qubit ids of dirty qubits are registered in MainEngine’s dirty_qubits set.
 Parameters
dirty (bool) – If True, indicates that the allocated qubit may be dirty (i.e., in an arbitrary initial state).
 Returns
Qureg of length 1, where the first entry is the allocated qubit.
 allocate_qureg(n_qubits)[source]
Allocate n qubits and return them as a quantum register, which is a list of qubit objects.
 Parameters
n (int) – Number of qubits to allocate
 Returns
Qureg of length n, a list of n newly allocated qubits.
 deallocate_qubit(qubit)[source]
Deallocate a qubit (and sends the deallocation command down the pipeline).
If the qubit was allocated as a dirty qubit, add DirtyQubitTag() to Deallocate command.
 Parameters
qubit (BasicQubit) – Qubit to deallocate.
 Raises
ValueError – Qubit already deallocated. Caller likely has a bug.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Default implementation of is_available: Ask the next engine whether a command is available, i.e., whether it can be executed by the next engine(s).
 Parameters
cmd (Command) – Command for which to check availability.
 Returns
True if the command can be executed.
 Raises
LastEngineException – If is_last_engine is True but is_available is not implemented.
 is_meta_tag_supported(meta_tag)[source]
Check if there is a compiler engine handling the meta tag.
 Parameters
engine – First engine to check (then iteratively calls getNextEngine)
meta_tag – Meta tag class for which to check support
 Returns
True if one of the further compiler engines is a meta tag handler, i.e., engine.is_meta_tag_handler(meta_tag) returns True.
 Return type
supported (bool)
 class projectq.cengines._basics.ForwarderEngine(engine, cmd_mod_fun=None)[source]
A ForwarderEngine is a trivial engine which forwards all commands to the next engine.
It is mainly used as a substitute for the MainEngine at lower levels such that meta operations still work (e.g., with Compute).
 exception projectq.cengines._basics.LastEngineException(engine)[source]
Exception thrown when the last engine tries to access the next one. (Next engine does not exist).
The default implementation of isAvailable simply asks the next engine whether the command is available. An engine which legally may be the last engine, this behavior needs to be adapted (see BasicEngine.isAvailable).
_cmdmodifier
A CommandModifier engine that can be used to apply a userdefined transformation to all incoming commands.
A CommandModifier engine can be used to, e.g., modify the tags of all commands which pass by (see the AutoReplacer for an example).
 class projectq.cengines._cmdmodifier.CommandModifier(cmd_mod_fun)[source]
Compiler engine applying a userdefined transformation to all incoming commands.
CommandModifier is a compiler engine which applies a function to all incoming commands, sending on the resulting command instead of the original one.
_ibm5qubitmapper
Contains a compiler engine to map to the 5qubit IBM chip.
 class projectq.cengines._ibm5qubitmapper.IBM5QubitMapper(connections=None)[source]
Mapper for the 5qubit IBM backend.
Maps a given circuit to the IBM Quantum Experience chip.
Note
The mapper has to be run once on the entire circuit.
Warning
If the provided circuit cannot be mapped to the hardware layout without performing Swaps, the mapping procedure raises an Exception.
 is_available(cmd)[source]
Check if the IBM backend can perform the Command cmd and return True if so.
 Parameters
cmd (Command) – The command to check
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and, for each command, stores it until completion.
 Parameters
command_list (list of Command objects) – list of commands to receive.
 Raises
Exception – If mapping the CNOT gates to 1 qubit would require Swaps. The current version only supports remapping of CNOT gates without performing any Swaps due to the large costs associated with Swapping given the CNOT constraints.
_linearmapper
Mapper for a quantum circuit to a linear chain of qubits.
 Input: Quantum circuit with 1 and 2 qubit gates on n qubits. Gates are assumed to be applied in parallel if they act
on disjoint qubit(s) and any pair of qubits can perform a 2 qubit gate (alltoall connectivity)
 Output: Quantum circuit in which qubits are placed in 1D chain in which only nearest neighbour qubits can perform a 2
qubit gate. The mapper uses Swap gates in order to move qubits next to each other.
 class projectq.cengines._linearmapper.LinearMapper(num_qubits, cyclic=False, storage=1000)[source]
Map a quantum circuit to a linear chain of nearest neighbour interactions.
Maps a quantum circuit to a linear chain of qubits with nearest neighbour interactions using Swap gates. It supports open or cyclic boundary conditions.
 current_mapping[source]
Stores the mapping: key is logical qubit id, value is mapped qubit id from 0,…,self.num_qubits
 cyclic
If chain is cyclic or not
 Type
Bool
 storage
Number of gate it caches before mapping.
 Type
int
 num_mappings
Number of times the mapper changed the mapping
 Type
int
 depth_of_swaps
Key are circuit depth of swaps, value is the number of such mappings which have been applied
 Type
dict
 num_of_swaps_per_mapping
Key are the number of swaps per mapping, value is the number of such mappings which have been applied
 Type
dict
Note
Gates are cached and only mapped from time to time. A FastForwarding gate doesn’t empty the cache, only a FlushGate does.
Only 1 and two qubit gates allowed.
Does not optimize for dirty qubits.
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and, for each command, stores it until we do a mapping (FlushGate or Cache of stored commands is full).
 Parameters
command_list (list of Command objects) – list of commands to receive.
 static return_new_mapping(num_qubits, cyclic, currently_allocated_ids, stored_commands, current_mapping)[source]
Build a mapping of qubits to a linear chain.
It goes through stored_commands and tries to find a mapping to apply these gates on a first come first served basis. More compilicated scheme could try to optimize to apply as many gates as possible between the Swaps.
 Parameters
num_qubits (int) – Total number of qubits in the linear chain
cyclic (bool) – If linear chain is a cycle.
currently_allocated_ids (set of int) – Logical qubit ids for which the Allocate gate has already been processed and sent to the next engine but which are not yet deallocated and hence need to be included in the new mapping.
stored_commands (list of Command objects) – Future commands which should be applied next.
current_mapping – A current mapping as a dict. key is logical qubit id, value is placement id. If there are different possible maps, this current mapping is used to minimize the swaps to go to the new mapping by a heuristic.
 Returns: A new mapping as a dict. key is logical qubit id,
value is placement id
_main
The main engine of every compiler engine pipeline, called MainEngine.
 class projectq.cengines._main.MainEngine(backend=None, engine_list=None, verbose=False)[source]
The MainEngine class provides all functionality of the main compiler engine.
It initializes all further compiler engines (calls, e.g., .next_engine=…) and keeps track of measurement results and active qubits (and their IDs).
 next_engine
Next compiler engine (or the backend).
 Type
 main_engine
Self.
 Type
 active_qubits
WeakSet containing all active qubits
 Type
WeakSet
 dirty_qubits
Containing all dirty qubit ids
 Type
Set
 backend
Access the backend.
 Type
 mapper
Access to the mapper if there is one.
 Type
 n_engines
Current number of compiler engines in the engine list
 Type
int
 n_engines_max
Maximum number of compiler engines allowed in the engine list. Defaults to 100.
 Type
int
 flush(deallocate_qubits=False)[source]
Flush the entire circuit down the pipeline, clearing potential buffers (of, e.g., optimizers).
 Parameters
deallocate_qubits (bool) – If True, deallocates all qubits that are still alive (invalidating references to them by setting their id to 1).
 get_measurement_result(qubit)[source]
Return the classical value of a measured qubit, given that an engine registered this result previously.
See also setMeasurementResult.
 Parameters
qubit (BasicQubit) – Qubit of which to get the measurement result.
Example
from projectq.ops import H, Measure from projectq import MainEngine eng = MainEngine() qubit = eng.allocate_qubit() # quantum register of size 1 H  qubit Measure  qubit eng.get_measurement_result(qubit[0]) == int(qubit)
 get_new_qubit_id()[source]
Return a unique qubit id to be used for the next qubit allocation.
 Returns
New unique qubit id.
 Return type
new_qubit_id (int)
 receive(command_list)[source]
Forward the list of commands to the first engine.
 Parameters
command_list (list<Command>) – List of commands to receive (and then send on)
 send(command_list)[source]
Forward the list of commands to the next engine in the pipeline.
It also shortens exception stack traces if self.verbose is False.
 set_measurement_result(qubit, value)[source]
Register a measurement result.
The engine being responsible for measurement results needs to register these results with the master engine such that they are available when the user calls an int() or bool() conversion operator on a measured qubit.
 Parameters
qubit (BasicQubit) – Qubit for which to register the measurement result.
value (bool) – Boolean value of the measurement outcome (True / False = 1 / 0 respectively).
_manualmapper
A compiler engine to add mapping information.
 class projectq.cengines._manualmapper.ManualMapper(map_fun=<function ManualMapper.<lambda>>)[source]
Manual Mapper which adds QubitPlacementTags to Allocate gate commands according to a userspecified mapping.
 map
The function which maps a given qubit id to its location. It gets set when initializing the mapper.
 Type
function
_optimize
A local optimizer engine.
 class projectq.cengines._optimize.LocalOptimizer(cache_size=5, m=None)[source]
Circuit optimization compiler engine.
LocalOptimizer is a compiler engine which optimizes locally (merging rotations, cancelling gates with their inverse) in a local window of user defined size.
It stores all commands in a dict of lists, where each qubit has its own gate pipeline. After adding a gate, it tries to merge / cancel successive gates using the get_merged and get_inverse functions of the gate (if available). For examples, see BasicRotationGate. Once a list corresponding to a qubit contains >=m gates, the pipeline is sent on to the next engine.
_replacer
_swapandcnotflipper
A compiler engine which flips the directionality of CNOTs according to the given connectivity graph.
It also translates Swap gates to CNOTs if necessary.
 class projectq.cengines._swapandcnotflipper.SwapAndCNOTFlipper(connectivity)[source]
Flip CNOTs and translates Swaps to CNOTs where necessary.
Warning
This engine assumes that CNOT and Hadamard gates are supported by the following engines.
Warning
This engine cannot be used as a backend.
 is_available(cmd)[source]
Check if the IBM backend can perform the Command cmd and return True if so.
 Parameters
cmd (Command) – The command to check
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and if the command is a CNOT gate, it flips it using Hadamard gates if necessary; if it is a Swap gate, it decomposes it using 3 CNOTs. All other gates are simply sent to the next engine.
 Parameters
command_list (list of Command objects) – list of commands to receive.
_tagremover
The TagRemover compiler engine.
A TagRemover engine removes temporary command tags (such as Compute/Uncompute), thus enabling optimization across meta statements (loops after unrolling, compute/uncompute, …)
 class projectq.cengines._tagremover.TagRemover(tags=None)[source]
Compiler engine that remove temporary command tags.
TagRemover is a compiler engine which removes temporary command tags (see the tag classes such as LoopTag in projectq.meta._loop).
Removing tags is important (after having handled them if necessary) in order to enable optimizations across metafunction boundaries (compute/ action/uncompute or loops after unrolling)
 receive(command_list)[source]
Receive a list of commands.
Receive a list of commands from the previous engine, remove all tags which are an instance of at least one of the meta tags provided in the constructor, and then send them on to the next compiler engine.
 Parameters
command_list (list<Command>) – List of commands to receive and then (after removing tags) send on.
_testengine
TestEngine and DummyEngine.
 class projectq.cengines._testengine.CompareEngine[source]
Command list comparison compiler engine for testing purposes.
CompareEngine is an engine which saves all commands. It is only intended for testing purposes. Two CompareEngine backends can be compared and return True if they contain the same commmands.
 class projectq.cengines._testengine.DummyEngine(save_commands=False)[source]
DummyEngine used for testing.
The DummyEngine forwards all commands directly to next engine. If self.is_last_engine == True it just discards all gates. By setting save_commands == True all commands get saved as a list in self.received_commands. Elements are appended to this list so they are ordered according to when they are received.
_twodmapper
Mapper for a quantum circuit to a 2D square grid.
 Input: Quantum circuit with 1 and 2 qubit gates on n qubits. Gates are assumed to be applied in parallel if they act
on disjoint qubit(s) and any pair of qubits can perform a 2 qubit gate (alltoall connectivity)
 Output: Quantum circuit in which qubits are placed in 2D square grid in which only nearest neighbour qubits can
perform a 2 qubit gate. The mapper uses Swap gates in order to move qubits next to each other.
 class projectq.cengines._twodmapper.GridMapper(num_rows, num_columns, mapped_ids_to_backend_ids=None, storage=1000, optimization_function=<function return_swap_depth>, num_optimization_steps=50)[source]
Mapper to a 2D grid graph.
Mapped qubits on the grid are numbered in rowmajor order. E.g. for 3 rows and 2 columns:
0  1   2  3   4  5
The numbers are the mapped qubit ids. The backend might number the qubits on the grid differently (e.g. not rowmajor), we call these backend qubit ids. If the backend qubit ids are not rowmajor, one can pass a dictionary translating from our rowmajor mapped ids to these backend ids.
Note: The algorithm sorts twice inside each column and once inside each row.
 storage
Number of gate it caches before mapping.
 Type
int
 num_rows
Number of rows in the grid
 Type
int
 num_columns
Number of columns in the grid
 Type
int
 num_qubits
num_rows x num_columns = number of qubits
 Type
int
 num_mappings
Number of times the mapper changed the mapping
 Type
int
 depth_of_swaps
Key are circuit depth of swaps, value is the number of such mappings which have been applied
 Type
dict
 num_of_swaps_per_mapping
Key are the number of swaps per mapping, value is the number of such mappings which have been applied
 Type
dict
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and, for each command, stores it until we do a mapping (FlushGate or Cache of stored commands is full).
 Parameters
command_list (list of Command objects) – list of commands to receive.
 return_swaps(old_mapping, new_mapping, permutation=None)[source]
Return the swap operation to change mapping.
 Parameters
old_mapping – dict: keys are logical ids and values are mapped qubit ids
new_mapping – dict: keys are logical ids and values are mapped qubit ids
permutation – list of int from 0, 1, …, self.num_rows1. It is used to permute the found perfect matchings. Default is None which keeps the original order.
 Returns
List of tuples. Each tuple is a swap operation which needs to be applied. Tuple contains the two mapped qubit ids for the Swap.
Module contents
ProjectQ module containing all compiler engines.
 class projectq.cengines.AutoReplacer(decomposition_rule_se, decomposition_chooser=<function AutoReplacer.<lambda>>)[source]
A compiler engine to automatically replace certain commands.
The AutoReplacer is a compiler engine which uses engine.is_available in order to determine which commands need to be replaced/decomposed/compiled further. The loaded setup is used to find decomposition rules appropriate for each command (e.g., setups.default).
 __init__(decomposition_rule_se, decomposition_chooser=<function AutoReplacer.<lambda>>)[source]
Initialize an AutoReplacer.
 Parameters
decomposition_chooser (function) – A function which, given the Command to decompose and a list of potential Decomposition objects, determines (and then returns) the ‘best’ decomposition.
The default decomposition chooser simply returns the first list element, i.e., calling
repl = AutoReplacer()
Amounts to
def decomposition_chooser(cmd, decomp_list): return decomp_list[0] repl = AutoReplacer(decomposition_chooser)
 class projectq.cengines.BasicEngine[source]
Basic compiler engine: All compiler engines are derived from this class.
It provides basic functionality such as qubit allocation/deallocation and functions that provide information about the engine’s position (e.g., next engine).
This information is provided by the MainEngine, which initializes all further engines.
 next_engine
Next compiler engine (or the backend).
 Type
 main_engine
Reference to the main compiler engine.
 Type
 is_last_engine
True for the last engine, which is the backend.
 Type
bool
 __init__()[source]
Initialize the basic engine.
Initializes local variables such as _next_engine, _main_engine, etc. to None.
 allocate_qubit(dirty=False)[source]
Return a new qubit as a list containing 1 qubit object (quantum register of size 1).
Allocates a new qubit by getting a (new) qubit id from the MainEngine, creating the qubit object, and then sending an AllocateQubit command down the pipeline. If dirty=True, the fresh qubit can be replaced by a preallocated one (in an unknown, dirty, initial state). Dirty qubits must be returned to their initial states before they are deallocated / freed.
All allocated qubits are added to the MainEngine’s set of active qubits as weak references. This allows proper cleanup at the end of the Python program (using atexit), deallocating all qubits which are still alive. Qubit ids of dirty qubits are registered in MainEngine’s dirty_qubits set.
 Parameters
dirty (bool) – If True, indicates that the allocated qubit may be dirty (i.e., in an arbitrary initial state).
 Returns
Qureg of length 1, where the first entry is the allocated qubit.
 allocate_qureg(n_qubits)[source]
Allocate n qubits and return them as a quantum register, which is a list of qubit objects.
 Parameters
n (int) – Number of qubits to allocate
 Returns
Qureg of length n, a list of n newly allocated qubits.
 deallocate_qubit(qubit)[source]
Deallocate a qubit (and sends the deallocation command down the pipeline).
If the qubit was allocated as a dirty qubit, add DirtyQubitTag() to Deallocate command.
 Parameters
qubit (BasicQubit) – Qubit to deallocate.
 Raises
ValueError – Qubit already deallocated. Caller likely has a bug.
 is_available(cmd)[source]
Test whether a Command is supported by a compiler engine.
Default implementation of is_available: Ask the next engine whether a command is available, i.e., whether it can be executed by the next engine(s).
 Parameters
cmd (Command) – Command for which to check availability.
 Returns
True if the command can be executed.
 Raises
LastEngineException – If is_last_engine is True but is_available is not implemented.
 is_meta_tag_supported(meta_tag)[source]
Check if there is a compiler engine handling the meta tag.
 Parameters
engine – First engine to check (then iteratively calls getNextEngine)
meta_tag – Meta tag class for which to check support
 Returns
True if one of the further compiler engines is a meta tag handler, i.e., engine.is_meta_tag_handler(meta_tag) returns True.
 Return type
supported (bool)
 class projectq.cengines.BasicMapperEngine[source]
Parent class for all Mappers.
 self.current_mapping
Keys are the logical qubit ids and values are the mapped qubit ids.
 Type
dict
 class projectq.cengines.CommandModifier(cmd_mod_fun)[source]
Compiler engine applying a userdefined transformation to all incoming commands.
CommandModifier is a compiler engine which applies a function to all incoming commands, sending on the resulting command instead of the original one.
 class projectq.cengines.CompareEngine[source]
Command list comparison compiler engine for testing purposes.
CompareEngine is an engine which saves all commands. It is only intended for testing purposes. Two CompareEngine backends can be compared and return True if they contain the same commmands.
 class projectq.cengines.DecompositionRule(gate_class, gate_decomposer, gate_recognizer=<function DecompositionRule.<lambda>>)[source]
A rule for breaking down specific gates into sequences of simpler gates.
 __init__(gate_class, gate_decomposer, gate_recognizer=<function DecompositionRule.<lambda>>)[source]
Initialize a DecompositionRule object.
 Parameters
gate_class (type) –
The type of gate that this rule decomposes.
The gate class is redundant information used to make lookups faster when iterating over a circuit and deciding “which rules apply to this gate?” again and again.
Note that this parameter is a gate type, not a gate instance. You supply gate_class=MyGate or gate_class=MyGate().__class__, not gate_class=MyGate().
gate_decomposer (function[projectq.ops.Command]) – Function which, given the command to decompose, applies a sequence of gates corresponding to the highlevel function of a gate of type gate_class.
(function[projectq.ops.Command] (gate_recognizer) –
boolean): A predicate that determines if the decomposition applies to the given command (on top of the filtering by gate_class).
For example, a decomposition rule may only to apply rotation gates that rotate by a specific angle.
If no gate_recognizer is given, the decomposition applies to all gates matching the gate_class.
 class projectq.cengines.DecompositionRuleSet(rules=None, modules=None)[source]
A collection of indexed decomposition rules.
 __init__(rules=None, modules=None)[source]
Initialize a DecompositionRuleSet object.
 Parameters
list[DecompositionRule] (rules) – Initial decomposition rules.
modules (iterable[ModuleWithDecompositionRuleSet]) – A list of things with an “all_defined_decomposition_rules” property containing decomposition rules to add to the rule set.
 class projectq.cengines.DummyEngine(save_commands=False)[source]
DummyEngine used for testing.
The DummyEngine forwards all commands directly to next engine. If self.is_last_engine == True it just discards all gates. By setting save_commands == True all commands get saved as a list in self.received_commands. Elements are appended to this list so they are ordered according to when they are received.
 class projectq.cengines.ForwarderEngine(engine, cmd_mod_fun=None)[source]
A ForwarderEngine is a trivial engine which forwards all commands to the next engine.
It is mainly used as a substitute for the MainEngine at lower levels such that meta operations still work (e.g., with Compute).
 __init__(engine, cmd_mod_fun=None)[source]
Initialize a ForwarderEngine.
 Parameters
engine (BasicEngine) – Engine to forward all commands to.
cmd_mod_fun (function) – Function which is called before sending a command. Each command cmd is replaced by the command it returns when getting called with cmd.
 class projectq.cengines.GridMapper(num_rows, num_columns, mapped_ids_to_backend_ids=None, storage=1000, optimization_function=<function return_swap_depth>, num_optimization_steps=50)[source]
Mapper to a 2D grid graph.
Mapped qubits on the grid are numbered in rowmajor order. E.g. for 3 rows and 2 columns:
0  1   2  3   4  5
The numbers are the mapped qubit ids. The backend might number the qubits on the grid differently (e.g. not rowmajor), we call these backend qubit ids. If the backend qubit ids are not rowmajor, one can pass a dictionary translating from our rowmajor mapped ids to these backend ids.
Note: The algorithm sorts twice inside each column and once inside each row.
 storage
Number of gate it caches before mapping.
 Type
int
 num_rows
Number of rows in the grid
 Type
int
 num_columns
Number of columns in the grid
 Type
int
 num_qubits
num_rows x num_columns = number of qubits
 Type
int
 num_mappings
Number of times the mapper changed the mapping
 Type
int
 depth_of_swaps
Key are circuit depth of swaps, value is the number of such mappings which have been applied
 Type
dict
 num_of_swaps_per_mapping
Key are the number of swaps per mapping, value is the number of such mappings which have been applied
 Type
dict
 __init__(num_rows, num_columns, mapped_ids_to_backend_ids=None, storage=1000, optimization_function=<function return_swap_depth>, num_optimization_steps=50)[source]
Initialize a GridMapper compiler engine.
 Parameters
num_rows (int) – Number of rows in the grid
num_columns (int) – Number of columns in the grid.
mapped_ids_to_backend_ids (dict) – Stores a mapping from mapped ids which are 0,…,self.num_qubits1 in rowmajor order on the grid to the corresponding qubit ids of the backend. Key: mapped id. Value: corresponding backend id. Default is None which means backend ids are identical to mapped ids.
storage – Number of gates to temporarily store
optimization_function – Function which takes a list of swaps and returns a cost value. Mapper chooses a permutation which minimizes this cost. Default optimizes for circuit depth.
num_optimization_steps (int) – Number of different permutations to of the matching to try and minimize the cost.
 Raises
RuntimeError – if incorrect mapped_ids_to_backend_ids parameter
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and, for each command, stores it until we do a mapping (FlushGate or Cache of stored commands is full).
 Parameters
command_list (list of Command objects) – list of commands to receive.
 return_swaps(old_mapping, new_mapping, permutation=None)[source]
Return the swap operation to change mapping.
 Parameters
old_mapping – dict: keys are logical ids and values are mapped qubit ids
new_mapping – dict: keys are logical ids and values are mapped qubit ids
permutation – list of int from 0, 1, …, self.num_rows1. It is used to permute the found perfect matchings. Default is None which keeps the original order.
 Returns
List of tuples. Each tuple is a swap operation which needs to be applied. Tuple contains the two mapped qubit ids for the Swap.
 class projectq.cengines.IBM5QubitMapper(connections=None)[source]
Mapper for the 5qubit IBM backend.
Maps a given circuit to the IBM Quantum Experience chip.
Note
The mapper has to be run once on the entire circuit.
Warning
If the provided circuit cannot be mapped to the hardware layout without performing Swaps, the mapping procedure raises an Exception.
 __init__(connections=None)[source]
Initialize an IBM 5qubit mapper compiler engine.
Resets the mapping.
 is_available(cmd)[source]
Check if the IBM backend can perform the Command cmd and return True if so.
 Parameters
cmd (Command) – The command to check
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and, for each command, stores it until completion.
 Parameters
command_list (list of Command objects) – list of commands to receive.
 Raises
Exception – If mapping the CNOT gates to 1 qubit would require Swaps. The current version only supports remapping of CNOT gates without performing any Swaps due to the large costs associated with Swapping given the CNOT constraints.
 class projectq.cengines.InstructionFilter(filterfun)[source]
A compiler engine that implements a userdefined is_available() method.
The InstructionFilter is a compiler engine which changes the behavior of is_available according to a filter function. All commands are passed to this function, which then returns whether this command can be executed (True) or needs replacement (False).
 __init__(filterfun)[source]
Initialize an InstructionFilter object.
Initializer: The provided filterfun returns True for all commands which do not need replacement and False for commands that do.
 Parameters
filterfun (function) – Filter function which returns True for available commands, and False otherwise. filterfun will be called as filterfun(self, cmd).
 exception projectq.cengines.LastEngineException(engine)[source]
Exception thrown when the last engine tries to access the next one. (Next engine does not exist).
The default implementation of isAvailable simply asks the next engine whether the command is available. An engine which legally may be the last engine, this behavior needs to be adapted (see BasicEngine.isAvailable).
 class projectq.cengines.LinearMapper(num_qubits, cyclic=False, storage=1000)[source]
Map a quantum circuit to a linear chain of nearest neighbour interactions.
Maps a quantum circuit to a linear chain of qubits with nearest neighbour interactions using Swap gates. It supports open or cyclic boundary conditions.
 current_mapping[source]
Stores the mapping: key is logical qubit id, value is mapped qubit id from 0,…,self.num_qubits
 cyclic
If chain is cyclic or not
 Type
Bool
 storage
Number of gate it caches before mapping.
 Type
int
 num_mappings
Number of times the mapper changed the mapping
 Type
int
 depth_of_swaps
Key are circuit depth of swaps, value is the number of such mappings which have been applied
 Type
dict
 num_of_swaps_per_mapping
Key are the number of swaps per mapping, value is the number of such mappings which have been applied
 Type
dict
Note
Gates are cached and only mapped from time to time. A FastForwarding gate doesn’t empty the cache, only a FlushGate does.
Only 1 and two qubit gates allowed.
Does not optimize for dirty qubits.
 __init__(num_qubits, cyclic=False, storage=1000)[source]
Initialize a LinearMapper compiler engine.
 Parameters
num_qubits (int) – Number of physical qubits in the linear chain
cyclic (bool) – If 1D chain is a cycle. Default is False.
storage (int) – Number of gates to temporarily store, default is 1000
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and, for each command, stores it until we do a mapping (FlushGate or Cache of stored commands is full).
 Parameters
command_list (list of Command objects) – list of commands to receive.
 static return_new_mapping(num_qubits, cyclic, currently_allocated_ids, stored_commands, current_mapping)[source]
Build a mapping of qubits to a linear chain.
It goes through stored_commands and tries to find a mapping to apply these gates on a first come first served basis. More compilicated scheme could try to optimize to apply as many gates as possible between the Swaps.
 Parameters
num_qubits (int) – Total number of qubits in the linear chain
cyclic (bool) – If linear chain is a cycle.
currently_allocated_ids (set of int) – Logical qubit ids for which the Allocate gate has already been processed and sent to the next engine but which are not yet deallocated and hence need to be included in the new mapping.
stored_commands (list of Command objects) – Future commands which should be applied next.
current_mapping – A current mapping as a dict. key is logical qubit id, value is placement id. If there are different possible maps, this current mapping is used to minimize the swaps to go to the new mapping by a heuristic.
 Returns: A new mapping as a dict. key is logical qubit id,
value is placement id
 class projectq.cengines.LocalOptimizer(cache_size=5, m=None)[source]
Circuit optimization compiler engine.
LocalOptimizer is a compiler engine which optimizes locally (merging rotations, cancelling gates with their inverse) in a local window of user defined size.
It stores all commands in a dict of lists, where each qubit has its own gate pipeline. After adding a gate, it tries to merge / cancel successive gates using the get_merged and get_inverse functions of the gate (if available). For examples, see BasicRotationGate. Once a list corresponding to a qubit contains >=m gates, the pipeline is sent on to the next engine.
 class projectq.cengines.MainEngine(backend=None, engine_list=None, verbose=False)[source]
The MainEngine class provides all functionality of the main compiler engine.
It initializes all further compiler engines (calls, e.g., .next_engine=…) and keeps track of measurement results and active qubits (and their IDs).
 next_engine
Next compiler engine (or the backend).
 Type
 main_engine
Self.
 Type
 active_qubits
WeakSet containing all active qubits
 Type
WeakSet
 dirty_qubits
Containing all dirty qubit ids
 Type
Set
 backend
Access the backend.
 Type
 mapper
Access to the mapper if there is one.
 Type
 n_engines
Current number of compiler engines in the engine list
 Type
int
 n_engines_max
Maximum number of compiler engines allowed in the engine list. Defaults to 100.
 Type
int
 __init__(backend=None, engine_list=None, verbose=False)[source]
Initialize the main compiler engine and all compiler engines.
Sets ‘next_engine’ and ‘main_engine’attributes of all compiler engines and adds the backend as the last engine.
 Parameters
backend (BasicEngine) – Backend to send the compiled circuit to.
engine_list (list<BasicEngine>) – List of engines / backends to use as compiler engines. Note: The engine list must not contain multiple mappers (instances of BasicMapperEngine). Default: projectq.setups.default.get_engine_list()
verbose (bool) – Either print full or compact error messages. Default: False (i.e. compact error messages).
Example
from projectq import MainEngine eng = MainEngine() # uses default engine_list and the Simulator
Instead of the default engine_list one can use, e.g., one of the IBM setups which defines a custom engine_list useful for one of the IBM chips
Example
import projectq.setups.ibm as ibm_setup from projectq import MainEngine eng = MainEngine(engine_list=ibm_setup.get_engine_list()) # eng uses the default Simulator backend
Alternatively, one can specify all compiler engines explicitly, e.g.,
Example
from projectq.cengines import ( TagRemover, AutoReplacer, LocalOptimizer, DecompositionRuleSet, ) from projectq.backends import Simulator from projectq import MainEngine rule_set = DecompositionRuleSet() engines = [AutoReplacer(rule_set), TagRemover(), LocalOptimizer(3)] eng = MainEngine(Simulator(), engines)
 flush(deallocate_qubits=False)[source]
Flush the entire circuit down the pipeline, clearing potential buffers (of, e.g., optimizers).
 Parameters
deallocate_qubits (bool) – If True, deallocates all qubits that are still alive (invalidating references to them by setting their id to 1).
 get_measurement_result(qubit)[source]
Return the classical value of a measured qubit, given that an engine registered this result previously.
See also setMeasurementResult.
 Parameters
qubit (BasicQubit) – Qubit of which to get the measurement result.
Example
from projectq.ops import H, Measure from projectq import MainEngine eng = MainEngine() qubit = eng.allocate_qubit() # quantum register of size 1 H  qubit Measure  qubit eng.get_measurement_result(qubit[0]) == int(qubit)
 get_new_qubit_id()[source]
Return a unique qubit id to be used for the next qubit allocation.
 Returns
New unique qubit id.
 Return type
new_qubit_id (int)
 receive(command_list)[source]
Forward the list of commands to the first engine.
 Parameters
command_list (list<Command>) – List of commands to receive (and then send on)
 send(command_list)[source]
Forward the list of commands to the next engine in the pipeline.
It also shortens exception stack traces if self.verbose is False.
 set_measurement_result(qubit, value)[source]
Register a measurement result.
The engine being responsible for measurement results needs to register these results with the master engine such that they are available when the user calls an int() or bool() conversion operator on a measured qubit.
 Parameters
qubit (BasicQubit) – Qubit for which to register the measurement result.
value (bool) – Boolean value of the measurement outcome (True / False = 1 / 0 respectively).
 class projectq.cengines.ManualMapper(map_fun=<function ManualMapper.<lambda>>)[source]
Manual Mapper which adds QubitPlacementTags to Allocate gate commands according to a userspecified mapping.
 map
The function which maps a given qubit id to its location. It gets set when initializing the mapper.
 Type
function
 __init__(map_fun=<function ManualMapper.<lambda>>)[source]
Initialize the mapper to a given mapping.
If no mapping function is provided, the qubit id is used as the location.
 Parameters
map_fun (function) – Function which, given the qubit id, returns an integer describing the physical location (must be constant).
 exception projectq.cengines.NotYetMeasuredError[source]
Exception raised when trying to access the measurement value of a qubit that has not yet been measured.
 class projectq.cengines.SwapAndCNOTFlipper(connectivity)[source]
Flip CNOTs and translates Swaps to CNOTs where necessary.
Warning
This engine assumes that CNOT and Hadamard gates are supported by the following engines.
Warning
This engine cannot be used as a backend.
 __init__(connectivity)[source]
Initialize the engine.
 Parameters
connectivity (set) – Set of tuples (c, t) where if (c, t) is an element of the set means that a CNOT can be performed between the physical ids (c, t) with c being the control and t being the target qubit.
 is_available(cmd)[source]
Check if the IBM backend can perform the Command cmd and return True if so.
 Parameters
cmd (Command) – The command to check
 receive(command_list)[source]
Receive a list of commands.
Receive a command list and if the command is a CNOT gate, it flips it using Hadamard gates if necessary; if it is a Swap gate, it decomposes it using 3 CNOTs. All other gates are simply sent to the next engine.
 Parameters
command_list (list of Command objects) – list of commands to receive.
 class projectq.cengines.TagRemover(tags=None)[source]
Compiler engine that remove temporary command tags.
TagRemover is a compiler engine which removes temporary command tags (see the tag classes such as LoopTag in projectq.meta._loop).
Removing tags is important (after having handled them if necessary) in order to enable optimizations across metafunction boundaries (compute/ action/uncompute or loops after unrolling)
 __init__(tags=None)[source]
Initialize a TagRemover object.
 Parameters
tags – A list of meta tag classes (e.g., [ComputeTag, UncomputeTag]) denoting the tags to remove
 receive(command_list)[source]
Receive a list of commands.
Receive a list of commands from the previous engine, remove all tags which are an instance of at least one of the meta tags provided in the constructor, and then send them on to the next compiler engine.
 Parameters
command_list (list<Command>) – List of commands to receive and then (after removing tags) send on.
libs
The library collection of ProjectQ which, for now, consists of a tiny math library and an interface library to RevKit. Soon, more libraries will be added.
Subpackages
libs.math
A tiny math library which will be extended thoughout the next weeks. Right now, it only contains the math functions necessary to run Beauregard’s implementation of Shor’s algorithm.
Module containing constant math quantum operations. 

