"""
\********************************************************************************
* Copyright (c) 2023 the Qrisp authors
*
* This program and the accompanying materials are made available under the
* terms of the Eclipse Public License 2.0 which is available at
* http://www.eclipse.org/legal/epl-2.0.
*
* This Source Code may also be made available under the following Secondary
* Licenses when the conditions for such availability set forth in the Eclipse
* Public License, v. 2.0 are satisfied: GNU General Public License, version 2
* with the GNU Classpath Exception which is
* available at https://www.gnu.org/software/classpath/license.html.
*
* SPDX-License-Identifier: EPL-2.0 OR GPL-2.0 WITH Classpath-exception-2.0
********************************************************************************/
"""
import traceback
import numpy as np
import sympy
def bin_rep(n, bits):
if n < 0:
raise Exception("Only positive numbers are supported")
if n >= 2**bits:
raise Exception(
str(n) + " can't be represented as a " + str(bits) + " bit number"
)
return bin(n)[2:].zfill(bits)
zero_string = "".join(["0" for k in range(bits)])
return (zero_string + bin(n)[2:])[-bits:]
def int_encoder(qv, encoding_number):
if encoding_number > 2 ** len(qv) - 1:
raise Exception("Not enough qubits to encode integer " + str(encoding_number))
binary_rep = bin_rep(encoding_number, len(qv))[::-1]
from qrisp import x
for i in range(len(binary_rep)):
if int(binary_rep[i]):
x(qv[i])
# Calculates the binary expression of a given integer and returns it as an array of
# length bits
def int_as_array(k, bit):
bin_str = bin_rep(k, bit)
return np.array([int(c) for c in bin_str])
def array_as_int(array):
result = 0
for k in range(len(array)):
if array[::-1][k]:
result += 2 ** (k)
return result
# Decomposes the circuit qc until no more decompositions are possible and then counts
# the cnot operations
def cnot_count(qc):
qc = qc.transpile()
gate_count_dic = qc.count_ops()
cnot_count = 0
for gate_name in ["cx", "cy", "cz"]:
cnot_count += gate_count_dic.get(gate_name, 0)
return cnot_count
def is_inv(x, bit):
# return (math.gcd(int(np.round(x, 3)),2**bit) == 1)
# The only divisors 2**bit has is powers of 2
# ie. if tha factorization of x doesn't contain any powers of 2 it is invertible
# in other words: x is invertible if it is uneven
return bool(int(x) % 2)
def get_depth_dic(qc, transpile_qc=True, depth_indicator=lambda x: 1):
if len(qc.qubits) == 0:
return {}
if transpile_qc:
qc = qc.transpile()
# Assign each bit in the circuit a unique integer
# to index into op_stack.
bit_indices = {bit: idx for idx, bit in enumerate(qc.qubits + qc.clbits)}
# If no bits, return 0
if not bit_indices:
return 0
# A list that holds the height of each qubit
# and classical bit.
op_stack = [0] * len(bit_indices)
# Here we are playing a modified version of
# Tetris where we stack gates, but multi-qubit
# gates, or measurements have a block for each
# qubit or cbit that are connected by a virtual
# line so that they all stacked at the same depth.
# Conditional gates act on all cbits in the register
# they are conditioned on.
# We treat barriers or snapshots different as
# They are transpiler and simulator directives.
# The max stack height is the circuit depth.
for instr in qc.data:
if instr.op.name in ["qb_alloc", "qb_dealloc", "gphase"]:
continue
qargs = instr.qubits
cargs = instr.clbits
levels = []
reg_ints = []
# If count then add one to stack heights
gate_depth = depth_indicator(instr.op)
for ind, reg in enumerate(qargs + cargs):
# Add to the stacks of the qubits and
# cbits used in the gate.
reg_ints.append(bit_indices[reg])
levels.append(op_stack[reg_ints[ind]] + gate_depth)
max_level = max(levels)
for ind in reg_ints:
op_stack[ind] = max_level
return {qc.qubits[i]: op_stack[i] for i in range(len(qc.qubits))}
[docs]
def gate_wrap(*args, permeability=None, is_qfree=None, name=None, verify=False):
"""
Decorator to bundle up the quantum instructions of a function into a single gate
object. Bundled gate objects can help debugging as it allows for a more clear
QuantumCircuit visualisation.
Furthermore, bundling up functions is relevant for Qrisps uncomputation algorithm.
When bundling up for uncomputation, this decorator provides the means to annotate
the gate objects with information about its permeability and qfree-ness. For further
information about these concepts check the
:ref:`uncomputation documentation<uncomputation>`. Specifying this information
allows to skip the computationally costly automatic determination at runtime.
Note that the specified information is not checked for correctness as this would
defy the purpose.
A shorthand for ``gate_wrap(permeability = "args", is_qfree = True)`` is the
:meth:`lifted <qrisp.lifted>` decorator.
.. warning::
Using ``gate_wrap`` without specifying permeability and ``qfree``-ness on
functions processing a lot of qubits, can causes long compile times, since the
unitaries of these gates have to be determined numerically.
.. warning::
Incorrect information about permeability and ``qfree``-ness can yield incorrect
compilation results. If you are unsure, use the ``verify`` keyword on a small
scale first.
Parameters
----------
permeability : string or list, optional
Specify the permeability behavior of the function. When given "args", it is
assumed that the gate is permeable only on the qubits of the arguments. When
given "full", it is assumed that the gate is permeable on every qubit it acts on
(i.e. also the result). When given a list of integers it is assumed, that the
gate is permeable on the qubits of the arguments corresponding to the integers.
The default is None.
is_qfree : bool, optional
Specify the qfree-ness of the function. The default is None.
name : string, optional
String which will be used for naming the gate object. The default is None.
verify : bool, optional
If set to ``True``, the specified information about permeability and
``qfree``-ness will be checked numerically. The default is ``False``.
