new version black

Author: refraction-ray

docs/source/tutorials/circuit_basics.ipynb

@@ -799,7 +799,7 @@
     }
    ],
    "source": [
-    "input_state = K.ones([2 ** n])\n",
+    "input_state = K.ones([2**n])\n",
     "input_state /= K.norm(input_state)\n",
     "\n",
     "c = tc.Circuit(n, inputs=input_state)\n",

docs/source/tutorials/circuit_basics_cn.ipynb

@@ -798,7 +798,7 @@
     }
    ],
    "source": [
-    "input_state = K.ones([2 ** n])\n",
+    "input_state = K.ones([2**n])\n",
     "input_state /= K.norm(input_state)\n",
     "\n",
     "c = tc.Circuit(n, inputs=input_state)\n",

docs/source/tutorials/operator_spreading.ipynb

@@ -121,7 +121,7 @@
     "def get_unitary_of_circuit(v):\n",
     "    layers = len(v)\n",
     "    c = tc.Circuit(\n",
-    "        N, inputs=np.eye(2 ** N, dtype=dtype)\n",
+    "        N, inputs=np.eye(2**N, dtype=dtype)\n",
     "    )  # when we choose inputs=np.eye(2**N,...), the output wavefunction\n",
     "    # is the unitary matrix U(t) of the quantum circuit\n",
     "\n",
@@ -138,7 +138,7 @@
     "                c.ry((i + 1) % N, theta=v[layer, (i + 1) % N])\n",
     "                c.cz(i, (i + 1) % N)\n",
     "    U = c.wavefunction()  # the output of wavefunction is a vector with dim=2^(2N)\n",
-    "    U = tf.reshape(U, [2 ** N, 2 ** N])  # reshape vector to matrix\n",
+    "    U = tf.reshape(U, [2**N, 2**N])  # reshape vector to matrix\n",
     "    return U"
    ]
   },
@@ -215,7 +215,7 @@
     "    σ_x_t = U_dagger @ σ_x @ U  # O(t)\n",
     "    σ_x_t_dagger = tf.transpose(σ_x_t, conjugate=True)\n",
     "    C = tf.linalg.trace(σ_y @ σ_x_t_dagger @ σ_y @ σ_x_t)\n",
-    "    C = 1 - tf.math.real(C) / (2 ** N)\n",
+    "    C = 1 - tf.math.real(C) / (2**N)\n",
     "    return C"
    ]
   },

docs/source/tutorials/operator_spreading_cn.ipynb

@@ -121,7 +121,7 @@
     "def get_unitary_of_circuit(v):\n",
     "    layers = len(v)\n",
     "    c = tc.Circuit(\n",
-    "        N, inputs=np.eye(2 ** N, dtype=dtype)\n",
+    "        N, inputs=np.eye(2**N, dtype=dtype)\n",
     "    )  # 当我们选择inputs=np.eye(2**N,...)时，输出波函数\n",
     "    # 是量子电路的幺正矩阵 U(t)\n",
     "\n",
@@ -138,7 +138,7 @@
     "                c.ry((i + 1) % N, theta=v[layer, (i + 1) % N])\n",
     "                c.cz(i, (i + 1) % N)\n",
     "    U = c.wavefunction()  # 波函数的输出是一个向量，它的 dim=2^(2N)\n",
-    "    U = tf.reshape(U, [2 ** N, 2 ** N])  # 将向量重塑为矩阵\n",
+    "    U = tf.reshape(U, [2**N, 2**N])  # 将向量重塑为矩阵\n",
     "    return U"
    ]
   },
@@ -215,7 +215,7 @@
     "    σ_x_t = U_dagger @ σ_x @ U  # O(t)\n",
     "    σ_x_t_dagger = tf.transpose(σ_x_t, conjugate=True)\n",
     "    C = tf.linalg.trace(σ_y @ σ_x_t_dagger @ σ_y @ σ_x_t)\n",
-    "    C = 1 - tf.math.real(C) / (2 ** N)\n",
+    "    C = 1 - tf.math.real(C) / (2**N)\n",
     "    return C"
    ]
   },

docs/source/tutorials/optimization_and_expressibility.ipynb

@@ -242,7 +242,7 @@
     "@tf.function\n",
     "def energy_loss(v, hamiltonian):\n",
     "    ψ = get_state(v)\n",
-    "    ψ = tf.reshape(ψ, [2 ** N, 1])\n",
+    "    ψ = tf.reshape(ψ, [2**N, 1])\n",
     "    loss = tf.transpose(ψ, conjugate=True) @ hamiltonian @ ψ\n",
     "    loss = tc.backend.real(loss)\n",
     "    return loss"