Registers a few default replacement rules for Shor's algorithm to work (see Examples). 

Quantum number math gates for ProjectQ. 

Definition of some mathematical quantum operations. 

Add a constant to a quantum number represented by a quantum register, stored from low to highbit. 

Add a constant to a quantum number represented by a quantum register modulo N. 


Builtin mutable sequence. 
Multiply a quantum number represented by a quantum register by a constant modulo N. 

Subtract a constant from a quantum number represented by a quantum register, stored from low to highbit. 

Subtract a constant from a quantum number represented by a quantum register modulo N. 
Submodules
Module containing constant math quantum operations.
 projectq.libs.math._constantmath.add_constant(eng, constant, quint)[source]
Add a classical constant c to the quantum integer (qureg) quint using Draper addition.
Note
Uses the Fouriertransform adder from https://arxiv.org/abs/quantph/0008033.
 projectq.libs.math._constantmath.add_constant_modN(eng, constant, N, quint)[source]
Add a classical constant c to a quantum integer (qureg) quint modulo N using Draper addition.
This function uses Draper addition and the construction from https://arxiv.org/abs/quantph/0205095.
Registers a few default replacement rules for Shor’s algorithm to work (see Examples).
Quantum number math gates for ProjectQ.
 class projectq.libs.math._gates.AddConstant(a)[source]
Add a constant to a quantum number represented by a quantum register, stored from low to highbit.
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[1] # qunum is now equal to 2 AddConstant(3)  qunum # qunum is now equal to 5
Important: if you run with conditional and carry, carry needs to be a quantum register for the compiler/decomposition to work.
 class projectq.libs.math._gates.AddConstantModN(a, N)[source]
Add a constant to a quantum number represented by a quantum register modulo N.
The number is stored from low to highbit, i.e., qunum[0] is the LSB.
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[1] # qunum is now equal to 2 AddConstantModN(3, 4)  qunum # qunum is now equal to 1
Note
Preconditions:
c < N
c >= 0
The value stored in the quantum register must be lower than N
 class projectq.libs.math._gates.AddQuantumGate[source]
Add up two quantum numbers represented by quantum registers.
The numbers are stored from low to highbit, i.e., qunum[0] is the LSB.
Example
qunum_a = eng.allocate_qureg(5) # 5qubit number qunum_b = eng.allocate_qureg(5) # 5qubit number carry_bit = eng.allocate_qubit() X  qunum_a[2] # qunum_a is now equal to 4 X  qunum_b[3] # qunum_b is now equal to 8 AddQuantum  (qunum_a, qunum_b, carry) # qunum_a remains 4, qunum_b is now 12 and carry_bit is 0
 class projectq.libs.math._gates.ComparatorQuantumGate[source]
Flip a compare qubit if the binary value of first imput is higher than the second input.
The numbers are stored from low to highbit, i.e., qunum[0] is the LSB. .. rubric:: Example
qunum_a = eng.allocate_qureg(5) # 5qubit number qunum_b = eng.allocate_qureg(5) # 5qubit number compare_bit = eng.allocate_qubit() X  qunum_a[4] # qunum_a is now equal to 16 X  qunum_b[3] # qunum_b is now equal to 8 ComparatorQuantum  (qunum_a, qunum_b, compare_bit) # qunum_a and qunum_b remain 16 and 8, qunum_b is now 12 and compare bit is now 1
 class projectq.libs.math._gates.DivideQuantumGate[source]
Divide one quantum number from another.
Takes three inputs which should be quantum registers of equal size; a dividend, a remainder and a divisor. The remainder should be in the state 0…0> and the dividend should be bigger than the divisor.The gate returns (in this order): the remainder, the quotient and the divisor.
The numbers are stored from low to highbit, i.e., qunum[0] is the LSB.
Example: .. codeblock:: python
qunum_a = eng.allocate_qureg(5) # 5qubit number qunum_b = eng.allocate_qureg(5) # 5qubit number qunum_c = eng.allocate_qureg(5) # 5qubit number
All(X)  [qunum_a[0], qunum_a[3]] # qunum_a is now equal to 9 X  qunum_c[2] # qunum_c is now equal to 4
DivideQuantum  (qunum_a, qunum_b, qunum_c) # qunum_a is now equal to 1 (remainder), qunum_b is now # equal to 2 (quotient) and qunum_c remains 4 (divisor)
# dividend>remainder>divisor> > remainder>quotient>divisor>
 class projectq.libs.math._gates.MultiplyByConstantModN(a, N)[source]
Multiply a quantum number represented by a quantum register by a constant modulo N.
The number is stored from low to highbit, i.e., qunum[0] is the LSB.
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[2] # qunum is now equal to 4 MultiplyByConstantModN(3, 5)  qunum # qunum is now 2.
Note
Preconditions:
c < N
c >= 0
gcd(c, N) == 1
The value stored in the quantum register must be lower than N
 class projectq.libs.math._gates.MultiplyQuantumGate[source]
Multiply two quantum numbers represented by a quantum registers.
Requires three quantum registers as inputs, the first two are the numbers to be multiplied and should have the same size (n qubits). The third register will hold the product and should be of size 2n+1. The numbers are stored from low to highbit, i.e., qunum[0] is the LSB.
Example
qunum_a = eng.allocate_qureg(4) qunum_b = eng.allocate_qureg(4) qunum_c = eng.allocate_qureg(9) X  qunum_a[2] # qunum_a is now 4 X  qunum_b[3] # qunum_b is now 8 MultiplyQuantum()  (qunum_a, qunum_b, qunum_c) # qunum_a remains 4 and qunum_b remains 8, qunum_c is now equal to 32
 projectq.libs.math._gates.SubConstant(a)[source]
Subtract a constant from a quantum number represented by a quantum register, stored from low to highbit.
 Parameters
a (int) – Constant to subtract
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[2] # qunum is now equal to 4 SubConstant(3)  qunum # qunum is now equal to 1
 projectq.libs.math._gates.SubConstantModN(a, N)[source]
Subtract a constant from a quantum number represented by a quantum register modulo N.
The number is stored from low to highbit, i.e., qunum[0] is the LSB.
 Parameters
a (int) – Constant to add
N (int) – Constant modulo which the addition of a should be carried out.
Example
qunum = eng.allocate_qureg(3) # 3qubit number X  qunum[1] # qunum is now equal to 2 SubConstantModN(4, 5)  qunum # qunum is now 2 = 6 = 1 (mod 5)
Note
Preconditions:
c < N
c >= 0
The value stored in the quantum register must be lower than N
 class projectq.libs.math._gates.SubtractQuantumGate[source]
Subtract one quantum number from another quantum number both represented by quantum registers.
Example: .. codeblock:: python
qunum_a = eng.allocate_qureg(5) # 5qubit number qunum_b = eng.allocate_qureg(5) # 5qubit number X  qunum_a[2] # qunum_a is now equal to 4 X  qunum_b[3] # qunum_b is now equal to 8 SubtractQuantum  (qunum_a, qunum_b) # qunum_a remains 4, qunum_b is now 4
Definition of some mathematical quantum operations.
 projectq.libs.math._quantummath.add_quantum(eng, quint_a, quint_b, carry=None)[source]
Add two quantum integers.
i.e.,
a0…a(n1)>b(0)…b(n1)>c> > a0…a(n1)>b+a(0)…b+a(n)>
(only works if quint_a and quint_b are the same size and carry is a single qubit)
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
carry (list) – Carry qubit
Notes
Ancilla: 0, size: 7n6, toffoli: 2n1, depth: 5n3.
References
Quantum addition using ripple carry from: https://arxiv.org/pdf/0910.2530.pdf
 projectq.libs.math._quantummath.comparator(eng, quint_a, quint_b, comp)[source]
Compare the size of two quantum integers.
i.e,
if a>b: a>b>c> > a>b>c+1>
(only works if quint_a and quint_b are the same size and the comparator is 1 qubit)
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
comp (Qubit) – Comparator qubit
Notes
Comparator flipping a compare qubit by computing the high bit of ba, which is 1 if and only if a > b. The high bit is computed using the first half of circuit in AddQuantum (such that the high bit is written to the carry qubit) and then undoing the first half of the circuit. By complementing b at the start and b+a at the end the high bit of ba is calculated.
Ancilla: 0, size: 8n3, toffoli: 2n+1, depth: 4n+3.
 projectq.libs.math._quantummath.inverse_add_quantum_carry(eng, quint_a, quint_b)[source]
Inverse of quantum addition with carry.
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
 projectq.libs.math._quantummath.inverse_quantum_division(eng, remainder, quotient, divisor)[source]
Perform the inverse of a restoring integer division.
i.e.,
remainder>quotient>divisor> > dividend>remainder(0)>divisor>
 Parameters
eng (MainEngine) – ProjectQ MainEngine
dividend (list) – Quantum register (or list of qubits)
remainder (list) – Quantum register (or list of qubits)
divisor (list) – Quantum register (or list of qubits)
 projectq.libs.math._quantummath.inverse_quantum_multiplication(eng, quint_a, quint_b, product)[source]
Inverse of the multiplication of two quantum integers.
i.e,
(only works if quint_a and quint_b are of the same size, n qubits and product has size 2n+1)
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
product (list) – Quantum register (or list of qubits) storing the result
 projectq.libs.math._quantummath.quantum_conditional_add(eng, quint_a, quint_b, conditional)[source]
Add up two quantum integers if conditional is high.
i.e.,
a>b>c> > a>b+a>c> (without a carry out qubit)
if conditional is low, no operation is performed, i.e., a>b>c> > a>b>c>
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
conditional (list) – Conditional qubit
Notes
Ancilla: 0, Size: 7n7, Toffoli: 3n3, Depth: 5n3.
References
Quantum Conditional Add from https://arxiv.org/pdf/1609.01241.pdf
 projectq.libs.math._quantummath.quantum_conditional_add_carry(eng, quint_a, quint_b, ctrl, z)[source]
Add up two quantum integers if the control qubit is 1>.
i.e.,
a>b>ctrl>z(0)z(1)> > a>s(0)…s(n1)>ctrl>s(n)z(1)> (where s denotes the sum of a and b)
If the control qubit is 0> no operation is performed:
a>b>ctrl>z(0)z(1)> > a>b>ctrl>z(0)z(1)>
(only works if quint_a and quint_b are of the same size, ctrl is a single qubit and z is a quantum register with 2 qubits.
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
ctrl (list) – Control qubit
z (list) – Quantum register with 2 qubits
Notes
Ancilla: 2, size: 7n  4, toffoli: 3n + 2, depth: 5n.
References
Quantum conditional add with no input carry from: https://arxiv.org/pdf/1706.05113.pdf
 projectq.libs.math._quantummath.quantum_division(eng, dividend, remainder, divisor)[source]
Perform restoring integer division.
i.e.,
dividend>remainder>divisor> > remainder>quotient>divisor>
(only works if all three qubits are of equal length)
 Parameters
eng (MainEngine) – ProjectQ MainEngine
dividend (list) – Quantum register (or list of qubits)
remainder (list) – Quantum register (or list of qubits)
divisor (list) – Quantum register (or list of qubits)
Notes
Ancilla: n, size 16n^2  13, toffoli: 5n^2 5 , depth: 10n^26.
References
Quantum Restoring Integer Division from: https://arxiv.org/pdf/1609.01241.pdf.
 projectq.libs.math._quantummath.quantum_multiplication(eng, quint_a, quint_b, product)[source]
Multiplies two quantum integers.
i.e,
(only works if quint_a and quint_b are of the same size, n qubits and product has size 2n+1).
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
product (list) – Quantum register (or list of qubits) storing the result
Notes
Ancilla: 2n + 1, size: 7n^2  9n + 4, toffoli: 5n^2  4n, depth: 3n^2  2.
References
Quantum multiplication from: https://arxiv.org/abs/1706.05113.
 projectq.libs.math._quantummath.subtract_quantum(eng, quint_a, quint_b)[source]
Subtract two quantum integers.
i.e.,
(only works if quint_a and quint_b are the same size)
 Parameters
eng (MainEngine) – ProjectQ MainEngine
quint_a (list) – Quantum register (or list of qubits)
quint_b (list) – Quantum register (or list of qubits)
Notes
Quantum subtraction using bitwise complementation of quantum adder: ba = (a + b’)’. Same as the quantum addition circuit except that the steps involving the carry qubit are left out and complement b at the start and at the end of the circuit is added.
Ancilla: 0, size: 9n8, toffoli: 2n2, depth: 5n5.
References
Quantum addition using ripple carry from: https://arxiv.org/pdf/0910.2530.pdf
Module contents
 class projectq.libs.math.AddConstant(a)[source]
Add a constant to a quantum number represented by a quantum register, stored from low to highbit.
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[1] # qunum is now equal to 2 AddConstant(3)  qunum # qunum is now equal to 5
Important: if you run with conditional and carry, carry needs to be a quantum register for the compiler/decomposition to work.
 class projectq.libs.math.AddConstantModN(a, N)[source]
Add a constant to a quantum number represented by a quantum register modulo N.
The number is stored from low to highbit, i.e., qunum[0] is the LSB.
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[1] # qunum is now equal to 2 AddConstantModN(3, 4)  qunum # qunum is now equal to 1
Note
Preconditions:
c < N
c >= 0
The value stored in the quantum register must be lower than N
 __init__(a, N)[source]
Initialize the gate to the number to add modulo N.
 Parameters
a (int) – Number to add to a quantum register (0 <= a < N).
N (int) – Number modulo which the addition is carried out.
It also initializes its base class, BasicMathGate, with the corresponding function, so it can be emulated efficiently.
 class projectq.libs.math.MultiplyByConstantModN(a, N)[source]
Multiply a quantum number represented by a quantum register by a constant modulo N.
The number is stored from low to highbit, i.e., qunum[0] is the LSB.
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[2] # qunum is now equal to 4 MultiplyByConstantModN(3, 5)  qunum # qunum is now 2.
Note
Preconditions:
c < N
c >= 0
gcd(c, N) == 1
The value stored in the quantum register must be lower than N
 __init__(a, N)[source]
Initialize the gate to the number to multiply with modulo N.
 Parameters
a (int) – Number by which to multiply a quantum register (0 <= a < N).
N (int) – Number modulo which the multiplication is carried out.
It also initializes its base class, BasicMathGate, with the corresponding function, so it can be emulated efficiently.
 projectq.libs.math.SubConstant(a)[source]
Subtract a constant from a quantum number represented by a quantum register, stored from low to highbit.
 Parameters
a (int) – Constant to subtract
Example
qunum = eng.allocate_qureg(5) # 5qubit number X  qunum[2] # qunum is now equal to 4 SubConstant(3)  qunum # qunum is now equal to 1
 projectq.libs.math.SubConstantModN(a, N)[source]
Subtract a constant from a quantum number represented by a quantum register modulo N.
The number is stored from low to highbit, i.e., qunum[0] is the LSB.
 Parameters
a (int) – Constant to add
N (int) – Constant modulo which the addition of a should be carried out.
Example
qunum = eng.allocate_qureg(3) # 3qubit number X  qunum[1] # qunum is now equal to 2 SubConstantModN(4, 5)  qunum # qunum is now 2 = 6 = 1 (mod 5)
Note
Preconditions:
c < N
c >= 0
The value stored in the quantum register must be lower than N
libs.revkit
This library integrates RevKit into ProjectQ to allow some automatic synthesis routines for reversible logic. The library adds the following operations that can be used to construct quantum circuits:
ControlFunctionOracle
: Synthesizes a reversible circuit from Boolean control functionPermutationOracle
: Synthesizes a reversible circuit for a permutationPhaseOracle
: Synthesizes phase circuit from an arbitrary Boolean function
RevKit can be installed from PyPi with pip install revkit.
Note
The RevKit Python module must be installed in order to use this ProjectQ library.
There exist precompiled binaries in PyPi, as well as a source distribution. Note that a C++ compiler with C++17 support is required to build the RevKit python module from source. Examples for compatible compilers are Clang 6.0, GCC 7.3, and GCC 8.1.
The integration of RevKit into ProjectQ and other quantum programming languages is described in the paper
Mathias Soeken, Thomas Haener, and Martin Roetteler “Programming Quantum Computers Using Design Automation,” in: Design Automation and Test in Europe (2018) [arXiv:1803.01022]
RevKit support for control function oracles. 

RevKit support for permutation oracles. 

RevKit support for phase oracles. 

Module containing some utility functions 

Synthesize a negation controlled by an arbitrary control function. 

Synthesize a permutation using RevKit. 