Examples
--------
We create a simple function wrapping up multiple gates: ::
from qrisp import QuantumVariable, cx, x, h, z, gate_wrap
@gate_wrap
def example_function(a, b):
cx(a,b)
x(a)
cx(b,a)
h(b)
a = QuantumVariable(3)
b = QuantumVariable(3)
example_function(a, b)
>>> print(a.qs)
::
QuantumCircuit:
--------------
┌───────────────────┐
b.0: ┤0 ├
│ │
b.1: ┤1 ├
│ │
b.2: ┤2 ├
│ example_function │
a.0: ┤3 ├
│ │
a.1: ┤4 ├
│ │
a.2: ┤5 ├
└───────────────────┘
Live QuantumVariables:
---------------------
QuantumVariable a
QuantumVariable b
>>> print(a.qs.transpile())
::
┌───┐ ┌───┐
b.0: ┤ X ├─────────────────■──┤ H ├──────────
└─┬─┘┌───┐ │ └───┘┌───┐
b.1: ──┼──┤ X ├────────────┼────■──┤ H ├─────
│ └─┬─┘┌───┐ │ │ └───┘┌───┐
b.2: ──┼────┼──┤ X ├───────┼────┼────■──┤ H ├
│ │ └─┬─┘┌───┐┌─┴─┐ │ │ └───┘
a.0: ──■────┼────┼──┤ X ├┤ X ├──┼────┼───────
│ │ ├───┤└───┘┌─┴─┐ │
a.1: ───────■────┼──┤ X ├─────┤ X ├──┼───────
│ ├───┤ └───┘┌─┴─┐
a.2: ────────────■──┤ X ├──────────┤ X ├─────
└───┘ └───┘
In the next example, we create a function that performs no quantum gates and specify
that it is qfree and permeable on the second argument but not on the first. ::
from qrisp import QuantumCircuit
@gate_wrap(permeability = [1], is_qfree = True)
def example_function(arg_0, arg_1):
res = QuantumVariable(1)
#Append an identity gate
res.qs.append(QuantumCircuit(3).to_gate(), [arg_0, arg_1, res])
return res
qv_0 = QuantumVariable(1)
qv_1 = QuantumVariable(1)
res = example_function(qv_0, qv_1)
>>> print(qv_0.qs)
::
QuantumCircuit:
--------------
┌───────────────────┐
qv_0.0: ┤0 ├
│ │
qv_1.0: ┤1 example_function ├
│ │
res.0: ┤2 ├
└───────────────────┘
Live QuantumVariables:
---------------------
QuantumVariable qv_0
QuantumVariable qv_1
QuantumVariable res
>>> qv_1.uncompute()
>>> print(qv_0.qs)
::
QuantumCircuit:
--------------
┌───────────────────┐
qv_0.0: ┤0 ├
│ │
qv_1.0: ┤1 example_function ├
│ │
res.0: ┤2 ├
└───────────────────┘
Live QuantumVariables:
---------------------
QuantumVariable qv_0
QuantumVariable res
Since ``arg_1`` is marked as permeable, there are no further gates required for
uncomputation. The situation is different for the other two QuantumVariables, where
#the qubits are not marked as permeable.
>>> qv_0.uncompute(do_it = False)
>>> res.uncompute()
>>> print(qv_0.qs)
::
QuantumCircuit:
--------------
┌───────────────────┐┌──────────────────────┐
qv_0.0: ┤0 ├┤0 ├
│ ││ │
qv_1.0: ┤1 example_function ├┤1 example_function_dg ├
│ ││ │
res.0: ┤2 ├┤2 ├
└───────────────────┘└──────────────────────┘
Live QuantumVariables:
---------------------
"""
if len(args):
return gate_wrap_inner(args[0])
else:
def gate_wrap_helper(function):
return gate_wrap_inner(
function,
permeability=permeability,
is_qfree=is_qfree,
name=name,
verify=verify,
)
return gate_wrap_helper
def gate_wrap_inner(
function, permeability=None, is_qfree=None, name=None, verify=False
):
def wrapped_function(
*args, permeability=permeability, is_qfree=is_qfree, verify=verify, **kwargs
):
wrapped_function.__name__ = function.__name__
from qrisp.circuit import Qubit
from qrisp.core import recursive_qs_search, recursive_qv_search
from qrisp.environments import GateWrapEnvironment
from qrisp import merge, QuantumVariable, QuantumArray
try:
qs = find_qs(args)
except:
qs_list = recursive_qs_search([args, kwargs])
qs = qs_list[0]
initial_qubits = set(qs.qubits)
if name is None:
gwe = GateWrapEnvironment(name=function.__name__)
else:
gwe = GateWrapEnvironment(name=name)
with gwe:
result = function(*args, **kwargs)
if len(qs.env_stack):
gwe.compile()
if gwe in qs.data:
qs.data.remove(gwe)
if gwe.instruction is None:
return result
created_qubits = set(qs.qubits) - initial_qubits
ancillas = []
for qb in created_qubits:
if qb.allocated is False:
ancillas.append(qb)
if is_qfree is not None:
if verify and is_qfree:
from qrisp.permeability import is_qfree as is_qfree_function
if not is_qfree_function(gwe.instruction.op):
raise Exception(
f"Verification of qfree-ness for function {function.__name__} "
f"failed"
)
gwe.instruction.op.is_qfree = is_qfree
if permeability is not None:
permeability_dict = {i: None for i in range(gwe.instruction.op.num_qubits)}
permeable_qubits = []
not_permeable_qubits = []
if isinstance(permeability, list):
for i in range(len(args)):
if i in permeability:
extension_list = permeable_qubits
else:
extension_list = not_permeable_qubits
arg = args[i]
if isinstance(arg, QuantumVariable):
extension_list += arg.reg
elif isinstance(arg, (tuple, list)):
for item in arg:
if isinstance(item, Qubit):
extension_list.append(item)
elif isinstance(item, QuantumVariable):
extension_list += item.reg
elif isinstance(arg, QuantumArray):
for qv in arg.flatten():
extension_list += qv.reg
if isinstance(result, QuantumVariable):
not_permeable_qubits += result.reg
elif isinstance(result, (tuple, list)):
for item in result:
if isinstance(item, Qubit):
not_permeable_qubits.append(item)
elif isinstance(item, QuantumVariable):
not_permeable_qubits += item.reg
elif isinstance(result, QuantumArray):
for qv in result.flatten():
not_permeable_qubits += qv.reg
elif isinstance(permeability, str):
for arg in args:
if isinstance(arg, QuantumVariable):
permeable_qubits += arg.reg
elif isinstance(arg, (tuple, list)):
for item in arg:
if isinstance(item, Qubit):
permeable_qubits.append(item)
elif isinstance(item, QuantumVariable):
permeable_qubits += item.reg
elif isinstance(arg, QuantumArray):
for qv in arg.flatten():
permeable_qubits += qv.reg
if permeability == "full":
extension_list = permeable_qubits
elif permeability == "args":
extension_list = not_permeable_qubits
else:
raise Exception(f"Don't know permeability option {permeability}")
if isinstance(result, QuantumVariable):
extension_list += result.reg
elif isinstance(result, (tuple, list)):
for item in result:
if isinstance(item, Qubit):
extension_list.append(item)
elif isinstance(item, QuantumVariable):
extension_list += item.reg
elif isinstance(result, QuantumArray):
for qv in result.flatten():
extension_list += qv.reg
for i in range(len(gwe.instruction.qubits)):
qb = gwe.instruction.qubits[i]
if qb in permeable_qubits:
permeability_dict[i] = True
elif qb in not_permeable_qubits:
permeability_dict[i] = False
elif qb in ancillas:
# Even though ancilla qubits are permeable, we want to be able to
# use the gate_wrap decorator as an interface to perform
# recomputation. If we mark them as permeable, Unqomp won't wrap
# the uncomputed gate in alloc/dealloc gates but instead wrap both
# the computation gate and the recomputation in alloc/dealloc gates
# To undestand this behavior better consider the following example
# from qrisp import QuantumFloat
# a = QuantumFloat(2)
# qf_res = a * a
# qf_res.uncompute()
# print(qf_res.qs)
# If we mark the ancilla qubits as permeable, this gives
# QuantumCircuit:
# ---------------
# ┌──────────┐┌──────────┐┌─────────────┐
# a.0: ┤ qb_alloc ├┤0 ├┤0 ├──────────────
# ├──────────┤│ ││ │
# a.1: ┤ qb_alloc ├┤1 ├┤1 ├──────────────
# ├──────────┤│ ││ │┌────────────┐
# return.0: ┤ qb_alloc ├┤2 ├┤2 ├┤ qb_dealloc ├
# ├──────────┤│ ││ │├────────────┤
# return.1: ┤ qb_alloc ├┤3 __mul__ ├┤3 __mul___dg ├┤ qb_dealloc ├
# ├──────────┤│ ││ │├────────────┤
# return.2: ┤ qb_alloc ├┤4 ├┤4 ├┤ qb_dealloc ├
# ├──────────┤│ ││ │├────────────┤
# return.3: ┤ qb_alloc ├┤5 ├┤5 ├┤ qb_dealloc ├
# ├──────────┤│ ││ │├────────────┤
# sbp_anc_0.0: ┤ qb_alloc ├┤6 ├┤6 ├┤ qb_dealloc ├
# └──────────┘└──────────┘└─────────────┘└────────────┘
# Live QuantumVariables:
# ----------------------
# QuantumFloat a
# If we set them to non-permeable, we get instead
# QuantumCircuit:
# ---------------
# ┌──────────┐┌──────────┐ ┌─────────────┐»
# a.0: ┤ qb_alloc ├┤0 ├──────────────────────────┤0 ├»
# ├──────────┤│ │ │ │»
# a.1: ┤ qb_alloc ├┤1 ├──────────────────────────┤1 ├»
# ├──────────┤│ │ │ │»
# return.0: ┤ qb_alloc ├┤2 ├──────────────────────────┤2 ├»
# ├──────────┤│ │ │ │»
# return.1: ┤ qb_alloc ├┤3 __mul__ ├──────────────────────────┤3 __mul___dg ├»
# ├──────────┤│ │ │ │»
# return.2: ┤ qb_alloc ├┤4 ├──────────────────────────┤4 ├»
# ├──────────┤│ │ │ │»
# return.3: ┤ qb_alloc ├┤5 ├──────────────────────────┤5 ├»
# ├──────────┤│ │┌────────────┐┌──────────┐│ │»
# sbp_anc_0.0: ┤ qb_alloc ├┤6 ├┤ qb_dealloc ├┤ qb_alloc ├┤6 ├»
# └──────────┘└──────────┘└────────────┘└──────────┘└─────────────┘»
# «
# « a.0: ──────────────
# «
# « a.1: ──────────────
# « ┌────────────┐
# « return.0: ┤ qb_dealloc ├
# « ├────────────┤
# « return.1: ┤ qb_dealloc ├
# « ├────────────┤
# « return.2: ┤ qb_dealloc ├
# « ├────────────┤
# « return.3: ┤ qb_dealloc ├
# « ├────────────┤
# «sbp_anc_0.0: ┤ qb_dealloc ├
# « └────────────┘
# Live QuantumVariables:
# ----------------------
# QuantumFloat a
# In both cases, sbp_anc is recomputed, but in the second case, the
# reallocation allows the compiler to use the free qubit elsewhere
# and afterwards pick a potentially different qubit for the
# recomputation.
permeability_dict[i] = False
# for i in range(len(gwe.instruction.qubits)):
# if permeability_dict[i] is None:
# permeability_dict[i] = False
if verify:
from qrisp.permeability import is_permeable
permeable_qubit_indices = []
for i in range(gwe.instruction.op.num_qubits):
if permeability_dict[i]:
permeable_qubit_indices.append(i)
if not is_permeable(gwe.instruction.op, permeable_qubit_indices):
raise Exception(
f"Verification of permeability for function "
f"{function.__name__} failed"
)
gwe.instruction.op.permeability = permeability_dict
return result
return wrapped_function
def find_qs(args):
if hasattr(args, "qs"):
return args.qs()
from qrisp import QuantumVariable, QuantumArray, Qubit
for arg in args:
if isinstance(arg, (QuantumVariable, QuantumArray)):
return arg.qs
if isinstance(arg, Qubit):
return arg.qs()
else:
for arg in args:
if isinstance(arg, (list, tuple)):
try:
return find_qs(arg)
except:
pass
if isinstance(arg, dict):
try:
return find_qs(arg.items())
except:
pass
raise Exception("Couldn't find QuantumSession")
# Function to measure multiple quantum variables at once to assess their entanglement
[docs]
def multi_measurement(qv_list, shots=None, backend=None):
"""
This functions facilitates the measurement of multiple QuantumVariables at the same
time. This can be used if the entanglement structure between several
QuantumVariables is of interest.
Parameters
----------
qv_list : list[QuantumVariable]
A list of QuantumVariables.
shots : int, optional
The amount of shots to perform. The default is given by the backend used.
backend : BackendClient, optional
The backend to evaluate the compiled QuantumCircuit on. By default, the backend
from default_backend.py will be used.
Raises
------
Exception
Tried to perform measurement with open environments.
Returns
-------
counts_list : list
A list of tuples. The first element of each tuple is a tuple again and contains
the labels of the QuantumVariables. The second element is a float and indicates
the probability of measurement.
Examples
--------
We entangle three QuantumFloats via addition and perform a multi-measurement:
>>> from qrisp import QuantumFloat, h, multi_measurement
>>> qf_0 = QuantumFloat(4)
>>> qf_1 = QuantumFloat(4)
>>> qf_0[:] = 3
>>> qf_1[:] = 2
>>> h(qf_1[0])
>>> qf_sum = qf_0 + qf_1
>>> multi_measurement([qf_0, qf_1, qf_sum])
{(3, 2, 5): 0.5, (3, 3, 6): 0.5}
"""
if backend is None:
if qv_list[0].qs.backend is None:
from qrisp.default_backend import def_backend
backend = def_backend
else:
backend = qv_list[0].qs.backend
if len(qv_list[0].qs.env_stack) != 0:
raise Exception("Tried to perform measurement with open environments")
from qrisp import merge
merge(qv_list)
# Copy circuit in order to prevent modification
from qrisp import QuantumArray, QuantumVariable, recursive_qv_search
from qrisp.core.compilation import qompiler
temp = recursive_qv_search(qv_list)
compiled_qc = qompiler(
qv_list[0].qs, intended_measurements=sum([qv.reg for qv in temp], [])
)
# compiled_qc = qv_list[0].qs.copy()
# Add classical registers for the measurement results to be stored in
cl_reg_list = []
for var in qv_list[::-1]:
cl_reg = []
if isinstance(var, QuantumArray):
qubits = sum([qv.reg for qv in var.flatten()[::-1]], [])
elif isinstance(var, QuantumVariable):
qubits = var.reg
else:
raise Exception(f"Found type {type(var)} in measurement list")
for i in range(len(qubits)):
cl_reg.append(compiled_qc.add_clbit())
cl_reg_list.append(cl_reg)
# Add measurement instruction
compiled_qc.measure(qubits, cl_reg)
# counts = execute(qs_temp, backend, basis_gates = basis_gates,
# noise_model = noise_model, shots = shots).result().get_counts()
counts = backend.run(compiled_qc, shots)
counts = {k: counts[k] for k in sorted(counts)}
shots = sum(counts.values())
# Convert the labeling bistrings of counts into list of labels
new_counts = {}
for i in range(len(counts)):
# Retrieve the separated strings of each measurement variable
counts_strings = []
counts_bitstring = list(counts.keys())[i]
bitstring_adress = 0
for j in range(len(cl_reg_list)):
cl_reg = cl_reg_list[::-1][j]
counts_strings.append(
counts_bitstring[bitstring_adress : bitstring_adress + len(cl_reg)][
::-1
]
)
bitstring_adress += len(cl_reg)
# Convert to integers and insert outcome labels
counts_values = []
for j in range(len(counts_strings)):
outcome_int = int(counts_strings[j][::-1], 2)
try:
label = qv_list[j].decoder(outcome_int)
if isinstance(label, np.ndarray):
from qrisp import OutcomeArray
label = OutcomeArray(label)
counts_values.append(label)
except AttributeError:
counts_values.append(outcome_int)
# Create array
array_state = tuple(counts_values)
try:
no_of_shots_executed = sum(counts.values())
new_counts[array_state] = counts[list(counts.keys())[i]] / no_of_shots_executed
except TypeError:
raise Exception(
"Tried to create measurement outcome dic for QuantumVariable "
"with unhashable labels"
)
# Append to the counts list
# counts_list.append((array_state, counts[list(counts.keys())[i]]/shots))
# Sort counts_list such the most probable values come first
new_counts = dict(sorted(new_counts.items(), key=lambda item: -item[1]))