docs/source/tutorials/optimization_and_expressibility_cn.ipynb

@@ -241,7 +241,7 @@
     "@tf.function\n",
     "def energy_loss(v, hamiltonian):\n",
     "    ψ = get_state(v)\n",
-    "    ψ = tf.reshape(ψ, [2 ** N, 1])\n",
+    "    ψ = tf.reshape(ψ, [2**N, 1])\n",
     "    loss = tf.transpose(ψ, conjugate=True) @ hamiltonian @ ψ\n",
     "    loss = tc.backend.real(loss)\n",
     "    return loss"

docs/source/tutorials/qml_scenarios.ipynb

@@ -289,7 +289,7 @@
     "\n",
     "qml_layer = tc.keras.QuantumLayer(qml, weights_shape=[blocks, n, 3])\n",
     "\n",
-    "inputs = tf.keras.Input(shape=(2 ** n), dtype=tf.complex64)\n",
+    "inputs = tf.keras.Input(shape=(2**n), dtype=tf.complex64)\n",
     "measurements = qml_layer(inputs)\n",
     "output = tf.keras.layers.Dense(1, activation=\"sigmoid\")(measurements)\n",
     "model = tf.keras.Model(inputs=inputs, outputs=output)"

docs/source/tutorials/qml_scenarios_cn.ipynb

@@ -289,7 +289,7 @@
     "\n",
     "qml_layer = tc.keras.QuantumLayer(qml, weights_shape=[blocks, n, 3])\n",
     "\n",
-    "inputs = tf.keras.Input(shape=(2 ** n), dtype=tf.complex64)\n",
+    "inputs = tf.keras.Input(shape=(2**n), dtype=tf.complex64)\n",
     "measurements = qml_layer(inputs)\n",
     "output = tf.keras.layers.Dense(1, activation=\"sigmoid\")(measurements)\n",
     "model = tf.keras.Model(inputs=inputs, outputs=output)"

docs/source/tutorials/vqe_h2o.ipynb

@@ -430,7 +430,7 @@
    "outputs": [],
    "source": [
     "mb_tf = tc.backend.coo_sparse_matrix(\n",
-    "    np.transpose(np.stack([mb.row, mb.col])), mb.data, shape=(2 ** n, 2 ** n)\n",
+    "    np.transpose(np.stack([mb.row, mb.col])), mb.data, shape=(2**n, 2**n)\n",
     ")"
    ]
   },

docs/source/tutorials/vqe_h2o_cn.ipynb

@@ -430,7 +430,7 @@
    "outputs": [],
    "source": [
     "mb_tf = tc.backend.coo_sparse_matrix(\n",
-    "    np.transpose(np.stack([mb.row, mb.col])), mb.data, shape=(2 ** n, 2 ** n)\n",
+    "    np.transpose(np.stack([mb.row, mb.col])), mb.data, shape=(2**n, 2**n)\n",
     ")"
    ]
   },

docs/source/tutorials/vqex_mbl.ipynb

@@ -280,7 +280,7 @@
     "        tape.watch(v)\n",
     "        ρ = get_ρ_circuit(v, epochs)\n",
     "        hexp = tf.linalg.trace(ρ @ hamiltonian)\n",
-    "        loss = tf.linalg.trace(ρ @ hamiltonian @ hamiltonian) - hexp ** 2\n",
+    "        loss = tf.linalg.trace(ρ @ hamiltonian @ hamiltonian) - hexp**2\n",
     "        loss = tf.math.real(loss)\n",
     "    gr = tape.gradient(loss, v)\n",
     "    return loss, gr, ρ"

docs/source/tutorials/vqex_mbl_cn.ipynb

@@ -280,7 +280,7 @@
     "        tape.watch(v)\n",
     "        ρ = get_ρ_circuit(v, epochs)\n",
     "        hexp = tf.linalg.trace(ρ @ hamiltonian)\n",
-    "        loss = tf.linalg.trace(ρ @ hamiltonian @ hamiltonian) - hexp ** 2\n",
+    "        loss = tf.linalg.trace(ρ @ hamiltonian @ hamiltonian) - hexp**2\n",
     "        loss = tf.math.real(loss)\n",
     "    gr = tape.gradient(loss, v)\n",
     "    return loss, gr, ρ"

examples/checkpoint_memsave.py

@@ -97,7 +97,7 @@ def totallayer(s, param):
 
 
 def vqe_forward(param):
-    s = tc.backend.ones([2 ** nwires])
+    s = tc.backend.ones([2**nwires])
     s /= tc.backend.norm(s)
     s = totallayer(s, param)
     e = tc.expectation((tc.gates.x(), [1]), ket=s)

examples/variational_dynamics.py

@@ -96,7 +96,7 @@ def update(theta, lhs, rhs, tau):
     # TFIM Hamiltonian defined on lattice graph g (1D OBC chain)
     h = tc.array_to_tensor(h)
 