Synthesize phase circuit from an arbitrary Boolean function. 
Submodules
RevKit support for control function oracles.
 class projectq.libs.revkit._control_function.ControlFunctionOracle(function, **kwargs)[source]
Synthesize a negation controlled by an arbitrary control function.
This creates a circuit for a NOT gate which is controlled by an arbitrary Boolean control function. The control function is provided as integer representation of the function’s truth table in binary notation. For example, for the majorityofthree function, which truth table 11101000, the value for function can be, e.g.,
0b11101000
,0xe8
, or232
.Example
This example creates a circuit that causes to invert qubit
d
, the majorityofthree function evaluates to true for the control qubitsa
,b
, andc
.ControlFunctionOracle(0x8E)  ([a, b, c], d)
RevKit support for permutation oracles.
 class projectq.libs.revkit._permutation.PermutationOracle(permutation, **kwargs)[source]
Synthesize a permutation using RevKit.
Given a permutation over 2**q elements (starting from 0), this class helps to automatically find a reversible circuit over q qubits that realizes that permutation.
Example
PermutationOracle([0, 2, 1, 3])  (a, b)
RevKit support for phase oracles.
 class projectq.libs.revkit._phase.PhaseOracle(function, **kwargs)[source]
Synthesize phase circuit from an arbitrary Boolean function.
This creates a phase circuit from a Boolean function. It inverts the phase of all amplitudes for which the function evaluates to 1. The Boolean function is provided as integer representation of the function’s truth table in binary notation. For example, for the majorityofthree function, which truth table 11101000, the value for function can be, e.g.,
0b11101000
,0xe8
, or232
.Note that a phase circuit can only accurately be found for a normal function, i.e., a function that maps the input pattern 0, 0, …, 0 to 0. The circuits for a function and its inverse are the same.
Example
This example creates a phase circuit based on the majorityofthree function on qubits
a
,b
, andc
.PhaseOracle(0x8E)  (a, b, c)
Module containing some utility functions
Module contents
Module containing code to interface with RevKit
 class projectq.libs.revkit.ControlFunctionOracle(function, **kwargs)[source]
Synthesize a negation controlled by an arbitrary control function.
This creates a circuit for a NOT gate which is controlled by an arbitrary Boolean control function. The control function is provided as integer representation of the function’s truth table in binary notation. For example, for the majorityofthree function, which truth table 11101000, the value for function can be, e.g.,
0b11101000
,0xe8
, or232
.Example
This example creates a circuit that causes to invert qubit
d
, the majorityofthree function evaluates to true for the control qubitsa
,b
, andc
.ControlFunctionOracle(0x8E)  ([a, b, c], d)
 __init__(function, **kwargs)[source]
Initialize a control function oracle.
 Parameters
function (int) – Function truth table.
 Keyword Arguments
synth – A RevKit synthesis command which creates a reversible circuit based on a truth table and requires no additional ancillae (e.g.,
revkit.esopbs
). Can also be a nullary lambda that calls several RevKit commands. Default:revkit.esopbs
 class projectq.libs.revkit.PermutationOracle(permutation, **kwargs)[source]
Synthesize a permutation using RevKit.
Given a permutation over 2**q elements (starting from 0), this class helps to automatically find a reversible circuit over q qubits that realizes that permutation.
Example
PermutationOracle([0, 2, 1, 3])  (a, b)
 __init__(permutation, **kwargs)[source]
Initialize a permutation oracle.
 Parameters
permutation (list<int>) – Permutation (starting from 0).
 Keyword Arguments
synth – A RevKit synthesis command which creates a reversible circuit based on a reversible truth table (e.g.,
revkit.tbs
orrevkit.dbs
). Can also be a nullary lambda that calls several RevKit commands. Default:revkit.tbs
 class projectq.libs.revkit.PhaseOracle(function, **kwargs)[source]
Synthesize phase circuit from an arbitrary Boolean function.
This creates a phase circuit from a Boolean function. It inverts the phase of all amplitudes for which the function evaluates to 1. The Boolean function is provided as integer representation of the function’s truth table in binary notation. For example, for the majorityofthree function, which truth table 11101000, the value for function can be, e.g.,
0b11101000
,0xe8
, or232
.Note that a phase circuit can only accurately be found for a normal function, i.e., a function that maps the input pattern 0, 0, …, 0 to 0. The circuits for a function and its inverse are the same.
Example
This example creates a phase circuit based on the majorityofthree function on qubits
a
,b
, andc
.PhaseOracle(0x8E)  (a, b, c)
 __init__(function, **kwargs)[source]
Initialize a phase oracle.
 Parameters
function (int) – Function truth table.
 Keyword Arguments
synth – A RevKit synthesis command which creates a reversible circuit based on a truth table and requires no additional ancillae (e.g.,
revkit.esopps
). Can also be a nullary lambda that calls several RevKit commands. Default:revkit.esopps
Submodules
contains a function to plot measurement outcome probabilities as a histogram for the simulator 
hist
contains a function to plot measurement outcome probabilities as a histogram for the simulator
Module contents
ProjectQ module containing libraries expanding the basic functionalities of ProjectQ
meta
Contains meta statements which allow more optimal code while making it easier for users to write their code. Examples are with Compute, followed by an automatic uncompute or with Control, which allows the user to condition an entire code block upon the state of a qubit.
Definition of Compute, Uncompute and CustomUncompute. 

Contains the tools to make an entire section of operations controlled. 

Tools to easily invert a sequence of gates. 

Define the DirtyQubitTag meta tag. 

Exception classes for projectq.meta. 

Definition of LogicalQubitIDTag to annotate a MeasureGate for mapped qubits. 

Tools to implement loops. 

Tools to add/remove compiler engines to the MainEngine list. 

Return canonical form for control state. 


Start a computesection. 
Compute meta tag. 


Condition an entire code block on the value of qubits being 1. 

Start a custom uncomputesection. 

Invert an entire code block. 
Dirty qubit meta tag. 


Remove an engine from the singlylinked list of engines. 
Return the number of control qubits of the command object cmd. 

Return whether a command has negatively controlled qubits. 


Insert an engine into the singlylinked list of engines. 

LogicalQubitIDTag for a mapped qubit to annotate a MeasureGate. 

Loop n times over an entire code block. 
Loop meta tag. 


Uncompute automatically. 
Uncompute meta tag. 
Submodules
_compute
Definition of Compute, Uncompute and CustomUncompute.
Contains Compute, Uncompute, and CustomUncompute classes which can be used to annotate Compute / Action / Uncompute sections, facilitating the conditioning of the entire operation on the value of a qubit / register (only Action needs controls). This file also defines the corresponding meta tags.
 class projectq.meta._compute.Compute(engine)[source]
Start a computesection.
Example
with Compute(eng): do_something(qubits) action(qubits) Uncompute(eng) # runs inverse of the compute section
Warning
If qubits are allocated within the compute section, they must either be uncomputed and deallocated within that section or, alternatively, uncomputed and deallocated in the following uncompute section.
This means that the following examples are valid:
with Compute(eng): anc = eng.allocate_qubit() do_something_with_ancilla(anc) ... uncompute_ancilla(anc) del anc do_something_else(qubits) Uncompute(eng) # will allocate a new ancilla (with a different id) # and then deallocate it again
with Compute(eng): anc = eng.allocate_qubit() do_something_with_ancilla(anc) ... do_something_else(qubits) Uncompute(eng) # will deallocate the ancilla!
After the uncompute section, ancilla qubits allocated within the compute section will be invalid (and deallocated). The same holds when using CustomUncompute.
Failure to comply with these rules results in an exception being thrown.
 class projectq.meta._compute.ComputeEngine[source]
Add Computetags to all commands and stores them (to later uncompute them automatically).
 end_compute()[source]
End the compute step (exit the with Compute()  statement).
Will tell the Computeengine to stop caching. It then waits for the uncompute instruction, which is when it sends all cached commands inverted and in reverse order down to the next compiler engine.
 Raises
QubitManagementError – If qubit has been deallocated in Compute section which has not been allocated in Compute section
 receive(command_list)[source]
Receive a list of commands.
If in computemode, receive commands and store deepcopy of each cmd. Add ComputeTag to received cmd and send it on. Otherwise, send all received commands directly to next_engine.
 Parameters
command_list (list<Command>) – List of commands to receive.
 run_uncompute()[source]
Send uncomputing gates.
Sends the inverse of the stored commands in reverse order down to the next engine. And also deals with allocated qubits in Compute section. If a qubit has been allocated during compute, it will be deallocated during uncompute. If a qubit has been allocated and deallocated during compute, then a new qubit is allocated and deallocated during uncompute.
 class projectq.meta._compute.CustomUncompute(engine)[source]
Start a custom uncomputesection.
Example
with Compute(eng): do_something(qubits) action(qubits) with CustomUncompute(eng): do_something_inverse(qubits)
 Raises
QubitManagementError – If qubits are allocated within Compute or within CustomUncompute context but are not deallocated.
 exception projectq.meta._compute.NoComputeSectionError[source]
Exception raised if uncompute is called but no compute section found.
 projectq.meta._compute.Uncompute(engine)[source]
Uncompute automatically.
Example
with Compute(eng): do_something(qubits) action(qubits) Uncompute(eng) # runs inverse of the compute section
_control
Contains the tools to make an entire section of operations controlled.
Example
with Control(eng, qubit1):
H  qubit2
X  qubit3
 class projectq.meta._control.Control(engine, qubits, ctrl_state=CtrlAll.One)[source]
Condition an entire code block on the value of qubits being 1.
Example
with Control(eng, ctrlqubits): do_something(otherqubits)
 class projectq.meta._control.ControlEngine(qubits, ctrl_state=CtrlAll.One)[source]
Add control qubits to all commands that have no compute / uncompute tags.
 projectq.meta._control.canonical_ctrl_state(ctrl_state, num_qubits)[source]
Return canonical form for control state.
 Parameters
ctrl_state (int,str,CtrlAll) – Initial control state representation
num_qubits (int) – number of control qubits
 Returns
Canonical form of control state (currently a string composed of ‘0’ and ‘1’)
Note
In case of integer values for ctrl_state, the least significant bit applies to the first qubit in the qubit register, e.g. if ctrl_state == 2, its binary representation if ‘10’ with the least significan bit being 0.
This means in particular that the followings are equivalent:
canonical_ctrl_state(6, 3) == canonical_ctrl_state(6, '110')
_dagger
Tools to easily invert a sequence of gates.
with Dagger(eng):
H  qubit1
Rz(0.5)  qubit2
 class projectq.meta._dagger.Dagger(engine)[source]
Invert an entire code block.
Use it with a withstatement, i.e.,
with Dagger(eng): # [code to invert] pass
Warning
If the code to invert contains allocation of qubits, those qubits have to be deleted prior to exiting the ‘with Dagger()’ context.
This code is NOT VALID:
with Dagger(eng): qb = eng.allocate_qubit() H  qb # qb is still available!!!
The correct way of handling qubit (de)allocation is as follows:
with Dagger(eng): qb = eng.allocate_qubit() ... del qb # sends deallocate gate (which becomes an allocate)
_dirtyqubit
Define the DirtyQubitTag meta tag.
_exceptions
Exception classes for projectq.meta.
_logicalqubit
Definition of LogicalQubitIDTag to annotate a MeasureGate for mapped qubits.
_loop
Tools to implement loops.
Example
with Loop(eng, 4):
H  qb
Rz(M_PI / 3.0)  qb
 class projectq.meta._loop.Loop(engine, num)[source]
Loop n times over an entire code block.
Example
with Loop(eng, 4): # [quantum gates to be executed 4 times] pass
Warning
If the code in the loop contains allocation of qubits, those qubits have to be deleted prior to exiting the ‘with Loop()’ context.
This code is NOT VALID:
with Loop(eng, 4): qb = eng.allocate_qubit() H  qb # qb is still available!!!
The correct way of handling qubit (de)allocation is as follows:
with Loop(eng, 4): qb = eng.allocate_qubit() # ... del qb # sends deallocate gate
 class projectq.meta._loop.LoopEngine(num)[source]
A compiler engine to represent executing part of the code multiple times.
Stores all commands and, when done, executes them num times if no loop tag handler engine is available. If there is one, it adds a loop_tag to the commands and sends them on.
 receive(command_list)[source]
Receive (and potentially temporarily store) all commands.
Add LoopTag to all receiving commands and send to the next engine if a further engine is a LoopTaghandling engine. Otherwise store all commands (to later unroll them). Check that within the loop body, all allocated qubits have also been deallocated. If loop needs to be unrolled and ancilla qubits have been allocated within the loop body, then store a reference all these qubit ids (to change them when unrolling the loop)
 Parameters
command_list (list<Command>) – List of commands to store and later unroll or, if there is a LoopTaghandling engine, add the LoopTag.
_util
Tools to add/remove compiler engines to the MainEngine list.
 projectq.meta._util.drop_engine_after(prev_engine)[source]
Remove an engine from the singlylinked list of engines.
 Parameters
prev_engine (projectq.cengines.BasicEngine) – The engine just before the engine to drop.
 Returns
The dropped engine.
 Return type
Engine
 projectq.meta._util.insert_engine(prev_engine, engine_to_insert)[source]
Insert an engine into the singlylinked list of engines.
It also sets the correct main_engine for engine_to_insert.
 Parameters
prev_engine (projectq.cengines.BasicEngine) – The engine just before the insertion point.
engine_to_insert (projectq.cengines.BasicEngine) – The engine to insert at the insertion point.
Module contents
The projectq.meta package features meta instructions which help both the user and the compiler in writing/producing efficient code. It includes, e.g.,
Loop (with Loop(eng): …)
Compute/Uncompute (with Compute(eng): …, […], Uncompute(eng))
Control (with Control(eng, ctrl_qubits): …)
Dagger (with Dagger(eng): …)
 class projectq.meta.Compute(engine)[source]
Start a computesection.
Example
with Compute(eng): do_something(qubits) action(qubits) Uncompute(eng) # runs inverse of the compute section
Warning
If qubits are allocated within the compute section, they must either be uncomputed and deallocated within that section or, alternatively, uncomputed and deallocated in the following uncompute section.
This means that the following examples are valid:
with Compute(eng): anc = eng.allocate_qubit() do_something_with_ancilla(anc) ... uncompute_ancilla(anc) del anc do_something_else(qubits) Uncompute(eng) # will allocate a new ancilla (with a different id) # and then deallocate it again
with Compute(eng): anc = eng.allocate_qubit() do_something_with_ancilla(anc) ... do_something_else(qubits) Uncompute(eng) # will deallocate the ancilla!
After the uncompute section, ancilla qubits allocated within the compute section will be invalid (and deallocated). The same holds when using CustomUncompute.
Failure to comply with these rules results in an exception being thrown.
 __init__(engine)[source]
Initialize a Compute context.
 Parameters
engine (BasicEngine) – Engine which is the first to receive all commands (normally: MainEngine).
 class projectq.meta.Control(engine, qubits, ctrl_state=CtrlAll.One)[source]
Condition an entire code block on the value of qubits being 1.
Example
with Control(eng, ctrlqubits): do_something(otherqubits)
 class projectq.meta.CustomUncompute(engine)[source]
Start a custom uncomputesection.
Example
with Compute(eng): do_something(qubits) action(qubits) with CustomUncompute(eng): do_something_inverse(qubits)
 Raises
QubitManagementError – If qubits are allocated within Compute or within CustomUncompute context but are not deallocated.
 __init__(engine)[source]
Initialize a CustomUncompute context.
 Parameters
engine (BasicEngine) – Engine which is the first to receive all commands (normally: MainEngine).
 class projectq.meta.Dagger(engine)[source]
Invert an entire code block.
Use it with a withstatement, i.e.,
with Dagger(eng): # [code to invert] pass
Warning
If the code to invert contains allocation of qubits, those qubits have to be deleted prior to exiting the ‘with Dagger()’ context.
This code is NOT VALID:
with Dagger(eng): qb = eng.allocate_qubit() H  qb # qb is still available!!!
The correct way of handling qubit (de)allocation is as follows:
with Dagger(eng): qb = eng.allocate_qubit() ... del qb # sends deallocate gate (which becomes an allocate)
 class projectq.meta.LogicalQubitIDTag(logical_qubit_id)[source]
LogicalQubitIDTag for a mapped qubit to annotate a MeasureGate.
 logical_qubit_id
Logical qubit id
 Type
int
 class projectq.meta.Loop(engine, num)[source]
Loop n times over an entire code block.
Example
with Loop(eng, 4): # [quantum gates to be executed 4 times] pass
Warning
If the code in the loop contains allocation of qubits, those qubits have to be deleted prior to exiting the ‘with Loop()’ context.
This code is NOT VALID:
with Loop(eng, 4): qb = eng.allocate_qubit() H  qb # qb is still available!!!
The correct way of handling qubit (de)allocation is as follows:
with Loop(eng, 4): qb = eng.allocate_qubit() # ... del qb # sends deallocate gate
 __init__(engine, num)[source]
Enter a looped section.
 Parameters
engine – Engine handling the commands (usually MainEngine)
num (int) – Number of loop iterations
Example
with Loop(eng, 4): H  qb Rz(M_PI / 3.0)  qb
 Raises
TypeError – If number of iterations (num) is not an integer
ValueError – If number of iterations (num) is not >= 0
 projectq.meta.Uncompute(engine)[source]
Uncompute automatically.
Example
with Compute(eng): do_something(qubits) action(qubits) Uncompute(eng) # runs inverse of the compute section
 projectq.meta.canonical_ctrl_state(ctrl_state, num_qubits)[source]
Return canonical form for control state.
 Parameters
ctrl_state (int,str,CtrlAll) – Initial control state representation
num_qubits (int) – number of control qubits
 Returns
Canonical form of control state (currently a string composed of ‘0’ and ‘1’)
Note
In case of integer values for ctrl_state, the least significant bit applies to the first qubit in the qubit register, e.g. if ctrl_state == 2, its binary representation if ‘10’ with the least significan bit being 0.
This means in particular that the followings are equivalent:
canonical_ctrl_state(6, 3) == canonical_ctrl_state(6, '110')
 projectq.meta.drop_engine_after(prev_engine)[source]
Remove an engine from the singlylinked list of engines.
 Parameters
prev_engine (projectq.cengines.BasicEngine) – The engine just before the engine to drop.
 Returns
The dropped engine.
 Return type
Engine
 projectq.meta.get_control_count(cmd)[source]
Return the number of control qubits of the command object cmd.
 projectq.meta.has_negative_control(cmd)[source]
Return whether a command has negatively controlled qubits.
 projectq.meta.insert_engine(prev_engine, engine_to_insert)[source]
Insert an engine into the singlylinked list of engines.
It also sets the correct main_engine for engine_to_insert.
 Parameters
prev_engine (projectq.cengines.BasicEngine) – The engine just before the insertion point.
engine_to_insert (projectq.cengines.BasicEngine) – The engine to insert at the insertion point.
ops
The operations collection consists of various default gates and is a workinprogress, as users start to work with ProjectQ.
Definitions of some of the most basic quantum gates. 