return new_counts
# Function to apply a phase function of signature phase_function(x,y,z..) -> float
# which specifies the phase for each constellation of outcome labels of the quantum
# variables in qv_list.
def app_phase_function(qv_list, phase_function, t=1, **kwargs):
# Prepare the list of index tuples
# For this we first create a list of outcome indices of each qv first
index_lists = [list(range(2**qv.size)) for qv in qv_list]
# We now calculate the direct product in order to obtain every possible combination
from itertools import product
product_index_list = list(product(*index_lists))
# The next step is to iterate over every combination in order to determine the
# phases.
phases = []
for i in range(len(product_index_list)):
# Calculate the outcome labels of the current constellation of indices
labels = [
qv_list[j].decoder(product_index_list[i][j]) for j in range(len(qv_list))
]
# Calculate the phase
phases.append(phase_function(*labels, **kwargs) * t)
# Synthesize phase
from qrisp import gray_phase_synth_qb_list
gray_phase_synth_qb_list(
qv_list[0].qs, sum([qv.reg[::-1] for qv in qv_list], []), phases
)
[docs]
def as_hamiltonian(hamiltonian):
r"""
Decorator that recieves a regular Python function (returning a float) and returns a
function of QuantumVariables, applying phases based on the function's output.
Parameters
----------
hamiltonian : function
A function of arbitrary (non-quantum) variables returning a float.
Returns
-------
hamiltonian_application : function
A function of QuantumVariables, which applies the phase dictated by the
hamiltonian to the corresponding states.
Examples
--------
In this example we will demonstrate how a phase function with multiple arguments can
be synthesized. For this we will create a phase function which encodes the fourier
transform of different integers on the QuantumFloat x conditioned on the value of a
QuantumChar ch. We will then apply the inverse Fourier transform to x and measure
the results.
::
import numpy as np
from qrisp import QuantumChar, QuantumFloat, QFT, h
#Create Variables
x_size = 3
x = QuantumFloat(x_size, 0, signed = False)
ch = QuantumChar()
# Bring x into uniform superposition so the phase function application yields
# a fourier transformed computation basis state
h(x)
#Bring ch into partial superposition (here |a> + |b> + |c> + |d>)
h(ch[0])
h(ch[1])
from qrisp import multi_measurement, as_hamiltonian
#In order to define the hamiltonian, we use regular Python syntax
#The decorator "as_hamiltonian" turns it into a function
#that takes Quantum Variables as arguments. The decorator will add the
#keyword argument t to the function which mimics the t in exp(i*H*t)
@as_hamiltonian
def apply_multi_var_hamiltonian(ch, x):
if ch == "a":
k = 2
elif ch == "b":
k = 7
elif ch == "c":
k = 3
else:
k = 4
#Return phase value
#This is the phase distribution of the Fourier-transform
#of the computational basis state |k>
return k*x * 2*np.pi/2**x_size
#Apply Hamiltonian
apply_multi_var_hamiltonian(ch,x, t = 1)
#Apply inverse Fourier transform
QFT(x, inv = True)
Acquire measurement results
>>> multi_measurement([ch, x])
{('a', 2): 0.25, ('b', 7): 0.25, ('c', 3): 0.25, ('d', 4): 0.25}
We see that the measurement results correspond to what we specified in the
Hamiltonian.
In Bra-Ket notation, before applying the Hamiltonian, we are in the state
.. math::
\ket{\psi} = \frac{1}{\sqrt{4}}(\ket{a} + \ket{b} + \ket{c} + \ket{d})
\left( \frac{1}{\sqrt{8}} \sum_{x = 0}^8 \ket{x} \right)
We then apply the Hamiltonian:
.. math::
\text{exp}(i\text{H(ch, x)})\ket{\psi} = \frac{1}{\sqrt{32}}
( &\ket{a} \sum_{x = 0}^{2^3-1} \text{exp}(2x \frac{2 \pi i}{2^3}) \ket{x} \\
+&\ket{b} \sum_{x = 0}^{2^3-1} \text{exp}(7x \frac{2 \pi i}{2^3}) \ket{x} \\
+&\ket{c} \sum_{x = 0}^{2^3-1} \text{exp}(3x \frac{2 \pi i}{2^3}) \ket{x} \\
+&\ket{d} \sum_{x = 0}^{2^3-1} \text{exp}(4x \frac{2 \pi i}{2^3}) \ket{x})
For each branch, the QuantumFloat tensor-factor is in a Fourier-transformed
computational basis state. Thus, if we apply the inverse QFT, we receive:
.. math::
&\text{QFT}^{-1}\text{exp}(i\text{H(ch, x)})\ket{\psi} \\
= & \frac{1}{\sqrt{4}} (\ket{a} \ket{2} + \ket{b} \ket{7} +\ket{c} \ket{3}
+ \ket{d} \ket{4})
"""
def hamiltonian_application(*args, t=1, **kwargs):
gate_wrap(app_phase_function)(args, hamiltonian, t=t, **kwargs)
# app_phase_function(args, hamiltonian, t = t, **kwargs)
return hamiltonian_application
[docs]
def perm_lock(qubits):
"""
Locks a list of qubits such that only permeable gates can be executed on these
qubits. This means that an error will be raised if the user attempts to perform any
operation involving these qubits if the operation does not commute with the
Z-operator of this qubit. For more information, what a permeable gate is, check the
:ref:`uncomputation documentation <uncomputation>`.
This can be helpfull as it forbids all operations that change that computational
basis state of this qubit but still allow controling on this qubit or applying
phase gates.
The effect of this function can be reversed using perm_unlock.
Parameters
----------
qubits : list[Qubit] or QuantumVariable
The qubits to phase-tolerantly lock.