-    psi0 = np.zeros(2 ** N)
+    psi0 = np.zeros(2**N)
     psi0[0] = 1.0
     psi0 = tc.array_to_tensor(psi0)
 

examples/vqeh2o_benchmark.py

@@ -36,7 +36,7 @@
 mb = tc.quantum.PauliStringSum2COO_numpy(lsb, wb)
 mbd = mb.todense()
 mb = K.coo_sparse_matrix(
-    np.transpose(np.stack([mb.row, mb.col])), mb.data, shape=(2 ** n, 2 ** n)
+    np.transpose(np.stack([mb.row, mb.col])), mb.data, shape=(2**n, 2**n)
 )
 mbd = tc.array_to_tensor(mbd)
 

requirements-dev.txt

@@ -3,7 +3,7 @@ pytest==6.2.4
 pytest-cov
 pytest-benchmark
 pytest-xdist
-black==21.9b0
+black==22.3.0
 sphinx>=4.0
 pytest-lazy-fixture
 pylint==2.11.1

requirements-docker.txt

@@ -28,7 +28,7 @@ pytest-cov
 pytest-benchmark
 pytest-xdist
 pytest-lazy-fixture
-black==21.9b0
+black==22.3.0
 sphinx>=4.0
 sphinx-intl
 sphinx-copybutton

tensorcircuit/applications/utils.py

@@ -54,11 +54,11 @@ def amplitude_encoding(
     norm = tf.linalg.norm(fig, axis=1)
     norm = norm[..., tf.newaxis]
     fig = fig / norm
-    if fig.shape[1] < 2 ** qubits:
+    if fig.shape[1] < 2**qubits:
         fig = tf.concat(
             [
                 fig,
-                tf.zeros([fig.shape[0], 2 ** qubits - fig.shape[1]], dtype=tf.float64),
+                tf.zeros([fig.shape[0], 2**qubits - fig.shape[1]], dtype=tf.float64),
             ],
             axis=1,
         )
@@ -205,7 +205,7 @@ def train_qml_vag(
     with tf.GradientTape() as tape:
         tape.watch(nnp)
         cnnp = tf.cast(nnp, dtype=tf.complex64)
-        c = Circuit(nqubits, inputs=np.ones([1024], dtype=np.complex64) / 2 ** 5)
+        c = Circuit(nqubits, inputs=np.ones([1024], dtype=np.complex64) / 2**5)
         for epoch in range(epochs):
             for i in range(nqubits):
                 c.rz(i, theta=cnnp[3 * epoch, i])  # type: ignore
@@ -275,7 +275,7 @@ def validate_qml_vag(
 ) -> Any:
     xs, ys = gdata
     cnnp = tf.cast(nnp, dtype=tf.complex64)
-    c = Circuit(nqubits, inputs=np.ones([1024], dtype=np.complex64) / 2 ** 5)
+    c = Circuit(nqubits, inputs=np.ones([1024], dtype=np.complex64) / 2**5)
     for epoch in range(epochs):
         for i in range(nqubits):
             c.rz(i, theta=cnnp[3 * epoch, i])  # type: ignore
@@ -401,7 +401,7 @@ def TFIM1Denergy(
         Jzz *= 4
     for i in range(L):
         q = np.pi * (2 * i - (1 + (-1) ** L) / 2) / L
-        e -= np.abs(Jx) / 2 * np.sqrt(1 + Jzz ** 2 / 4 / Jx ** 2 - Jzz / Jx * np.cos(q))
+        e -= np.abs(Jx) / 2 * np.sqrt(1 + Jzz**2 / 4 / Jx**2 - Jzz / Jx * np.cos(q))
     return e
 
 

tensorcircuit/applications/vags.py

@@ -54,7 +54,7 @@ def GHZ_vag(
     gdata: Any, nnp: Tensor, preset: Sequence[int], verbose: bool = False, n: int = 3
 ) -> Tuple[Tensor, Tensor]:
     # gdata = None
-    reference_state = np.zeros([2 ** n])
+    reference_state = np.zeros([2**n])
     #     W states benchmarks
     #     for i in range(n):
     #         reference_state[2**(i)] = 1/np.sqrt(n)
@@ -80,7 +80,7 @@ def GHZ_vag(
         s = circuit.wavefunction()
         s = tf.reshape(
             s,
-            [2 ** n],
+            [2**n],
         )
         loss = tf.math.reduce_sum(
             tf.math.abs(s - reference_state)
@@ -1053,7 +1053,6 @@ def quantum_mp_qaoa_vag(
         gmatrix = tf.constant(gmatrix)
         return loss[0], gmatrix
 