The apply_command function and the Command class. 

Definition of the basic set of quantum gates. 

Definition of some meta gates. 

Definition of the quantum amplitude amplification gate. 

Definition of the QFT gate. 

Definition of the quantum phase estimation gate. 

QubitOperator stores a sum of Pauli operators acting on qubits. 

A few shortcuts for certain gates. 

Definition of the state preparation gate. 

Definition of the time evolution gate. 

Definition of uniformly controlled Ry and Rzrotation gates. 

alias of 

Dirty qubit allocation gate class. 

Qubit allocation gate class. 

Apply a command. 

Barrier gate class. 

Base class of all gates. 


Base class for all math gates. 

Base class for all phase gates. 
Base class of for all rotation gates. 


Return ncontrolled version of the provided gate. 
Classical instruction gate. 


Class used as a container to store commands. 

Controlled version of a gate. 

Shortcut for C(Rz(angle), n_qubits=1). 

Enum type to initialise the control state of qubits. 
Wrapper class allowing to execute the inverse of a gate, even when it does not define one. 

Qubit deallocation gate class. 

Entangle gate class. 

Base class for fastforward gates. 


Gate for flipping qubits by means of XGates. 
Flush gate (denotes the end of the circuit). 


Return the inverse of a gate. 
Hadamard gate class. 

Exception thrown when trying to set two incompatible states for a control qubit. 


Return True if the gate is an identity gate. 

A gate class whose instances are defined by a matrix. 
Measurement gate class (for single qubits). 

Exception thrown when trying to invert a gate which is not invertable. 

Exception thrown when trying to merge two gates which are not mergeable. 


Phase gate (global phase). 

Quantum Aplitude Amplification gate. 
Quantum Fourier Transform gate. 


Quantum Phase Estimation gate. 

A sum of terms acting on qubits, e.g., 0.5 * 'X0 X5' + 0.3 * 'Z1 Z2'. 

Phaseshift gate (equivalent to Rz up to a global phase). 

RotationX gate class. 

RotationXX gate class. 

RotationY gate class. 

RotationYY gate class. 

RotationZ gate class. 

RotationZZ gate class. 
Selfinverse basic gate class. 

S gate class. 

Squareroot Swap gate class. 

Squareroot X gate class. 


Gate for transforming qubits in state 0> to any desired quantum state. 
Swap gate class (swaps 2 qubits). 


Wrapper class allowing to apply a (singlequbit) gate to every qubit in a quantum register. 
T gate class. 


Gate for time evolution under a Hamiltonian (QubitOperator object). 
Uniformly controlled Ry gate as introduced in arXiv:quantph/0312218. 

Uniformly controlled Rz gate as introduced in arXiv:quantph/0312218. 

PauliX gate class. 

PauliY gate class. 