Examples
--------
We create a QuantumChar, perm-lock it's Qubits and attempt to initialize.
>>> from qrisp import QuantumChar, perm_lock, cx, p
>>> q_ch_0 = QuantumChar()
>>> perm_lock(q_ch_0)
>>> q_ch_0[:] = "g"
Exception: Tried to perform non-permeable operations on perm_locked qubits
We now create a second QuantumChar and perform a CNOT gate
>>> q_ch_1 = QuantumChar()
>>> cx(q_ch_0[3], q_ch_1[2])
Phase-gates are possible, too
>>> p(0.1, q_ch_0)
"""
from qrisp.circuit.quantum_circuit import convert_to_qb_list
for qb in convert_to_qb_list(qubits):
if isinstance(qb, list):
for item in qb:
perm_lock(item)
continue
qb.perm_lock = True
[docs]
def perm_unlock(qubits):
"""
Reverses the effect of "perm_lock".
Parameters
----------
qubits : list[Qubit] or QuantumVariable
The qubits to phase-tolerantly unlock.
Examples
--------
We create a QuantumChar, perm-lock it's Qubits and attempt to initialize.
>>> from qrisp import QuantumChar, perm_lock, perm_unlock
>>> q_ch = QuantumChar()
>>> perm_lock(q_ch)
>>> q_ch[:] = "g"
Exception: Tried to perform non-permeable operations on perm_locked qubits
>>> perm_unlock(q_ch)
>>> q_ch[:] = "g"
>>> print(q_ch)
{'g': 1.0}
"""
from qrisp.circuit.quantum_circuit import convert_to_qb_list
for qb in convert_to_qb_list(qubits):
if isinstance(qb, list):
for item in qb:
perm_unlock(item)
continue
qb.perm_lock = False
[docs]
def lock(qubits):
"""
Locks a list of qubits, implying an error will be raised if the user tries to
perform any operation involving these qubits.
This can be reversed by calling unlock.
Parameters
----------
qubits : list[Qubit] or QuantumVariable
The list of Qubits to lock.
Examples
--------
We create a QuantumChar, lock it's Qubits and attempt to initialize.
>>> from qrisp import QuantumChar, lock
>>> q_ch = QuantumChar()
>>> lock(q_ch)
>>> q_ch[:] = "g"
Exception: Tried to operation on locked qubits
"""
from qrisp.circuit.quantum_circuit import convert_to_qb_list
for qb in convert_to_qb_list(qubits):
if isinstance(qb, list):
for item in qb:
lock(item)
continue
qb.lock = True
[docs]
def unlock(qubits):
"""
Reverses the effect of "lock".
Parameters
----------
qubits : list[Qubit] or QuantumVariable
The list of Qubits to lock.
Examples
--------
We create a QuantumChar, lock it's Qubits and attempt to initialize.
>>> from qrisp import QuantumChar, lock, unlock
>>> q_ch = QuantumChar()
>>> lock(q_ch)
>>> q_ch[:] = "g"
Exception: Tried to perform operations on locked qubits
We now unlock and try again
>>> unlock(q_ch)
>>> q_ch[:] = "g"
>>> print(q_ch)
{'g': 1.0}
"""
from qrisp.circuit.quantum_circuit import convert_to_qb_list
for qb in convert_to_qb_list(qubits):
if isinstance(qb, list):
for item in qb:
unlock(item)
continue
qb.lock = False
def benchmark_function(function):
def benchmarked_function(*args, sort_stats="tottime", stat_amount=20, **kwargs):
def slow_function():
function(*args, **kwargs)
import cProfile
import pstats
profile = cProfile.Profile()
profile.runcall(slow_function)
ps = pstats.Stats(profile)
ps.strip_dirs()
ps.sort_stats(sort_stats)
ps.print_stats(stat_amount)
return benchmarked_function
def custom_qv(labels, decoder=None, qs=None, name=None):
if not isinstance(labels, list):
raise Exception(
"Tried to create custom QuantumVariable without providing a list type"
)
if len(labels) == 0:
raise Exception(
"Tried to create custom QuantumVariable without providing labels"
)
elif len(labels) == 1:
n = 1
else:
n = int(np.ceil(np.log2(len(labels))))
from qrisp import QuantumVariable
class CustomQuantumVariable(QuantumVariable):
def __init__(self, qs=None, name=None):
super().__init__(n, qs=qs, name=name)
def decoder(self, x):
if decoder is None:
if x < len(labels):
return labels[x]
else:
return "undefined_label_" + str(x)
return decoder(x)
return CustomQuantumVariable(qs=qs, name=name)
def init_state(qv, target_array):
from qiskit.circuit.library.data_preparation.state_preparation import (
StatePreparation,
)
qiskit_qc = StatePreparation(target_array).definition
from qrisp import QuantumCircuit
init_qc = QuantumCircuit.from_qiskit(qiskit_qc)
# Find global phase correction
from qrisp.simulator import statevector_sim
init_qc.qubits.reverse()
sim_array = statevector_sim(init_qc)
init_qc.qubits.reverse()
arg_max = np.argmax(np.abs(sim_array))
gphase_dif = (np.angle(target_array[arg_max] / sim_array[arg_max])) % (2 * np.pi)
init_qc.gphase(gphase_dif, 0)
init_gate = init_qc.to_gate()
init_gate.name = "state_init"
qv.qs.append(init_gate, qv)
def get_statevector_function(qs, decimals=None):
if len(qs.qv_list) == 0:
return lambda x: 0
else:
from qrisp.simulator import statevector_sim
compiled_qc = qs.compile()
sv_array = statevector_sim(compiled_qc)
if decimals is not None:
sv_array = np.round(sv_array, decimals)
def statevector(label_constellation, round=None):
from qrisp import bin_rep
qs = list(label_constellation.keys())[0].qs
if len(label_constellation) != len(qs.qv_list):
missing_variables = set([qv.name for qv in qs.qv_list]) - set(
[qv.name for qv in label_constellation.keys()]
)
raise Exception(
"Tried to invoke statevector debugger without specifying an "
"outcome label for each QuantumVariable registered in "
"QuantumSession. Missing variables are: " + str(missing_variables)
)
bitstring = len(compiled_qc.qubits) * ["0"]
for qf in label_constellation.keys():
label_int = qf.encoder(label_constellation[qf])
bin_label_int = bin_rep(label_int, qf.size)[::-1]
for i in range(qf.size):
qubit_pos = compiled_qc.qubits.index(qf[i])
bitstring[qubit_pos] = bin_label_int[i]
bitstring = "".join(bitstring)
state_index = int(bitstring, base=2)
if round is None:
return sv_array[state_index]
else:
return np.around(sv_array[state_index], round)
return statevector
def check_if_fresh(qubits, qs, ignore_q_envs=True):
from qrisp import QuantumEnvironment
if not ignore_q_envs:
temp_data = list(qs.data)
qs.data = []
for i in range(len(temp_data)):
if isinstance(temp_data[i], QuantumEnvironment):
env = temp_data[i]
env.compile()
else:
qs.append(temp_data[i])
for qb in qubits:
reversed_data = qb.qs().data[::-1]
for instr in reversed_data:
if isinstance(instr, QuantumEnvironment) and ignore_q_envs:
continue
if qb in instr.qubits:
if instr.op.name == "qb_alloc":
break
else:
return False
else:
return False
return True
def get_measurement_from_qc(qc, qubits, backend, shots=None):
# Add classical registers for the measurement results to be stored in
cl = []
for i in range(len(qubits)):
cl.append(qc.add_clbit())
# Add measurement instruction
for i in range(len(qubits)):
qc.measure(qubits[i], cl[i])
# Execute circuit
counts = backend.run(qc, shots)
# Remove other measurements outcomes from counts dic
new_counts_dic = {}
no_of_shots_executed = 0
for key in counts.keys():
# Remove possible whitespaces
new_key = key.replace(" ", "")
# Remove other measurements
new_key = new_key[: len(cl)]
new_key = int(new_key, base=2)
try:
new_counts_dic[new_key] += counts[key]
except KeyError:
new_counts_dic[new_key] = counts[key]
no_of_shots_executed += counts[key]
counts = new_counts_dic
# Plot result (if needed)
# Normalize counts
for key in counts.keys():
counts[key] = counts[key] / abs(no_of_shots_executed)
return counts
def find_calling_line(level=0):
stack = traceback.extract_stack(limit=level + 3)
return str(
traceback.format_list(stack)[1].split("\n")[1].strip()
) # prints "a = fct1()"
def retarget_instructions(data, source_qubits, target_qubits):
from qrisp import QuantumEnvironment, recursive_qs_search, multi_session_merge
for i in range(len(data)):
instr = data[i]
if isinstance(instr, QuantumEnvironment):
retarget_instructions(instr.original_data, source_qubits, target_qubits)
retarget_instructions(instr.env_data, source_qubits, target_qubits)
continue
for j in range(len(instr.qubits)):
if instr.qubits[j] in source_qubits:
instr.qubits[j] = target_qubits[source_qubits.index(instr.qubits[j])]
[docs]
def redirect_qfunction(function_to_redirect):
"""
Decorator to turn a function returning a QuantumVariable into an in-place function.