-
 except NameError as e:
     logger.warning(e)
     logger.warning("tfq related vags disabled due to missing packages")

tensorcircuit/applications/van.py

@@ -43,7 +43,7 @@ def call(self, inputs):
         return tf.tensordot(inputs, w_mask, axes=[[-2, -1], [2, 3]]) + self.b
 
     def regularization(self, lbd_w=1.0, lbd_b=1.0):
-        return lbd_w * tf.reduce_sum(self.w ** 2) + lbd_b * tf.reduce_sum(self.b ** 2)
+        return lbd_w * tf.reduce_sum(self.w**2) + lbd_b * tf.reduce_sum(self.b**2)
 
 
 class MADE(tf.keras.Model):

tensorcircuit/applications/vqes.py

@@ -86,7 +86,7 @@ def construct_matrix_tf(ham: List[List[float]], dtype: Any = tf.complex128) -> T
 def construct_matrix_v2(ham: List[List[float]], dtype: Any = tf.complex128) -> Tensor:
     # deprecated
     s = len(ham[0]) - 1
-    h = tf.zeros([2 ** s, 2 ** s], dtype=dtype)
+    h = tf.zeros([2**s, 2**s], dtype=dtype)
     for term in tqdm(ham, desc="Hamiltonian building"):
         term = list(term)
         for i, t in enumerate(term):

tensorcircuit/densitymatrix.py

@@ -239,7 +239,7 @@ def densitymatrix(self, check: bool = False, reuse: bool = True) -> tn.Node.tens
         nodes, _ = self._copy_dm_tensor(conj=False, reuse=reuse)
         # t = contractor(nodes, output_edge_order=d_edges)
         dm = backend.reshape(
-            nodes[0].tensor, shape=[2 ** self._nqubits, 2 ** self._nqubits]
+            nodes[0].tensor, shape=[2**self._nqubits, 2**self._nqubits]
         )
         if check:
             self.check_density_matrix(dm)

tensorcircuit/mpscircuit.py

@@ -277,7 +277,7 @@ def apply_adjacent_double_gate(
             max_truncation_err=self.max_truncation_err,
             relative=self.relative,
         )
-        self._fidelity *= 1 - backend.real(backend.sum(err ** 2))
+        self._fidelity *= 1 - backend.real(backend.sum(err**2))
 
     def apply_double_gate(
         self,

tensorcircuit/quantum.py

@@ -1227,7 +1227,6 @@ def ps2coo_core(
         )
         return tf.SparseTensor(indices=indices, values=values, dense_shape=(s, s))  # type: ignore
 
-
 except NameError:
     logger.warning(
         "tensorflow is not installed, and sparse Hamiltonian generation utilities are disabled"
@@ -1654,7 +1653,7 @@ def spin_by_basis(n: int, m: int, elements: Tuple[int, int] = (1, -1)) -> Tensor
         backend.cast(
             backend.convert_to_tensor(np.array([[elements[0]], [elements[1]]])), "int32"
         ),
-        [2 ** m, int(2 ** (n - m - 1))],
+        [2**m, int(2 ** (n - m - 1))],
     )
     return backend.reshape(s, [-1])
 

tensorcircuit/templates/dataset.py

@@ -19,12 +19,12 @@ def amplitude_encoding(
     fig = backend.reshape(fig, shape=[-1])
     norm = backend.norm(fig)
     fig = fig / norm
-    if backend.shape_tuple(fig)[0] < 2 ** nqubits:
+    if backend.shape_tuple(fig)[0] < 2**nqubits:
         fig = backend.concat(
             [
                 fig,
                 backend.zeros(
-                    [2 ** nqubits - backend.shape_tuple(fig)[0]], dtype=fig.dtype
+                    [2**nqubits - backend.shape_tuple(fig)[0]], dtype=fig.dtype
                 ),
             ],
         )

tensorcircuit/translation.py

@@ -20,8 +20,8 @@
 
 
 def perm_matrix(n: int) -> Tensor:
-    p_mat = np.zeros([2 ** n, 2 ** n])
-    for i in range(2 ** n):
+    p_mat = np.zeros([2**n, 2**n])
+    for i in range(2**n):
         bit = i
         revs_i = 0
         for j in range(n):