PauliZ gate class. 
Submodules
_basics
Definitions of some of the most basic quantum gates.
Defines the BasicGate class, the base class of all gates, the BasicRotationGate class, the SelfInverseGate, the FastForwardingGate, the ClassicalInstruction gate, and the BasicMathGate class.
Gates overload the  operator to allow the following syntax:
Gate  (qureg1, qureg2, qureg2)
Gate  (qureg, qubit)
Gate  qureg
Gate  qubit
Gate  (qubit,)
This means that for more than one quantum argument (right side of  ), a tuple needs to be made explicitely, while for one argument it is optional.
 class projectq.ops._basics.BasicGate[source]
Base class of all gates. (Don’t use it directly but derive from it).
 generate_command(qubits)[source]
Generate a command.
The command object created consists of the gate and the qubits being acted upon.
 Parameters
qubits – see BasicGate.make_tuple_of_qureg(qubits)
 Returns
A Command object containing the gate and the qubits.
 get_inverse()[source]
Return the inverse gate.
Standard implementation of get_inverse:
 Raises
NotInvertible – inverse is not implemented
 get_merged(other)[source]
Return this gate merged with another gate.
Standard implementation of get_merged:
 Raises
NotMergeable – merging is not implemented
 is_identity()[source]
Return True if the gate is an identity gate. In this base class, always returns False.
 static make_tuple_of_qureg(qubits)[source]
Convert quantum input of “gate  quantum input” to internal formatting.
A Command object only accepts tuples of Quregs (list of Qubit objects) as qubits input parameter. However, with this function we allow the user to use a more flexible syntax:
Gate  qubit
Gate  [qubit0, qubit1]
Gate  qureg
Gate  (qubit, )
Gate  (qureg, qubit)
where qubit is a Qubit object and qureg is a Qureg object. This function takes the right hand side of  and transforms it to the correct input parameter of a Command object which is:
> Gate  ([qubit], )
> Gate  ([qubit0, qubit1], )
> Gate  (qureg, )
> Gate  ([qubit], )
> Gate  (qureg, [qubit])
 Parameters
qubits – a Qubit object, a list of Qubit objects, a Qureg object, or a tuple of Qubit or Qureg objects (can be mixed).
 Returns
A tuple containing Qureg (or list of Qubits) objects.
 Return type
Canonical representation (tuple<qureg>)
 class projectq.ops._basics.BasicMathGate(math_fun)[source]
Base class for all math gates.
It allows efficient emulation by providing a mathematical representation which is given by the concrete gate which derives from this base class. The AddConstant gate, for example, registers a function of the form
def add(x): return (x + a,)
upon initialization. More generally, the function takes integers as parameters and returns a tuple / list of outputs, each entry corresponding to the function input. As an example, consider outofplace multiplication, which takes two input registers and adds the result into a third, i.e., (a,b,c) > (a,b,c+a*b). The corresponding function then is
def multiply(a, b, c): return (a, b, c + a * b)
 get_math_function(qubits)[source]
Get the math function associated with a BasicMathGate.
Return the math function which corresponds to the action of this math gate, given the input to the gate (a tuple of quantum registers).
 Parameters
qubits (tuple<Qureg>) – Qubits to which the math gate is being applied.
 Returns
Python function describing the action of this gate. (See BasicMathGate.__init__ for an example).
 Return type
math_fun (function)
 class projectq.ops._basics.BasicPhaseGate(angle)[source]
Base class for all phase gates.
A phase gate has a continuous parameter (the angle), labeled ‘angle’ / self.angle. Its inverse is the same gate with the negated argument. Phase gates of the same class can be merged by adding the angles. The continuous parameter is modulo 2 * pi, self.angle is in the interval [0, 2 * pi).
 get_inverse()[source]
Return the inverse of this rotation gate (negate the angle, return new object).
 get_merged(other)[source]
Return self merged with another gate.
Default implementation handles rotation gate of the same type, where angles are simply added.
 Parameters
other – Rotation gate of same type.
 Raises
NotMergeable – For nonrotation gates or rotation gates of different type.
 Returns
New object representing the merged gates.
 class projectq.ops._basics.BasicRotationGate(angle)[source]
Base class of for all rotation gates.
A rotation gate has a continuous parameter (the angle), labeled ‘angle’ / self.angle. Its inverse is the same gate with the negated argument. Rotation gates of the same class can be merged by adding the angles. The continuous parameter is modulo 4 * pi, self.angle is in the interval [0, 4 * pi).
 get_inverse()[source]
Return the inverse of this rotation gate (negate the angle, return new object).
 get_merged(other)[source]
Return self merged with another gate.
Default implementation handles rotation gate of the same type, where angles are simply added.
 Parameters
other – Rotation gate of same type.
 Raises
NotMergeable – For nonrotation gates or rotation gates of different type.
 Returns
New object representing the merged gates.
 class projectq.ops._basics.ClassicalInstructionGate[source]
Classical instruction gate.
Base class for all gates which are not quantum gates in the typical sense, e.g., measurement, allocation/deallocation, …
 class projectq.ops._basics.FastForwardingGate[source]
Base class for fastforward gates.
Base class for classical instruction gates which require a fastforward through compiler engines that cache / buffer gates. Examples include Measure and Deallocate, which both should be executed asap, such that Measurement results are available and resources are freed, respectively.
Note
The only requirement is that FlushGate commands run the entire circuit. FastForwardingGate objects can be used but the user cannot expect a measurement result to be available for all backends when calling only Measure. E.g., for the IBM Quantum Experience backend, sending the circuit for each Measuregate would be too inefficient, which is why a final
is required before the circuit gets sent through the API.
 class projectq.ops._basics.MatrixGate(matrix=None)[source]
A gate class whose instances are defined by a matrix.
Note
Use this gate class only for gates acting on a small numbers of qubits. In general, consider instead using one of the provided ProjectQ gates or define a new class as this allows the compiler to work symbolically.
Example
gate = MatrixGate([[0, 1], [1, 0]]) gate  qubit
 exception projectq.ops._basics.NotInvertible[source]
Exception thrown when trying to invert a gate which is not invertable.
This exception is also thrown if the inverse is not implemented (yet).
 exception projectq.ops._basics.NotMergeable[source]
Exception thrown when trying to merge two gates which are not mergeable.
This exception is also thrown if the merging is not implemented (yet)).
_command
The apply_command function and the Command class.
When a gate is applied to qubits, e.g.,
CNOT  (qubit1, qubit2)
a Command object is generated which represents both the gate, qubits and control qubits. This Command object then gets sent down the compilation pipeline.
In detail, the Gate object overloads the operator (magic method __or__) to generate a Command object which stores the qubits in a canonical order using interchangeable qubit indices defined by the gate to allow the optimizer to cancel the following two gates
Swap  (qubit1, qubit2)
Swap  (qubit2, qubit1)
The command then gets sent to the MainEngine via the apply wrapper (apply_command).
 class projectq.ops._command.Command(engine, gate, qubits, controls=(), tags=(), control_state=CtrlAll.One)[source]
Class used as a container to store commands.
If a gate is applied to qubits, then the gate and qubits are saved in a command object. Qubits are copied into WeakQubitRefs in order to allow early deallocation (would be kept alive otherwise). WeakQubitRef qubits don’t send deallocate gate when destructed.
 gate
The gate to execute
 qubits[source]
Tuple of qubit lists (e.g. Quregs). Interchangeable qubits are stored in a unique order
 tags
The list of tag objects associated with this command (e.g., ComputeTag, UncomputeTag, LoopTag, …). tag objects need to support ==, != (__eq__ and __ne__) for comparison as used in e.g. TagRemover. New tags should always be added to the end of the list. This means that if there are e.g. two LoopTags in a command, tag[0] is from the inner scope while tag[1] is from the other scope as the other scope receives the command after the inner scope LoopEngine and hence adds its LoopTag to the end.
 add_control_qubits(qubits, state=CtrlAll.One)[source]
Add (additional) control qubits to this command object.
They are sorted to ensure a canonical order. Also Qubit objects are converted to WeakQubitRef objects to allow garbage collection and thus early deallocation of qubits.
 Parameters
qubits (list of Qubit objects) – List of qubits which control this gate
state (int,str,CtrlAll) – Control state (ie. positive or negative) for the qubits being added as control qubits.
 property all_qubits[source]
Get all qubits (gate and control qubits).
Returns a tuple T where T[0] is a quantum register (a list of WeakQubitRef objects) containing the control qubits and T[1:] contains the quantum registers to which the gate is applied.
 property control_state[source]
Return the state of the control qubits (ie. either positively or negativelycontrolled).
 property engine[source]
Return engine to which the qubits belong / on which the gates are executed.
 get_inverse()[source]
Get the command object corresponding to the inverse of this command.
Inverts the gate (if possible) and creates a new command object from the result.
 Raises
NotInvertible – If the gate does not provide an inverse (see BasicGate.get_inverse)
 get_merged(other)[source]
Merge this command with another one and return the merged command object.
 Parameters
other – Other command to merge with this one (self)
 Raises
NotMergeable – if the gates don’t supply a get_merged()function or can’t be merged for other reasons.
 property interchangeable_qubit_indices[source]
Return nested list of qubit indices which are interchangeable.
Certain qubits can be interchanged (e.g., the qubit order for a Swap gate). To ensure that only those are sorted when determining the ordering (see _order_qubits), self.interchangeable_qubit_indices is used.
Example
If we can interchange qubits 0,1 and qubits 3,4,5, then this function returns [[0,1],[3,4,5]]
 class projectq.ops._command.CtrlAll(value)[source]
Enum type to initialise the control state of qubits.
_gates
Definition of the basic set of quantum gates.
Contains definitions of standard gates such as * Hadamard (H) * PauliX (X / NOT) * PauliY (Y) * PauliZ (Z) * S and its inverse (S / Sdagger) * T and its inverse (T / Tdagger) * SqrtX gate (SqrtX) * Swap gate (Swap) * SqrtSwap gate (SqrtSwap) * Entangle (Entangle) * Phase gate (Ph) * RotationX (Rx) * RotationY (Ry) * RotationZ (Rz) * RotationXX on two qubits (Rxx) * RotationYY on two qubits (Ryy) * RotationZZ on two qubits (Rzz) * Phaseshift (R) * Measurement (Measure)
and meta gates, i.e., * Allocate / Deallocate qubits * Flush gate (end of circuit) * Barrier * FlipBits
 projectq.ops._gates.Allocate = <projectq.ops._gates.AllocateQubitGate object>
Shortcut (instance of)
projectq.ops.AllocateQubitGate
 projectq.ops._gates.AllocateDirty = <projectq.ops._gates.AllocateDirtyQubitGate object>
Shortcut (instance of)
projectq.ops.AllocateDirtyQubitGate
 projectq.ops._gates.Barrier = <projectq.ops._gates.BarrierGate object>
Shortcut (instance of)
projectq.ops.BarrierGate
 projectq.ops._gates.Deallocate = <projectq.ops._gates.DeallocateQubitGate object>
Shortcut (instance of)
projectq.ops.DeallocateQubitGate
 projectq.ops._gates.Entangle = <projectq.ops._gates.EntangleGate object>
Shortcut (instance of)
projectq.ops.EntangleGate
 class projectq.ops._gates.EntangleGate[source]
Entangle gate class.
(Hadamard on first qubit, followed by CNOTs applied to all other qubits).
 class projectq.ops._gates.FlipBits(bits_to_flip)[source]
Gate for flipping qubits by means of XGates.
 class projectq.ops._gates.FlushGate[source]
Flush gate (denotes the end of the circuit).
Note
All compiler engines (cengines) which cache/buffer gates are obligated to flush and send all gates to the next compiler engine (followed by the flush command).
Note
This gate is sent when calling
eng.flush()
on the MainEngine eng.
 projectq.ops._gates.H = <projectq.ops._gates.HGate object>
Shortcut (instance of)
projectq.ops.HGate
 projectq.ops._gates.Measure = <projectq.ops._gates.MeasureGate object>
Shortcut (instance of)
projectq.ops.MeasureGate
 projectq.ops._gates.NOT = <projectq.ops._gates.XGate object>
Shortcut (instance of)
projectq.ops.XGate
 class projectq.ops._gates.R(angle)[source]
Phaseshift gate (equivalent to Rz up to a global phase).
 projectq.ops._gates.S = <projectq.ops._gates.SGate object>
Shortcut (instance of)
projectq.ops.SGate
 projectq.ops._gates.Sdag = <projectq.ops._metagates.DaggeredGate object>
Inverse (and shortcut) of
projectq.ops.SGate
 projectq.ops._gates.Sdagger = <projectq.ops._metagates.DaggeredGate object>
Inverse (and shortcut) of
projectq.ops.SGate
 projectq.ops._gates.SqrtSwap = <projectq.ops._gates.SqrtSwapGate object>
Shortcut (instance of)
projectq.ops.SqrtSwapGate
 projectq.ops._gates.SqrtX = <projectq.ops._gates.SqrtXGate object>
Shortcut (instance of)
projectq.ops.SqrtXGate
 projectq.ops._gates.Swap = <projectq.ops._gates.SwapGate object>
Shortcut (instance of)
projectq.ops.SwapGate
 projectq.ops._gates.T = <projectq.ops._gates.TGate object>
Shortcut (instance of)
projectq.ops.TGate
 projectq.ops._gates.Tdag = <projectq.ops._metagates.DaggeredGate object>
Inverse (and shortcut) of
projectq.ops.TGate
 projectq.ops._gates.Tdagger = <projectq.ops._metagates.DaggeredGate object>
Inverse (and shortcut) of
projectq.ops.TGate
 projectq.ops._gates.X = <projectq.ops._gates.XGate object>
Shortcut (instance of)
projectq.ops.XGate
 projectq.ops._gates.Y = <projectq.ops._gates.YGate object>
Shortcut (instance of)
projectq.ops.YGate
 projectq.ops._gates.Z = <projectq.ops._gates.ZGate object>
Shortcut (instance of)
projectq.ops.ZGate
_metagates
Definition of some meta gates.
Contains meta gates, i.e., * DaggeredGate (Represents the inverse of an arbitrary gate) * ControlledGate (Represents a controlled version of an arbitrary gate) * Tensor/All (Applies a single qubit gate to all supplied qubits), e.g.,
 Example:
Tensor(H)  (qubit1, qubit2) # apply H to qubit #1 and #2
As well as the meta functions * get_inverse (Tries to access the get_inverse member function of a gate and upon failure returns a DaggeredGate) * C (Creates an nary controlled version of an arbitrary gate)
 projectq.ops._metagates.All[source]
Shortcut (instance of)
projectq.ops.Tensor
 projectq.ops._metagates.C(gate, n_qubits=1)[source]
Return ncontrolled version of the provided gate.
 Parameters
gate – Gate to turn into its controlled version
n_qubits – Number of controls (default: 1)
Example
C(NOT)  (c, q) # equivalent to CNOT  (c, q)
 exception projectq.ops._metagates.ControlQubitError[source]
Exception thrown when wrong number of control qubits are supplied.
 class projectq.ops._metagates.ControlledGate(gate, n=1)[source]
Controlled version of a gate.
Note
Use the meta function
C()
to create a controlled gateA wrapper class which enables (multi) controlled gates. It overloads the __or__operator, using the first qubits provided as control qubits. The n controlqubits need to be the first n qubits. They can be in separate quregs.
Example
ControlledGate(gate, 2)  (qb0, qb2, qb3) # qb0 & qb2 are controls C(gate, 2)  (qb0, qb2, qb3) # This is much nicer. C(gate, 2)  ([qb0, qb2], qb3) # Is equivalent
 class projectq.ops._metagates.DaggeredGate(gate)[source]
Wrapper class allowing to execute the inverse of a gate, even when it does not define one.
If there is a replacement available, then there is also one for the inverse, namely the replacement function run in reverse, while inverting all gates. This class enables using this emulation automatically.
A DaggeredGate is returned automatically when employing the get_inverse function on a gate which does not provide a get_inverse() member function.
Example
with Dagger(eng): MySpecialGate  qubits
will create a DaggeredGate if MySpecialGate does not implement get_inverse. If there is a decomposition function available, an auto replacer engine can automatically replace the inverted gate by a call to the decomposition function inside a “with Dagger”statement.
 class projectq.ops._metagates.Tensor(gate)[source]
Wrapper class allowing to apply a (singlequbit) gate to every qubit in a quantum register.
Allowed syntax is to supply either a qureg or a tuple which contains only one qureg.
Example
Tensor(H)  x # applies H to every qubit in the list of qubits x Tensor(H)  (x,) # alternative to be consistent with other syntax
 projectq.ops._metagates.get_inverse(gate)[source]
Return the inverse of a gate.
Tries to call gate.get_inverse and, upon failure, creates a DaggeredGate instead.
 Parameters
gate – Gate of which to get the inverse
Example
get_inverse(H) # returns a Hadamard gate (HGate object)
 projectq.ops._metagates.is_identity(gate)[source]
Return True if the gate is an identity gate.
Tries to call gate.is_identity and, upon failure, returns False
 Parameters
gate – Gate of which to get the inverse
Example
get_inverse(Rx(2 * math.pi)) # returns True get_inverse(Rx(math.pi)) # returns False
_qaagate
Definition of the quantum amplitude amplification gate.
 class projectq.ops._qaagate.QAA(algorithm, oracle)[source]
Quantum Aplitude Amplification gate.
(Quick reference https://en.wikipedia.org/wiki/Amplitude_amplification. Complete reference G. Brassard, P. Hoyer, M. Mosca, A. Tapp (2000) Quantum Amplitude Amplification and Estimation https://arxiv.org/abs/quantph/0005055)
Quantum Amplitude Amplification (QAA) executes the algorithm, but not the final measurement required to obtain the marked state(s) with high probability. The starting state on wich the QAA algorithm is executed is the one resulting of aplying the Algorithm on the 0> state.
Example
def func_algorithm(eng, system_qubits): All(H)  system_qubits def func_oracle(eng, system_qubits, qaa_ancilla): # This oracle selects the state 010> as the one marked with Compute(eng): All(X)  system_qubits[0::2] with Control(eng, system_qubits): X  qaa_ancilla Uncompute(eng) system_qubits = eng.allocate_qureg(3) # Prepare the qaa_ancilla qubit in the > state qaa_ancilla = eng.allocate_qubit() X  qaa_ancilla H  qaa_ancilla # Creates the initial state form the Algorithm func_algorithm(eng, system_qubits) # Apply Quantum Amplitude Amplification the correct number of times num_it = int(math.pi / 4.0 * math.sqrt(1 << 3)) with Loop(eng, num_it): QAA(func_algorithm, func_oracle)  (system_qubits, qaa_ancilla) All(Measure)  system_qubits
Warning
No qubit allocation/deallocation may take place during the call to the defined Algorithm
func_algorithm
 func_algorithm
Algorithm that initialite the state and to be used in the QAA algorithm
 func_oracle
The Oracle that marks the state(s) as “good”
 system_qubits
the system we are interested on
 qaa_ancilla