This can be helpful for manual uncomputation if we have a function returning some
QuantumVariable, but we want the result to operate on some other variable, which is
supposed to be uncomputed.
Parameters
----------
function_to_redirect : function
A function returning a QuantumVariable.
Raises
------
Exception
Given function did not return a QuantumVariable
Exception
Tried to redirect quantum function into QuantumVariable of differing size
Returns
-------
redirected_function : function
A function which performs the same operation as the input but now has the
keyword argument target. Every instruction that would have been executed on the
input functions result is executed on the QuantumVariable specified by target
instead.
Examples
--------
We create a function that determins the AND value of its inputs and redirect it
onto another QuantumBool. ::
from qrisp import QuantumBool, mcx, redirect_qfunction
#This function has only two arguments and returns its result
def AND(a, b):
res = QuantumBool()
mcx([a,b], res)
return res
a = QuantumBool(name = "a")
b = QuantumBool(name = "b")
c = QuantumBool(name = "c")
#This function has two arguments and the keyword argument target
redirected_AND = redirect_qfunction(AND)
redirected_AND(a, b, target = c)
>>> print(a.qs)
::
QuantumCircuit:
--------------
b.0: ──■──
│
a.0: ──■──
┌─┴─┐
c.0: ┤ X ├
└───┘
Live QuantumVariables:
---------------------
QuantumBool b
QuantumBool a
QuantumBool c
"""
from qrisp import QuantumEnvironment, QuantumVariable, merge, QuantumArray
import weakref
def redirected_qfunction(*args, target=None, **kwargs):
merge(
[
arg
for arg in list(args) + [target]
if isinstance(arg, (QuantumVariable, QuantumArray))
]
)
env = QuantumEnvironment()
env.manual_allocation_management = True
qs = target.qs
with env:
res = function_to_redirect(*args, **kwargs)
if not isinstance(res, QuantumVariable):
raise Exception("Given function did not return a QuantumVariable")
target = list(target)
if len(res) != len(target):
raise Exception(
"Tried to redirect quantum function into QuantumVariable of "
"differing size"
)
i = 0
res_is_new = False
while i < len(env.env_qs.data):
instr = env.env_qs.data[i]
if isinstance(instr, QuantumEnvironment):
pass
elif instr.op.name == "qb_alloc" and instr.qubits[0] in list(res):
env.env_qs.data.pop(i)
res_is_new = True
continue
else:
for qb in instr.qubits:
qb.qs = weakref.ref(qs)
i += 1
retarget_instructions(env.env_qs.data, list(res), target)
if res_is_new:
# Remove all traces of res
res.delete()
for i in range(res.size):
res.qs.qubits.remove(res[i])
res.qs.data.pop(-1)
for i in range(len(res.qs.deleted_qv_list)):
qv = res.qs.deleted_qv_list[i]
if qv.name == res.name:
res.qs.deleted_qv_list.pop(i)
break
return target
redirected_qfunction.__name__ = function_to_redirect.__name__
return redirected_qfunction
def get_sympy_state(qs, decimals):
from sympy import (
I,
Rational,
Symbol,
cancel,
cos,
count_ops,
exp,
factor,
nsimplify,
pi,
simplify,
sin,
)
from sympy.physics.quantum import Ket, OrthogonalKet
from qrisp.simulator import statevector_sim
qv_list = list(qs.qv_list)
labels = []
for qv in qv_list:
labels.append([qv.decoder(i) for i in range(2**qv.size)])
compiled_qc = qs.compile()
sv_array = statevector_sim(compiled_qc)
if not sv_array.dtype == np.dtype("O"):
angles = np.angle(sv_array) % (2 * np.pi) / (np.pi)
if decimals is not None:
sv_array = np.round(sv_array, decimals)
angles = np.round(angles, decimals)
else:
sv_array = np.round(sv_array, 5)
angles = np.round(angles, 5)
nz_indices = np.nonzero(sv_array)[0]
nnz = len(nz_indices)
else:
import sympy as sp
nz_indices = []
for i in range(len(sv_array)):
entry = simplify(sv_array[i])
for a in sp.preorder_traversal(entry):
if isinstance(a, sp.Float):
entry = entry.subs(a, round(a, 5))
sv_array[i] = entry
if not sv_array[i] == 0:
nz_indices.append(i)
nnz = len(nz_indices)
res = 0
for ind in list(nz_indices):
amplitude = sv_array[ind]
if not sv_array.dtype == np.dtype("O"):
if decimals is None:
try:
abs_amp = trigify_amp(amplitude, nnz)
except TypeError:
abs_amp = amplitude
# For some reason there is a sympy error, when the angle is equal to 1
if angles[ind] == 1:
phase = 1
else:
phase = nsimplify(float(angles[ind]), tolerance=10**-5)
if count_ops(phase) > 5:
phase = angles[ind]
ket_expr = exp(I * phase * pi) * abs_amp * nnz**0.5
else:
ket_expr = sympy.N(amplitude, decimals)
else:
process_stack = [amplitude]
while process_stack:
a = process_stack.pop(0)
if (
isinstance(a, (sympy.core.add.Add, sympy.core.mul.Mul))
and len(a.free_symbols) != 0
):
process_stack.extend(a.args)
elif len(a.free_symbols) == 0:
sub_float = np.round(complex(a.evalf()), 5)
if np.abs(sub_float - 1) < 10**-5:
abs_amp = 1
continue
elif np.abs(sub_float) < 10**-5:
entry = entry.subs(a, 0)
continue
elif np.abs(sub_float) > 1:
continue
else:
abs_amp = trigify_amp(sub_float, nnz)
if np.angle(complex(a.evalf())) / np.pi == 1:
phase = -1
else:
phase = sp.exp(
sp.I
* nsimplify(
np.angle(complex(a.evalf())) / np.pi,
tolerance=10**-5,
)
* Symbol("pi")
)
expr = abs_amp * phase
amplitude = amplitude.subs(a, expr)