auxiliary qubit that helps to invert the amplitude of the “good” states
_qftgate
Definition of the QFT gate.
 projectq.ops._qftgate.QFT = <projectq.ops._qftgate.QFTGate object>
Shortcut (instance of)
projectq.ops.QFTGate
_qpegate
Definition of the quantum phase estimation gate.
_qubit_operator
QubitOperator stores a sum of Pauli operators acting on qubits.
 class projectq.ops._qubit_operator.QubitOperator(term=None, coefficient=1.0)[source]
A sum of terms acting on qubits, e.g., 0.5 * ‘X0 X5’ + 0.3 * ‘Z1 Z2’.
A term is an operator acting on n qubits and can be represented as:
coefficent * local_operator[0] x … x local_operator[n1]
where x is the tensor product. A local operator is a Pauli operator (‘I’, ‘X’, ‘Y’, or ‘Z’) which acts on one qubit. In math notation a term is, for example, 0.5 * ‘X0 X5’, which means that a Pauli X operator acts on qubit 0 and 5, while the identity operator acts on all other qubits.
A QubitOperator represents a sum of terms acting on qubits and overloads operations for easy manipulation of these objects by the user.
Note for a QubitOperator to be a Hamiltonian which is a hermitian operator, the coefficients of all terms must be real.
hamiltonian = 0.5 * QubitOperator('X0 X5') + 0.3 * QubitOperator('Z0')
Our Simulator takes a hermitian QubitOperator to directly calculate the expectation value (see Simulator.get_expectation_value) of this observable.
A hermitian QubitOperator can also be used as input for the TimeEvolution gate.
If the QubitOperator is unitary, i.e., it contains only one term with a coefficient, whose absolute value is 1, then one can apply it directly to qubits:
eng = projectq.MainEngine() qureg = eng.allocate_qureg(6) QubitOperator('X0 X5', 1.0j)  qureg # Applies X to qubit 0 and 5 with an additional global phase of 1.j
 terms
key: A term represented by a tuple containing all nontrivial local Pauli operators (‘X’, ‘Y’, or ‘Z’). A nontrivial local Pauli operator is specified by a tuple with the first element being an integer indicating the qubit on which a nontrivial local operator acts and the second element being a string, either ‘X’, ‘Y’, or ‘Z’, indicating which nontrivial Pauli operator acts on that qubit. Examples: ((1, ‘X’),) or ((1, ‘X’), (4,’Z’)) or the identity (). The tuples representing the nontrivial local terms are sorted according to the qubit number they act on, starting from 0. value: Coefficient of this term as a (complex) float
 Type
dict
 compress(abs_tol=1e12)[source]
Compress the coefficient of a QubitOperator.
Eliminate all terms with coefficients close to zero and removes imaginary parts of coefficients that are close to zero.
 Parameters
abs_tol (float) – Absolute tolerance, must be at least 0.0
 get_inverse()[source]
Return the inverse gate of a QubitOperator if applied as a gate.
 Raises
NotInvertible – Not implemented for QubitOperators which have multiple terms or a coefficient with absolute value not equal to 1.
 get_merged(other)[source]
Return this gate merged with another gate.
Standard implementation of get_merged:
 Raises
NotMergeable – merging is not possible
 isclose(other, rel_tol=1e12, abs_tol=1e12)[source]
Return True if other (QubitOperator) is close to self.
Comparison is done for each term individually. Return True if the difference between each term in self and other is less than the relative tolerance w.r.t. either other or self (symmetric test) or if the difference is less than the absolute tolerance.
 Parameters
other (QubitOperator) – QubitOperator to compare against.
rel_tol (float) – Relative tolerance, must be greater than 0.0
abs_tol (float) – Absolute tolerance, must be at least 0.0
_shortcuts
A few shortcuts for certain gates.
These include: * CNOT = C(NOT) * CRz = C(Rz) * Toffoli = C(NOT,2) = C(CNOT)
_state_prep
Definition of the state preparation gate.
_time_evolution
Definition of the time evolution gate.
 exception projectq.ops._time_evolution.NotHermitianOperatorError[source]
Error raised if an operator is nonhermitian.
 class projectq.ops._time_evolution.TimeEvolution(time, hamiltonian)[source]
Gate for time evolution under a Hamiltonian (QubitOperator object).
This gate is the unitary time evolution propagator: exp(i * H * t), where H is the Hamiltonian of the system and t is the time. Note that i factor is stored implicitely.
Example
wavefunction = eng.allocate_qureg(5) hamiltonian = 0.5 * QubitOperator("X0 Z1 Y5") # Apply exp(i * H * t) to the wavefunction: TimeEvolution(time=2.0, hamiltonian=hamiltonian)  wavefunction
 time
time t
 Type
float, int
 hamiltonian
hamiltonaian H
 Type
 get_merged(other)[source]
Return self merged with another TimeEvolution gate if possible.
 Two TimeEvolution gates are merged if:
both have the same terms
the proportionality factor for each of the terms must have relative error <= 1e9 compared to the proportionality factors of the other terms.
Note
While one could merge gates for which both hamiltonians commute, we are not doing this as in general the resulting gate would have to be decomposed again.
Note
We are not comparing if terms are proportional to each other with an absolute tolerance. It is up to the user to remove terms close to zero because we cannot choose a suitable absolute error which works for everyone. Use, e.g., a decomposition rule for that.
 Parameters
other – TimeEvolution gate
 Raises
NotMergeable – If the other gate is not a TimeEvolution gate or hamiltonians are not suitable for merging.
 Returns
New TimeEvolution gate equivalent to the two merged gates.
_uniformly_controlled_rotation
Definition of uniformly controlled Ry and Rzrotation gates.
 class projectq.ops._uniformly_controlled_rotation.UniformlyControlledRy(angles)[source]
Uniformly controlled Ry gate as introduced in arXiv:quantph/0312218.
This is an nqubit gate. There are n1 control qubits and one target qubit. This gate applies Ry(angles(k)) to the target qubit if the n1 control qubits are in the classical state k. As there are 2^(n1) classical states for the control qubits, this gate requires 2^(n1) (potentially different) angle parameters.
Example
controls = eng.allocate_qureg(2) target = eng.allocate_qubit() UniformlyControlledRy(angles=[0.1, 0.2, 0.3, 0.4])  (controls, target)
Note
The first quantum register contains the control qubits. When converting the classical state k of the control qubits to an integer, we define controls[0] to be the least significant (qu)bit. controls can also be an empty list in which case the gate corresponds to an Ry.
 Parameters
angles (list[float]) – Rotation angles. Ry(angles[k]) is applied conditioned on the control qubits being in state k.
 class projectq.ops._uniformly_controlled_rotation.UniformlyControlledRz(angles)[source]
Uniformly controlled Rz gate as introduced in arXiv:quantph/0312218.
This is an nqubit gate. There are n1 control qubits and one target qubit. This gate applies Rz(angles(k)) to the target qubit if the n1 control qubits are in the classical state k. As there are 2^(n1) classical states for the control qubits, this gate requires 2^(n1) (potentially different) angle parameters.
Example
controls = eng.allocate_qureg(2) target = eng.allocate_qubit() UniformlyControlledRz(angles=[0.1, 0.2, 0.3, 0.4])  (controls, target)
Note
The first quantum register are the contains qubits. When converting the classical state k of the control qubits to an integer, we define controls[0] to be the least significant (qu)bit. controls can also be an empty list in which case the gate corresponds to an Rz.
 Parameters
angles (list[float]) – Rotation angles. Rz(angles[k]) is applied conditioned on the control qubits being in state k.
Module contents
ProjectQ module containing all basic gates (operations)
 class projectq.ops.BasicGate[source]
Base class of all gates. (Don’t use it directly but derive from it).
 __init__()[source]
Initialize a basic gate.
Note
Set interchangeable qubit indices! (gate.interchangeable_qubit_indices)
As an example, consider
ExampleGate  (a, b, c, d, e)
where a and b are interchangeable. Then, call this function as follows:
self.set_interchangeable_qubit_indices([[0, 1]])
As another example, consider
ExampleGate2  (a, b, c, d, e)
where a and b are interchangeable and, in addition, c, d, and e are interchangeable among themselves. Then, call this function as
self.set_interchangeable_qubit_indices([[0, 1], [2, 3, 4]])
 __or__(qubits)[source]
Operator overload which enables the syntax Gate  qubits.
Example
Gate  qubit
Gate  [qubit0, qubit1]
Gate  qureg
Gate  (qubit, )
Gate  (qureg, qubit)
 Parameters
qubits – a Qubit object, a list of Qubit objects, a Qureg object, or a tuple of Qubit or Qureg objects (can be mixed).
 generate_command(qubits)[source]
Generate a command.
The command object created consists of the gate and the qubits being acted upon.
 Parameters
qubits – see BasicGate.make_tuple_of_qureg(qubits)
 Returns
A Command object containing the gate and the qubits.
 get_inverse()[source]
Return the inverse gate.
Standard implementation of get_inverse:
 Raises
NotInvertible – inverse is not implemented
 get_merged(other)[source]
Return this gate merged with another gate.
Standard implementation of get_merged:
 Raises
NotMergeable – merging is not implemented
 is_identity()[source]
Return True if the gate is an identity gate. In this base class, always returns False.
 static make_tuple_of_qureg(qubits)[source]
Convert quantum input of “gate  quantum input” to internal formatting.
A Command object only accepts tuples of Quregs (list of Qubit objects) as qubits input parameter. However, with this function we allow the user to use a more flexible syntax:
Gate  qubit
Gate  [qubit0, qubit1]
Gate  qureg
Gate  (qubit, )
Gate  (qureg, qubit)
where qubit is a Qubit object and qureg is a Qureg object. This function takes the right hand side of  and transforms it to the correct input parameter of a Command object which is:
> Gate  ([qubit], )
> Gate  ([qubit0, qubit1], )
> Gate  (qureg, )
> Gate  ([qubit], )
> Gate  (qureg, [qubit])
 Parameters
qubits – a Qubit object, a list of Qubit objects, a Qureg object, or a tuple of Qubit or Qureg objects (can be mixed).
 Returns
A tuple containing Qureg (or list of Qubits) objects.
 Return type
Canonical representation (tuple<qureg>)
 class projectq.ops.BasicMathGate(math_fun)[source]
Base class for all math gates.
It allows efficient emulation by providing a mathematical representation which is given by the concrete gate which derives from this base class. The AddConstant gate, for example, registers a function of the form
def add(x): return (x + a,)
upon initialization. More generally, the function takes integers as parameters and returns a tuple / list of outputs, each entry corresponding to the function input. As an example, consider outofplace multiplication, which takes two input registers and adds the result into a third, i.e., (a,b,c) > (a,b,c+a*b). The corresponding function then is
def multiply(a, b, c): return (a, b, c + a * b)
 __init__(math_fun)[source]
Initialize a BasicMathGate by providing the mathematical function that it implements.
 Parameters
math_fun (function) – Function which takes as many int values as input, as the gate takes registers. For each of these values, it then returns the output (i.e., it returns a list/tuple of output values).
Example
def add(a, b): return (a, a + b) super().__init__(add)
If the gate acts on, e.g., fixed point numbers, the number of bits per register is also required in order to describe the action of such a mathematical gate. For this reason, there is
BasicMathGate.get_math_function(qubits)
which can be overwritten by the gate deriving from BasicMathGate.
Example
def get_math_function(self, qubits): n = len(qubits[0]) scal = 2.0**n def math_fun(a): return (int(scal * (math.sin(math.pi * a / scal))),) return math_fun
 get_math_function(qubits)[source]
Get the math function associated with a BasicMathGate.
Return the math function which corresponds to the action of this math gate, given the input to the gate (a tuple of quantum registers).
 Parameters
qubits (tuple<Qureg>) – Qubits to which the math gate is being applied.
 Returns
Python function describing the action of this gate. (See BasicMathGate.__init__ for an example).
 Return type
math_fun (function)
 class projectq.ops.BasicPhaseGate(angle)[source]
Base class for all phase gates.
A phase gate has a continuous parameter (the angle), labeled ‘angle’ / self.angle. Its inverse is the same gate with the negated argument. Phase gates of the same class can be merged by adding the angles. The continuous parameter is modulo 2 * pi, self.angle is in the interval [0, 2 * pi).
 __init__(angle)[source]
Initialize a basic rotation gate.
 Parameters
angle (float) – Angle of rotation (saved modulo 2 * pi)
 get_inverse()[source]
Return the inverse of this rotation gate (negate the angle, return new object).
 get_merged(other)[source]
Return self merged with another gate.
Default implementation handles rotation gate of the same type, where angles are simply added.
 Parameters
other – Rotation gate of same type.
 Raises
NotMergeable – For nonrotation gates or rotation gates of different type.
 Returns
New object representing the merged gates.
 class projectq.ops.BasicRotationGate(angle)[source]
Base class of for all rotation gates.
A rotation gate has a continuous parameter (the angle), labeled ‘angle’ / self.angle. Its inverse is the same gate with the negated argument. Rotation gates of the same class can be merged by adding the angles. The continuous parameter is modulo 4 * pi, self.angle is in the interval [0, 4 * pi).
 __init__(angle)[source]
Initialize a basic rotation gate.
 Parameters
angle (float) – Angle of rotation (saved modulo 4 * pi)
 get_inverse()[source]
Return the inverse of this rotation gate (negate the angle, return new object).
 get_merged(other)[source]
Return self merged with another gate.
Default implementation handles rotation gate of the same type, where angles are simply added.
 Parameters
other – Rotation gate of same type.
 Raises
NotMergeable – For nonrotation gates or rotation gates of different type.
 Returns
New object representing the merged gates.
 projectq.ops.C(gate, n_qubits=1)[source]
Return ncontrolled version of the provided gate.
 Parameters
gate – Gate to turn into its controlled version
n_qubits – Number of controls (default: 1)
Example
C(NOT)  (c, q) # equivalent to CNOT  (c, q)
 class projectq.ops.ClassicalInstructionGate[source]
Classical instruction gate.
Base class for all gates which are not quantum gates in the typical sense, e.g., measurement, allocation/deallocation, …
 class projectq.ops.Command(engine, gate, qubits, controls=(), tags=(), control_state=CtrlAll.One)[source]
Class used as a container to store commands.
If a gate is applied to qubits, then the gate and qubits are saved in a command object. Qubits are copied into WeakQubitRefs in order to allow early deallocation (would be kept alive otherwise). WeakQubitRef qubits don’t send deallocate gate when destructed.
 gate
The gate to execute
 qubits[source]
Tuple of qubit lists (e.g. Quregs). Interchangeable qubits are stored in a unique order
 tags
The list of tag objects associated with this command (e.g., ComputeTag, UncomputeTag, LoopTag, …). tag objects need to support ==, != (__eq__ and __ne__) for comparison as used in e.g. TagRemover. New tags should always be added to the end of the list. This means that if there are e.g. two LoopTags in a command, tag[0] is from the inner scope while tag[1] is from the other scope as the other scope receives the command after the inner scope LoopEngine and hence adds its LoopTag to the end.
 __init__(engine, gate, qubits, controls=(), tags=(), control_state=CtrlAll.One)[source]
Initialize a Command object.
Note
control qubits (Command.control_qubits) are stored as a list of qubits, and command tags (Command.tags) as a list of tagobjects. All functions within this class also work if WeakQubitRefs are supplied instead of normal Qubit objects (see WeakQubitRef).
 Parameters
engine (projectq.cengines.BasicEngine) – engine which created the qubit (mostly the MainEngine)
gate (projectq.ops.Gate) – Gate to be executed
qubits (tuple[Qureg]) – Tuple of quantum registers (to which the gate is applied)
controls (Qureglist[Qubit]) – Qubits that condition the command.
tags (list[object]) – Tags associated with the command.
control_state (int,str,projectq.meta.CtrlAll) –
 add_control_qubits(qubits, state=CtrlAll.One)[source]
Add (additional) control qubits to this command object.
They are sorted to ensure a canonical order. Also Qubit objects are converted to WeakQubitRef objects to allow garbage collection and thus early deallocation of qubits.
 Parameters
qubits (list of Qubit objects) – List of qubits which control this gate
state (int,str,CtrlAll) – Control state (ie. positive or negative) for the qubits being added as control qubits.
 property all_qubits[source]
Get all qubits (gate and control qubits).
Returns a tuple T where T[0] is a quantum register (a list of WeakQubitRef objects) containing the control qubits and T[1:] contains the quantum registers to which the gate is applied.
 property control_state[source]
Return the state of the control qubits (ie. either positively or negativelycontrolled).
 property engine[source]
Return engine to which the qubits belong / on which the gates are executed.
 get_inverse()[source]
Get the command object corresponding to the inverse of this command.
Inverts the gate (if possible) and creates a new command object from the result.
 Raises
NotInvertible – If the gate does not provide an inverse (see BasicGate.get_inverse)
 get_merged(other)[source]
Merge this command with another one and return the merged command object.
 Parameters
other – Other command to merge with this one (self)
 Raises
NotMergeable – if the gates don’t supply a get_merged()function or can’t be merged for other reasons.
 property interchangeable_qubit_indices[source]
Return nested list of qubit indices which are interchangeable.
Certain qubits can be interchanged (e.g., the qubit order for a Swap gate). To ensure that only those are sorted when determining the ordering (see _order_qubits), self.interchangeable_qubit_indices is used.
Example
If we can interchange qubits 0,1 and qubits 3,4,5, then this function returns [[0,1],[3,4,5]]
 class projectq.ops.ControlledGate(gate, n=1)[source]
Controlled version of a gate.
Note
Use the meta function
C()
to create a controlled gateA wrapper class which enables (multi) controlled gates. It overloads the __or__operator, using the first qubits provided as control qubits. The n controlqubits need to be the first n qubits. They can be in separate quregs.
Example
ControlledGate(gate, 2)  (qb0, qb2, qb3) # qb0 & qb2 are controls C(gate, 2)  (qb0, qb2, qb3) # This is much nicer. C(gate, 2)  ([qb0, qb2], qb3) # Is equivalent
 __init__(gate, n=1)[source]
Initialize a ControlledGate object.
 Parameters
gate – Gate to wrap.
n (int) – Number of control qubits.
 __or__(qubits)[source]
Apply the controlled gate to qubits, using the first n qubits as controls.
 Note: The control qubits can be split across the first quregs. However, the nth control qubit needs to be
the last qubit in a qureg. The following quregs belong to the gate.
 Parameters
qubits (tuple of lists of Qubit objects) – qubits to which to apply the gate.