amplitude = amplitude.subs(1j, sp.I)
ket_expr = sp.trigsimp(amplitude) * nnz**0.5
int_string = bin_rep(ind, len(compiled_qc.qubits))
labels = []
for qv in qv_list:
bit_string = ""
for qb in qv.reg:
bit_string += int_string[compiled_qc.qubits.index(qb)]
label = qv.decoder(int(bit_string[::-1], 2))
ket_expr *= OrthogonalKet((label))
res += ket_expr
if decimals is None or sv_array.dtype == np.dtype("O"):
res = cancel(nsimplify(1 / nnz**0.5) * res)
if isinstance(res, sympy.core.mul.Mul):
temp = 1
for arg in res.args[:-1]:
temp *= nsimplify(arg.subs({Symbol("pi"): pi}))
res = temp * res.args[-1]
res = res.subs({Symbol("pi"): pi})
return res
def trigify_amp(amplitude, nnz):
from sympy import (
I,
Rational,
Symbol,
cancel,
cos,
count_ops,
exp,
factor,
latex,
nsimplify,
pi,
simplify,
sin,
)
cos_expr = nsimplify(
float(np.arccos(np.abs(amplitude)) / np.pi), tolerance=10**-5
)
sin_expr = nsimplify(
float(np.arcsin(np.abs(amplitude)) / np.pi), tolerance=10**-5
)
# if count_ops(sin_expr) > count_ops(cos_expr):
if len(latex(sin_expr)) > len(latex(cos_expr)):
expr = "cos"
temp = cos_expr
# elif count_ops(sin_expr) < count_ops(cos_expr):
elif len(latex(sin_expr)) < len(latex(cos_expr)):
expr = "sin"
temp = sin_expr
elif len(sin_expr.free_symbols) == 0:
if sin_expr.evalf() > cos_expr.evalf():
expr = "cos"
temp = cos_expr
else:
expr = "sin"
temp = sin_expr
else:
temp = (
nsimplify(np.abs(amplitude) * nnz**0.5, tolerance=10**-5) / nnz**0.5
)
# if count_ops(temp) > 4:
if len(latex(temp)) > 20:
temp = (
nsimplify(float(np.abs(amplitude) * nnz**0.5), tolerance=10**-5)
/ nnz**0.5
)
if len(latex(temp)) > 20:
abs = np.abs(amplitude)
else:
abs = temp
else:
if expr == "cos":
abs = cos(cos_expr * Symbol("pi"))
else:
abs = sin(sin_expr * Symbol("pi"))
return abs
def render_qc(qc):
latex_str = qc.to_latex()
import os.path
import subprocess
import tempfile
from IPython.display import Image, display
with tempfile.TemporaryDirectory(prefix="texinpy_") as tmpdir:
path = os.path.join(tmpdir, "document.tex")
with open(path, "w") as fp:
fp.write(latex_str)
subprocess.run(["lualatex", path], cwd=tmpdir)
subprocess.run(
[
"pdftocairo",
"-singlefile",
"-transp",
"-r",
"100",
"-png",
"document.pdf",
"document",
],
cwd=tmpdir,
)
im = Image(filename=os.path.join(tmpdir, "document.png"))
display(im)
[docs]
def lifted(*args, verify=False):
"""
Shorthand for ``gate_wrap(permability = "args", is_qfree = True)``.
A lifted function is ``qfree`` and permeable on its inputs. The results of lifted
functions can be automatically uncomputed even if they contain functions that could
not be uncomputed on their own.
You can find more information about these concepts :ref:`here <Uncomputation>` or
`here <https://silq.ethz.ch/overview#/overview/3_uncomputation>`_. Note that the
concept of permeability in Qrisp is a more general version of Silq's ``const``.
.. warning::
Incorrect information about permeability and ``qfree``-ness can yield incorrect
compilation results. If you are unsure, use the ``verify`` keyword on a small
scale first.
Parameters
----------
verify : bool, optional
If set to ``True``, the specified information about permeability and
``qfree``-ness will be checked numerically. The default is ``False``.
Examples
--------
We create a function performing the `Margolus gate
<https://arxiv.org/abs/quant-ph/0312225>`_. As it contains ``ry`` rotations,
there are non-``qfree`` steps involved. Putting on the ``lifted`` decorator however
marks the function as ``qfree`` as a whole.
::
from qrisp import QuantumVariable, cx, ry, lifted
from numpy import pi
@lifted(verify = True)
def margolus(control):
res = QuantumVariable(1)
ry(pi/4, res)
cx(control[1], res)
ry(-pi/4, res)
cx(control[0], res)
ry(pi/4, res)
cx(control[1], res)
ry(-pi/4, res)
return res
control = QuantumVariable(2)
res = margolus(control)
>>> print(res.qs)
::
QuantumCircuit:
--------------
┌───────────┐
control.0: ┤0 ├
│ │
control.1: ┤1 margolus ├
│ │
res.0: ┤2 ├
└───────────┘
Live QuantumVariables:
---------------------
QuantumVariable control
QuantumVariable res
>>> res.uncompute()
>>> print(res.qs)
::
QuantumCircuit:
--------------
┌───────────┐┌──────────────┐
control.0: ┤0 ├┤0 ├
│ ││ │
control.1: ┤1 margolus ├┤1 margolus_dg ├
│ ││ │
res.0: ┤2 ├┤2 ├
└───────────┘└──────────────┘
Live QuantumVariables:
---------------------
QuantumVariable control
Note that we set the ``verify`` keyword to ``True`` in this example. In more complex
functions, involving many qubits this feature should only be used for bug-fixing on
a small scale, since the verification can be time-consuming.
"""
if len(args) == 0:
def lifted_helper(function):
return gate_wrap(permeability="args", is_qfree=True, verify=verify)(
function
)
return lifted_helper
else:
return gate_wrap(permeability="args", is_qfree=True)(args[0])
[docs]
def t_depth_indicator(op, epsilon):
r"""
This function returns the T-depth of an :ref:`Operation` object.
According to `this paper <https://arxiv.org/abs/1403.2975>`_, the synthesis of an $RZ(\phi)$
up to precision $\epsilon$ requires $3\text{log}_2(\frac{1}{\epsilon})$
T-gates.