 class projectq.ops.DaggeredGate(gate)[source]
Wrapper class allowing to execute the inverse of a gate, even when it does not define one.
If there is a replacement available, then there is also one for the inverse, namely the replacement function run in reverse, while inverting all gates. This class enables using this emulation automatically.
A DaggeredGate is returned automatically when employing the get_inverse function on a gate which does not provide a get_inverse() member function.
Example
with Dagger(eng): MySpecialGate  qubits
will create a DaggeredGate if MySpecialGate does not implement get_inverse. If there is a decomposition function available, an auto replacer engine can automatically replace the inverted gate by a call to the decomposition function inside a “with Dagger”statement.
 __init__(gate)[source]
Initialize a DaggeredGate representing the inverse of the gate ‘gate’.
 Parameters
gate – Any gate object of which to represent the inverse.
 class projectq.ops.EntangleGate[source]
Entangle gate class.
(Hadamard on first qubit, followed by CNOTs applied to all other qubits).
 class projectq.ops.FastForwardingGate[source]
Base class for fastforward gates.
Base class for classical instruction gates which require a fastforward through compiler engines that cache / buffer gates. Examples include Measure and Deallocate, which both should be executed asap, such that Measurement results are available and resources are freed, respectively.
Note
The only requirement is that FlushGate commands run the entire circuit. FastForwardingGate objects can be used but the user cannot expect a measurement result to be available for all backends when calling only Measure. E.g., for the IBM Quantum Experience backend, sending the circuit for each Measuregate would be too inefficient, which is why a final
is required before the circuit gets sent through the API.
 class projectq.ops.FlipBits(bits_to_flip)[source]
Gate for flipping qubits by means of XGates.
 __init__(bits_to_flip)[source]
Initialize a FlipBits gate.
Example
qureg = eng.allocate_qureg(2) FlipBits([0, 1])  qureg
 Parameters
bits_to_flip (list[int]list[bool]strint) – int or array of 0/1, True/False, or string of 0/1 identifying the qubits to flip. In case of int, the bits to flip are determined from the binary digits, with the least significant bit corresponding to qureg[0]. If bits_to_flip is negative, exactly all qubits which would not be flipped for the input bits_to_flip1 are flipped, i.e., bits_to_flip=1 flips all qubits.
 class projectq.ops.FlushGate[source]
Flush gate (denotes the end of the circuit).
Note
All compiler engines (cengines) which cache/buffer gates are obligated to flush and send all gates to the next compiler engine (followed by the flush command).
Note
This gate is sent when calling
eng.flush()
on the MainEngine eng.
 exception projectq.ops.IncompatibleControlState[source]
Exception thrown when trying to set two incompatible states for a control qubit.
 class projectq.ops.MatrixGate(matrix=None)[source]
A gate class whose instances are defined by a matrix.
Note
Use this gate class only for gates acting on a small numbers of qubits. In general, consider instead using one of the provided ProjectQ gates or define a new class as this allows the compiler to work symbolically.
Example
gate = MatrixGate([[0, 1], [1, 0]]) gate  qubit
 class projectq.ops.MeasureGate[source]
Measurement gate class (for single qubits).
 __or__(qubits)[source]
Operator overload which enables the syntax Gate  qubits.
Previously (ProjectQ <= v0.3.6) MeasureGate/Measure was allowed to be applied to any number of quantum registers. Now the MeasureGate/Measure is strictly a single qubit gate.
 Raises
RuntimeError – Since ProjectQ v0.6.0 if the gate is applied to multiple qubits.
 exception projectq.ops.NotInvertible[source]
Exception thrown when trying to invert a gate which is not invertable.
This exception is also thrown if the inverse is not implemented (yet).
 exception projectq.ops.NotMergeable[source]
Exception thrown when trying to merge two gates which are not mergeable.
This exception is also thrown if the merging is not implemented (yet)).
 class projectq.ops.QAA(algorithm, oracle)[source]
Quantum Aplitude Amplification gate.
(Quick reference https://en.wikipedia.org/wiki/Amplitude_amplification. Complete reference G. Brassard, P. Hoyer, M. Mosca, A. Tapp (2000) Quantum Amplitude Amplification and Estimation https://arxiv.org/abs/quantph/0005055)
Quantum Amplitude Amplification (QAA) executes the algorithm, but not the final measurement required to obtain the marked state(s) with high probability. The starting state on wich the QAA algorithm is executed is the one resulting of aplying the Algorithm on the 0> state.
Example
def func_algorithm(eng, system_qubits): All(H)  system_qubits def func_oracle(eng, system_qubits, qaa_ancilla): # This oracle selects the state 010> as the one marked with Compute(eng): All(X)  system_qubits[0::2] with Control(eng, system_qubits): X  qaa_ancilla Uncompute(eng) system_qubits = eng.allocate_qureg(3) # Prepare the qaa_ancilla qubit in the > state qaa_ancilla = eng.allocate_qubit() X  qaa_ancilla H  qaa_ancilla # Creates the initial state form the Algorithm func_algorithm(eng, system_qubits) # Apply Quantum Amplitude Amplification the correct number of times num_it = int(math.pi / 4.0 * math.sqrt(1 << 3)) with Loop(eng, num_it): QAA(func_algorithm, func_oracle)  (system_qubits, qaa_ancilla) All(Measure)  system_qubits
Warning
No qubit allocation/deallocation may take place during the call to the defined Algorithm
func_algorithm
 func_algorithm
Algorithm that initialite the state and to be used in the QAA algorithm
 func_oracle
The Oracle that marks the state(s) as “good”
 system_qubits
the system we are interested on
 qaa_ancilla
auxiliary qubit that helps to invert the amplitude of the “good” states
 class projectq.ops.QPE(unitary)[source]
Quantum Phase Estimation gate.
See setups.decompositions for the complete implementation
 class projectq.ops.QubitOperator(term=None, coefficient=1.0)[source]
A sum of terms acting on qubits, e.g., 0.5 * ‘X0 X5’ + 0.3 * ‘Z1 Z2’.
A term is an operator acting on n qubits and can be represented as:
coefficent * local_operator[0] x … x local_operator[n1]
where x is the tensor product. A local operator is a Pauli operator (‘I’, ‘X’, ‘Y’, or ‘Z’) which acts on one qubit. In math notation a term is, for example, 0.5 * ‘X0 X5’, which means that a Pauli X operator acts on qubit 0 and 5, while the identity operator acts on all other qubits.
A QubitOperator represents a sum of terms acting on qubits and overloads operations for easy manipulation of these objects by the user.
Note for a QubitOperator to be a Hamiltonian which is a hermitian operator, the coefficients of all terms must be real.
hamiltonian = 0.5 * QubitOperator('X0 X5') + 0.3 * QubitOperator('Z0')
Our Simulator takes a hermitian QubitOperator to directly calculate the expectation value (see Simulator.get_expectation_value) of this observable.
A hermitian QubitOperator can also be used as input for the TimeEvolution gate.
If the QubitOperator is unitary, i.e., it contains only one term with a coefficient, whose absolute value is 1, then one can apply it directly to qubits:
eng = projectq.MainEngine() qureg = eng.allocate_qureg(6) QubitOperator('X0 X5', 1.0j)  qureg # Applies X to qubit 0 and 5 with an additional global phase of 1.j
 terms
key: A term represented by a tuple containing all nontrivial local Pauli operators (‘X’, ‘Y’, or ‘Z’). A nontrivial local Pauli operator is specified by a tuple with the first element being an integer indicating the qubit on which a nontrivial local operator acts and the second element being a string, either ‘X’, ‘Y’, or ‘Z’, indicating which nontrivial Pauli operator acts on that qubit. Examples: ((1, ‘X’),) or ((1, ‘X’), (4,’Z’)) or the identity (). The tuples representing the nontrivial local terms are sorted according to the qubit number they act on, starting from 0. value: Coefficient of this term as a (complex) float
 Type
dict
 __init__(term=None, coefficient=1.0)[source]
Initialize a QubitOperator object.
The init function only allows to initialize one term. Additional terms have to be added using += (which is fast) or using + of two QubitOperator objects:
Example
ham = QubitOperator('X0 Y3', 0.5) + 0.6 * QubitOperator('X0 Y3') # Equivalently ham2 = QubitOperator('X0 Y3', 0.5) ham2 += 0.6 * QubitOperator('X0 Y3')
Note
Adding terms to QubitOperator is faster using += (as this is done by inplace addition). Specifying the coefficient in the __init__ is faster than by multiplying a QubitOperator with a scalar as calls an outofplace multiplication.
 Parameters
coefficient (complex float, optional) – The coefficient of the first term of this QubitOperator. Default is 1.0.
term (optional, empy tuple, a tuple of tuples, or a string) –
Default is None which means there are no terms in the QubitOperator hence it is the “zero” Operator
An empty tuple means there are no nontrivial Pauli operators acting on the qubits hence only identities with a coefficient (which by default is 1.0).
A sorted tuple of tuples. The first element of each tuple is an integer indicating the qubit on which a nontrivial local operator acts, starting from zero. The second element of each tuple is a string, either ‘X’, ‘Y’ or ‘Z’, indicating which local operator acts on that qubit.
A string of the form ‘X0 Z2 Y5’, indicating an X on qubit 0, Z on qubit 2, and Y on qubit 5. The string should be sorted by the qubit number. ‘’ is the identity.
 Raises
QubitOperatorError – Invalid operators provided to QubitOperator.
 __or__(qubits)[source]
Operator overload which enables the syntax Gate  qubits.
In particular, enable the following syntax:
QubitOperator(...)  qureg QubitOperator(...)  (qureg,) QubitOperator(...)  qubit QubitOperator(...)  (qubit,)
Unlike other gates, this gate is only allowed to be applied to one quantum register or one qubit and only if the QubitOperator is unitary, i.e., consists of one term with a coefficient whose absolute values is 1.
Example:
eng = projectq.MainEngine() qureg = eng.allocate_qureg(6) QubitOperator('X0 X5', 1.0j)  qureg # Applies X to qubit 0 and 5 # with an additional global # phase of 1.j
While in the above example the QubitOperator gate is applied to 6 qubits, it only acts nontrivially on the two qubits qureg[0] and qureg[5]. Therefore, the operator will create a new rescaled QubitOperator, i.e, it sends the equivalent of the following new gate to the MainEngine:
QubitOperator('X0 X1', 1.0j)  [qureg[0], qureg[5]]
which is only a two qubit gate.
 Parameters
qubits – one Qubit object, one list of Qubit objects, one Qureg object, or a tuple of the former three cases.
 Raises
TypeError – If QubitOperator is not unitary or applied to more than one quantum register.
ValueError – If quantum register does not have enough qubits
 compress(abs_tol=1e12)[source]
Compress the coefficient of a QubitOperator.
Eliminate all terms with coefficients close to zero and removes imaginary parts of coefficients that are close to zero.
 Parameters
abs_tol (float) – Absolute tolerance, must be at least 0.0
 get_inverse()[source]
Return the inverse gate of a QubitOperator if applied as a gate.
 Raises
NotInvertible – Not implemented for QubitOperators which have multiple terms or a coefficient with absolute value not equal to 1.
 get_merged(other)[source]
Return this gate merged with another gate.
Standard implementation of get_merged:
 Raises
NotMergeable – merging is not possible
 isclose(other, rel_tol=1e12, abs_tol=1e12)[source]
Return True if other (QubitOperator) is close to self.
Comparison is done for each term individually. Return True if the difference between each term in self and other is less than the relative tolerance w.r.t. either other or self (symmetric test) or if the difference is less than the absolute tolerance.
 Parameters
other (QubitOperator) – QubitOperator to compare against.
rel_tol (float) – Relative tolerance, must be greater than 0.0
abs_tol (float) – Absolute tolerance, must be at least 0.0
 class projectq.ops.SelfInverseGate[source]
Selfinverse basic gate class.
Automatic implementation of the get_inversemember function for selfinverse gates.
Example
# get_inverse(H) == H, it is a selfinverse gate: get_inverse(H)  qubit
 class projectq.ops.StatePreparation(final_state)[source]
Gate for transforming qubits in state 0> to any desired quantum state.
 __init__(final_state)[source]
Initialize a StatePreparation gate.
Example
qureg = eng.allocate_qureg(2) StatePreparation([0.5, 0.5j, 0.5, 0.5])  qureg
Note
final_state[k] is taken to be the amplitude of the computational basis state whose string is equal to the binary representation of k.
 Parameters
final_state (list[complex]) – wavefunction of the desired quantum state. len(final_state) must be 2**len(qureg). Must be normalized!
 class projectq.ops.Tensor(gate)[source]
Wrapper class allowing to apply a (singlequbit) gate to every qubit in a quantum register.
Allowed syntax is to supply either a qureg or a tuple which contains only one qureg.
Example
Tensor(H)  x # applies H to every qubit in the list of qubits x Tensor(H)  (x,) # alternative to be consistent with other syntax
 class projectq.ops.TimeEvolution(time, hamiltonian)[source]
Gate for time evolution under a Hamiltonian (QubitOperator object).
This gate is the unitary time evolution propagator: exp(i * H * t), where H is the Hamiltonian of the system and t is the time. Note that i factor is stored implicitely.
Example
wavefunction = eng.allocate_qureg(5) hamiltonian = 0.5 * QubitOperator("X0 Z1 Y5") # Apply exp(i * H * t) to the wavefunction: TimeEvolution(time=2.0, hamiltonian=hamiltonian)  wavefunction
 time
time t
 Type
float, int
 hamiltonian
hamiltonaian H
 Type
 __init__(time, hamiltonian)[source]
Initialize time evolution gate.
Note
The hamiltonian must be hermitian and therefore only terms with real coefficients are allowed. Coefficients are internally converted to float.
 Parameters
time (float, or int) – time to evolve under (can be negative).
hamiltonian (QubitOperator) – hamiltonian to evolve under.
 Raises
TypeError – If time is not a numeric type and hamiltonian is not a QubitOperator.
NotHermitianOperatorError – If the input hamiltonian is not hermitian (only real coefficients).
 __or__(qubits)[source]
Operator overload which enables the syntax Gate  qubits.
In particular, enable the following syntax:
TimeEvolution(...)  qureg TimeEvolution(...)  (qureg,) TimeEvolution(...)  qubit TimeEvolution(...)  (qubit,)
Unlike other gates, this gate is only allowed to be applied to one quantum register or one qubit.
Example: .. codeblock:: python
wavefunction = eng.allocate_qureg(5) hamiltonian = QubitOperator(“X1 Y3”, 0.5) TimeEvolution(time=2.0, hamiltonian=hamiltonian)  wavefunction
While in the above example the TimeEvolution gate is applied to 5 qubits, the hamiltonian of this TimeEvolution gate acts only nontrivially on the two qubits wavefunction[1] and wavefunction[3]. Therefore, the operator will rescale the indices in the hamiltonian and sends the equivalent of the following new gate to the MainEngine:
h = QubitOperator("X0 Y1", 0.5) TimeEvolution(2.0, h)  [wavefunction[1], wavefunction[3]]
which is only a two qubit gate.
 Parameters
qubits – one Qubit object, one list of Qubit objects, one Qureg object, or a tuple of the former three cases.
 get_merged(other)[source]
Return self merged with another TimeEvolution gate if possible.
 Two TimeEvolution gates are merged if:
both have the same terms
the proportionality factor for each of the terms must have relative error <= 1e9 compared to the proportionality factors of the other terms.
Note
While one could merge gates for which both hamiltonians commute, we are not doing this as in general the resulting gate would have to be decomposed again.
Note
We are not comparing if terms are proportional to each other with an absolute tolerance. It is up to the user to remove terms close to zero because we cannot choose a suitable absolute error which works for everyone. Use, e.g., a decomposition rule for that.
 Parameters
other – TimeEvolution gate
 Raises
NotMergeable – If the other gate is not a TimeEvolution gate or hamiltonians are not suitable for merging.
 Returns
New TimeEvolution gate equivalent to the two merged gates.
 class projectq.ops.UniformlyControlledRy(angles)[source]
Uniformly controlled Ry gate as introduced in arXiv:quantph/0312218.
This is an nqubit gate. There are n1 control qubits and one target qubit. This gate applies Ry(angles(k)) to the target qubit if the n1 control qubits are in the classical state k. As there are 2^(n1) classical states for the control qubits, this gate requires 2^(n1) (potentially different) angle parameters.
Example
controls = eng.allocate_qureg(2) target = eng.allocate_qubit() UniformlyControlledRy(angles=[0.1, 0.2, 0.3, 0.4])  (controls, target)
Note
The first quantum register contains the control qubits. When converting the classical state k of the control qubits to an integer, we define controls[0] to be the least significant (qu)bit. controls can also be an empty list in which case the gate corresponds to an Ry.
 Parameters
angles (list[float]) – Rotation angles. Ry(angles[k]) is applied conditioned on the control qubits being in state k.
 class projectq.ops.UniformlyControlledRz(angles)[source]
Uniformly controlled Rz gate as introduced in arXiv:quantph/0312218.
This is an nqubit gate. There are n1 control qubits and one target qubit. This gate applies Rz(angles(k)) to the target qubit if the n1 control qubits are in the classical state k. As there are 2^(n1) classical states for the control qubits, this gate requires 2^(n1) (potentially different) angle parameters.
Example
controls = eng.allocate_qureg(2) target = eng.allocate_qubit() UniformlyControlledRz(angles=[0.1, 0.2, 0.3, 0.4])  (controls, target)
Note
The first quantum register are the contains qubits. When converting the classical state k of the control qubits to an integer, we define controls[0] to be the least significant (qu)bit. controls can also be an empty list in which case the gate corresponds to an Rz.
 Parameters
angles (list[float]) – Rotation angles. Rz(angles[k]) is applied conditioned on the control qubits being in state k.
 projectq.ops.apply_command(cmd)[source]
Apply a command.
Extracts the qubitsowning (target) engine from the Command object and sends the Command to it.
 Parameters
cmd (Command) – Command to apply
setups
The setups package contains a collection of setups which can be loaded by the MainEngine. Each setup contains a get_engine_list function which returns a list of compiler engines:
 Example:
import projectq.setups.ibm as ibm_setup from projectq import MainEngine eng = MainEngine(engine_list=ibm_setup.get_engine_list()) # eng uses the default Simulator backend
The subpackage decompositions contains all the individual decomposition rules which can be given to, e.g., an AutoReplacer.
Subpackages
setups.decompositions
The decomposition package is a collection of gate decomposition / replacement rules which can be used by, e.g., the AutoReplacer engine.
Registers a decomposition for quantum amplitude amplification. 

Register the ZY decomposition for an arbitrary one qubit gate. 

Registers a decomposition rule for barriers. 

Register the decomposition of an controlled arbitary single qubit gate. 

Registers a decomposition to for a CNOT gate in terms of CZ and Hadamard. 

Register a decomposition to for a CNOT gate in terms of Rxx, Rx and Ry gates. 

Register a decomposition rule for multicontrolled gates. 

Register a decomposition to replace turn negatively controlled qubits into positively controlled qubits. 

Registers a decomposition for controlled zrotation gates. 

Registers a decomposition for the Entangle gate. 

Registers a decomposition rule for global phases. 

Register a decomposition for the H gate into an Ry and Rx gate. 

Registers a decomposition for the controlled global phase gate. 

Registers a decomposition for phase estimation. 

Registers a decomposition rule for the quantum Fourier transform. 

Register a decomposition rule for a unitary QubitOperator to one qubit gates. 

Registers a decomposition rule for the phaseshift gate. 

Register a decomposition for the Rx gate into an Rz gate and Hadamard. 

Register a decomposition for the Ry gate into an Rz and Rx(pi/2) gate. 

Registers a decomposition for the Rz gate into an Rx and Ry(pi/2) or Ry(pi/2) gate. 

Register a decomposition to achieve a SqrtSwap gate. 

Register decomposition for StatePreparation. 

Registers a decomposition to achieve a Swap gate. 

Register decomposition for the TimeEvolution gates. 

Registers a decomposition rule for the Toffoli gate. 

Register decomposition for UnformlyControlledRy and UnformlyControlledRz. 


Builtin mutable sequence. 
Submodules
Registers a decomposition for quantum amplitude amplification.
(Quick reference https://en.wikipedia.org/wiki/Amplitude_amplification. Complete reference G. Brassard, P. Hoyer, M. Mosca, A. Tapp (2000) Quantum Amplitude Amplification and Estimation https://arxiv.org/abs/quantph/0005055)
Quantum Amplitude Amplification (QAA) executes the algorithm, but not the final measurement required to obtain the marked state(s) with high probability. The starting state on wich the QAA algorithm is executed is the one resulting of aplying the Algorithm on the 0> state.
Example
def func_algorithm(eng, system_qubits):
All(H)  system_qubits
def func_oracle(eng, system_qubits, qaa_ancilla):
# This oracle selects the state 010> as the one marked
with Compute(eng):
All(X)  system_qubits[0::2]
with Control(eng, system_qubits):
X  qaa_ancilla
Uncompute(eng)
system_qubits = eng.allocate_qureg(3)
# Prepare the qaa_ancilla qubit in the > state
qaa_ancilla = eng.allocate_qubit()
X  qaa_ancilla
H  qaa_ancilla
# Creates the initial state form the Algorithm
func_algorithm(eng, system_qubits)
# Apply Quantum Amplitude Amplification the correct number of times
num_it = int(math.pi / 4.0 * math.sqrt(1 << 3))
with Loop(eng, num_it):
QAA(func_algorithm, func_oracle)  (system_qubits, qaa_ancilla)
All(Measure)  system_qubits
Warning
No qubit allocation/deallocation may take place during the call
to the defined Algorithm func_algorithm
 projectq.setups.decompositions.amplitudeamplification.func_algorithm
Algorithm that initialite the state and to be used in the QAA algorithm
 projectq.setups.decompositions.amplitudeamplification.func_oracle
The Oracle that marks the state(s) as “good”
 projectq.setups.decompositions.amplitudeamplification.system_qubits
the system we are interested on
 projectq.setups.decompositions.amplitudeamplification.qaa_ancilla
auxiliary qubit that helps to invert the amplitude of the “good” states
 projectq.setups.decompositions.amplitudeamplification.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register the ZY decomposition for an arbitrary one qubit gate.
See paper “Elementary gates for quantum computing” by Adriano Barenco et al., arXiv:quantph/9503016v1. (Note: They use different gate definitions!) Or see theorem 4.1 in Nielsen and Chuang.
Decompose an arbitrary one qubit gate U into U = e^(i alpha) Rz(beta) Ry(gamma) Rz(delta). If a gate V is element of SU(2), i.e., determinant == 1, then V = Rz(beta) Ry(gamma) Rz(delta)
 projectq.setups.decompositions.arb1qubit2rzandry.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition rule for barriers.
Deletes all barriers if they are not supported.
 projectq.setups.decompositions.barrier.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register the decomposition of an controlled arbitary single qubit gate.
See paper “Elementary gates for quantum computing” by Adriano Barenco et al., arXiv:quantph/9503016v1. (Note: They use different gate definitions!) or Nielsen and Chuang chapter 4.3.
 projectq.setups.decompositions.carb1qubit2cnotrzandry.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition to for a CNOT gate in terms of CZ and Hadamard.
 projectq.setups.decompositions.cnot2cz.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition to for a CNOT gate in terms of Rxx, Rx and Ry gates.
 projectq.setups.decompositions.cnot2rxx.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>, <projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition rule for multicontrolled gates.
Implements the decomposition of Nielsen and Chuang (Fig. 4.10) which decomposes a C^n(U) gate into a sequence of 2 * (n1) Toffoli gates and one C(U) gate by using (n1) ancilla qubits and circuit depth of 2n1.
 projectq.setups.decompositions.cnu2toffoliandcu.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition to replace turn negatively controlled qubits into positively controlled qubits.
This achived by applying X gates to selected qubits.
 projectq.setups.decompositions.controlstate.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition for controlled zrotation gates.
It uses 2 zrotations and 2 C^n NOT gates to achieve this gate.
 projectq.setups.decompositions.crz2cxandrz.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition for the Entangle gate.
Applies a Hadamard gate to the first qubit and then, conditioned on this first qubit, CNOT gates to all others.
 projectq.setups.decompositions.entangle.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition rule for global phases.
Deletes global phase gates (which can be ignored).
 projectq.setups.decompositions.globalphase.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition for the H gate into an Ry and Rx gate.
 projectq.setups.decompositions.h2rx.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>, <projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition for the controlled global phase gate.
Turns the controlled global phase gate into a (controlled) phaseshift gate. Each time this rule is applied, one control can be shaved off.
 projectq.setups.decompositions.ph2r.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition for phase estimation.
(reference https://en.wikipedia.org/wiki/Quantum_phase_estimation_algorithm)
The Quantum Phase Estimation (QPE) executes the algorithm up to the inverse QFT included. The following steps measuring the ancillas and computing the phase should be executed outside of the QPE.
The decomposition uses as ancillas (qpe_ancillas) the first qubit/qureg in the Command and as system qubits teh second qubit/qureg in the Command.
The unitary operator for which the phase estimation is estimated (unitary) is the gate in Command
Example
# Example using a ProjectQ gate
n_qpe_ancillas = 3
qpe_ancillas = eng.allocate_qureg(n_qpe_ancillas)
system_qubits = eng.allocate_qureg(1)
angle = cmath.pi * 2.0 * 0.125
U = Ph(angle) # unitary_specfic_to_the_problem()
# Apply Quantum Phase Estimation
QPE(U)  (qpe_ancillas, system_qubits)
All(Measure)  qpe_ancillas
# Compute the phase from the ancilla measurement
# (https://en.wikipedia.org/wiki/Quantum_phase_estimation_algorithm)
phasebinlist = [int(q) for q in qpe_ancillas]
phase_in_bin = ''.join(str(j) for j in phasebinlist)
phase_int = int(phase_in_bin, 2)
phase = phase_int / (2**n_qpe_ancillas)
print(phase)
# Example using a function (two_qubit_gate).
# Instead of applying QPE on a gate U one could provide a function
def two_qubit_gate(system_q, time):
CNOT  (system_q[0], system_q[1])
Ph(2.0 * cmath.pi * (time * 0.125))  system_q[1]
CNOT  (system_q[0], system_q[1])
n_qpe_ancillas = 3
qpe_ancillas = eng.allocate_qureg(n_qpe_ancillas)
system_qubits = eng.allocate_qureg(2)
X  system_qubits[0]
# Apply Quantum Phase Estimation
QPE(two_qubit_gate)  (qpe_ancillas, system_qubits)
All(Measure)  qpe_ancillas
# Compute the phase from the ancilla measurement
# (https://en.wikipedia.org/wiki/Quantum_phase_estimation_algorithm)
phasebinlist = [int(q) for q in qpe_ancillas]
phase_in_bin = ''.join(str(j) for j in phasebinlist)
phase_int = int(phase_in_bin, 2)
phase = phase_int / (2**n_qpe_ancillas)
print(phase)
 projectq.setups.decompositions.phaseestimation.unitary
Unitary Operation either a ProjectQ gate or a function f.
 Type
 Calling the function with the parameters system_qubits
 Type
Qureg) and time (integer
 i.e. f
 Type
system_qubits, time
 with parameter time.
 projectq.setups.decompositions.phaseestimation.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition rule for the quantum Fourier transform.
Decomposes the QFT gate into Hadamard and controlled phaseshift gates (R).
Warning
The final Swaps are not included, as those are simply a reindexing of quantum registers.
 projectq.setups.decompositions.qft2crandhadamard.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition rule for a unitary QubitOperator to one qubit gates.
 projectq.setups.decompositions.qubitop2onequbit.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition rule for the phaseshift gate.
Decomposes the (controlled) phaseshift gate using zrotation and a global phase gate.
 projectq.setups.decompositions.r2rzandph.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition for the Rx gate into an Rz gate and Hadamard.
 projectq.setups.decompositions.rx2rz.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition for the Ry gate into an Rz and Rx(pi/2) gate.
 projectq.setups.decompositions.ry2rz.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition for the Rz gate into an Rx and Ry(pi/2) or Ry(pi/2) gate.
 projectq.setups.decompositions.rz2rx.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>, <projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register a decomposition to achieve a SqrtSwap gate.
 projectq.setups.decompositions.sqrtswap2cnot.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register decomposition for StatePreparation.
 projectq.setups.decompositions.stateprep2cnot.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition to achieve a Swap gate.
Decomposes a Swap gate using 3 CNOT gates, where the one in the middle features as many control qubits as the Swap gate has control qubits.
 projectq.setups.decompositions.swap2cnot.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Register decomposition for the TimeEvolution gates.
An exact straight forward decomposition of a TimeEvolution gate is possible if the hamiltonian has only one term or if all the terms commute with each other in which case one can implement each term individually.
 projectq.setups.decompositions.time_evolution.all_defined_decomposition_rules = [<projectq.cengines._replacer._decomposition_rule.DecompositionRule object>, <projectq.cengines._replacer._decomposition_rule.DecompositionRule object>]
Decomposition rules
Registers a decomposition rule for the Toffoli gate.
Decomposes the Toffoli g