Parameters
----------
op : :ref:`Operation`
The operation, whose T-depth should be estimated.
epsilon : float
The precision of the RZ gate simulation.
Returns
-------
float
The estimated T-depth of the Operation.
"""
from qrisp import ClControlledOperation
if isinstance(op, ClControlledOperation):
return t_depth_indicator(op.base_op, epsilon)
elif op.definition is not None:
return op.definition.t_depth(epsilon)
elif op.name in [
"cx",
"cx",
"cz",
"x",
"y",
"z",
"s",
"h",
"s_dg",
"sx",
"sx_dg",
"measure",
"reset",
"qb_alloc",
"qb_dealloc",
"barrier",
"gphase",
]:
return 0
elif op.name in ["rx", "ry", "rz", "p", "u1"]:
par = op.params[0] / (np.pi) % 1
if par in [0, 1 / 2]:
return 0
elif par in [1 / 4, 3 / 4]:
return 1
else:
return 3 * np.log2(1 / epsilon)
elif op.name in ["t", "t_dg"]:
return 1
elif op.name == "u3":
res = 0
for i in range(3):
par = op.params[0] / (np.pi) % 1
if par in [0, 1 / 2]:
pass
elif par in [1 / 4, 3 / 4]:
res += 1
else:
res += 3 * np.log2(1 / epsilon)
return res
else:
raise Exception(f"Gate {op.name} not implemented")
[docs]
def cnot_depth_indicator(op):
r"""
This function returns the CNOT-depth of an :ref:`Operation` object.
In NISQ-era devices, CNOT gates are the restricting bottleneck for quantum
circuit execution. This function can be used as a gate-speed specifier for
the :meth:`compile <qrisp.QuantumSession.compile>` method.
Parameters
----------
op : :ref:`Operation`
The operation, whose CNOT-depth should be computed.
Returns
-------
float
The CNOT-depth of the Operation.
"""
from qrisp import ClControlledOperation
if isinstance(op, ClControlledOperation):
return cnot_depth_indicator(op.base_op)
elif op.definition is not None:
return op.definition.cnot_depth()
if op.num_qubits == 1 or op.name == "barrier":
return 0
elif op.name in ["cx", "cx", "cz"]:
return 1
else:
raise Exception(f"Gate {op.name} not implemented")
[docs]
def inpl_adder_test(inpl_adder):
"""
This function runs tests on a desired inplace addition function.
An inplace addition function is a function mapping (a, b) to (a, a+b),
where a is a :ref:`QuantumVariable`, list[:ref:`Qubit`] or an integer
and b is either a :ref:`QuantumVariable` or a list[:ref:`Qubit`].
Parameters
----------
inpl_adder : callable
A quantum inplace addition function that can either act on single QuantumVariables or on lists of Qubits
by adding the first one to the second.
Returns
-------
Bool:
True if all tests are passed, else False/ Exceptions.
Examples
--------
We test the built-in Cuccaro adder:
::
from qrisp import cuccaro_adder, inpl_adder_test
inpl_adder_test(cuccaro_adder)
print("The cuccaro adder passed the tests without errors.")
And now a new user-defined qcla adder:
::
from qrisp import inpl_adder_test, qcla
qcla_2_0 = lambda x, y : qcla(x, y, radix_base = 2, radix_exponent = 0)
inpl_adder_test(qcla_2_0)
print("The qcla_2_0 adder passed the tests without errors.")
"""
from qrisp import QuantumFloat, multi_measurement, h, control, QuantumBool
for i in range(1, 7):
for j in range(1, i + 1):
a = QuantumFloat(j)
b = QuantumFloat(i)
c = QuantumFloat(i)
h(a)
h(b)
c[:] = b
inpl_adder(a, c)
statevector_arr = a.qs.compile().statevector_array()
angles = np.angle(
statevector_arr[
np.abs(statevector_arr) > 1 / 2 ** ((a.size + b.size) / 2 + 1)
]
)
# Test correct phase behavior
assert (
np.sum(np.abs(angles)) < 0.1
), f"Quantum-quantum adder produced a faulty phase shift on input sizes, {i},{j}."
mes_res = multi_measurement([a, b, c])
for a, b, c in mes_res.keys():
assert (a + b) % (
2**i
) == c, f"Quantum-quantum addition result was incorrect for input values {a} += {c} on input sizes, {i},{j}."
if i < 6:
for j in range(1, 2**i):
a = QuantumFloat(i)
b = QuantumFloat(i)
h(a)
b[:] = a
inpl_adder(j, a)
statevector_arr = a.qs.compile().statevector_array()
angles = np.angle(
statevector_arr[
np.abs(statevector_arr) > 1 / 2 ** ((a.size) / 2 + 1)
]
)
assert (
np.sum(np.abs(angles)) < 0.1
), f"Classical-quantum adder produced a faulty phase shift on input size {i}."
mes_res = multi_measurement([a, b])
for a, b in mes_res.keys():
assert (b + j) % (
2**i
) == a, f"Classical-quantum addition result was incorrect for input values {a} += {c} on input size {i}."
for i in range(1, 7):
for j in range(1, i + 1):
a = QuantumFloat(j)
b = QuantumFloat(i)
c = QuantumFloat(i)
qbl = QuantumBool()
h(qbl)
h(a)
h(b)
c[:] = b
with control(qbl):
inpl_adder(a, c)
statevector_arr = a.qs.compile().statevector_array()
angles = np.angle(
statevector_arr[
np.abs(statevector_arr) > 1 / 2 ** ((a.size + b.size) / 2 + 1)
]
)
assert (
np.sum(np.abs(angles)) < 0.1
), f"Controlled quantum-quantum adder produced a faulty phase shift on input sizes, {i},{j}."
mes_res = multi_measurement([a, b, c, qbl])
for a, b, c, qbl in mes_res.keys():
if qbl:
assert (a + b) % (
2**i
) == c, f"Controlled quantum-quantum addition result was incorrect for input values {a} += {c} on input sizes, {i},{j}."
else:
assert (
c == b
), f"Controlled quantum-quantum addition behaviour was incorrect; an operation was performed without the control qubit in |1> state.Faulty input sizes: {i},{j}"
if i < 6:
for j in range(1, 2**i):
a = QuantumFloat(i)
b = QuantumFloat(i)
qbl = QuantumBool()
h(qbl)
h(a)
b[:] = a
with control(qbl):
inpl_adder(j, a)
statevector_arr = a.qs.compile().statevector_array()
angles = np.angle(
statevector_arr[
np.abs(statevector_arr) > 1 / 2 ** ((a.size) / 2 + 1)
]
)
assert (
np.sum(np.abs(angles)) < 0.1
), f"Controlled classical-quantum adder produced a faulty phase shift on input size {i}."
mes_res = multi_measurement([a, b, qbl])
for a, b, qbl in mes_res.keys():
if qbl:
assert (b + j) % (
2**i
) == a, f"Controlled classical-quantum addition result was incorrect for input values {b} += {j} on input size, {i}."
else:
assert (
b == a
), f"Controlled classical-quantum addition behaviour was incorrect; an operation was performed without the control qubit in |1> state. Faulty input sizes: {i}"
