Merge pull request PaddlePaddle#10 from shjNT/shj0

test_examples

tests/test_examples/pytest.ini

@@ -0,0 +1,4 @@
+[pytest]
+addopts = -p no:warnings
+log_cli = true
+#log_cli_level = DEBUG

tests/test_examples/run.sh

@@ -0,0 +1,38 @@
+#!/bin/bash
+
+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+# 
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+# 
+#     http://www.apache.org/licenses/LICENSE-2.0
+# 
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+cases=`find . -maxdepth 1 -name "test_*.py" | sort `
+ignore=""
+bug=0
+
+for file in ${cases}
+do
+echo ${file}
+if [[ ${ignore} =~ ${file##*/} ]]; then
+    echo "skip"
+else
+    python3.7 -m pytest ${file}
+    if [ $? -ne 0 ]; then
+        echo ${file} >> result.txt
+        bug=`expr ${bug} + 1`
+    fi
+fi
+done
+
+echo "total bugs: "${bug} > result.txt
+cat result.txt
+exit ${bug}

tests/test_examples/test_darcy2d.py

@@ -0,0 +1,120 @@
+"""
+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+
+import pytest
+import paddlescience as psci
+import numpy as np
+import paddle
+
+
+# Analytical solution
+def DarcyRecSolution(x, y):
+    """
+    Dirichlet boundary condition
+    """
+    return np.sin(2.0 * np.pi * x) * np.cos(2.0 * np.pi * y)
+
+
+# Generate analytical Solution using Geometry points
+def GenSolution(xy, bc_index):
+    """
+    GenSolution
+    """
+    sol = np.zeros((len(xy), 1)).astype(np.float32)
+    bc_value = np.zeros((len(bc_index), 1)).astype(np.float32)
+    length1 = len(xy)
+    for i in range(length1):
+        sol[i][0] = DarcyRecSolution(xy[i][0], xy[i][1])
+
+    length2 = len(bc_index)
+    for i in range(length2):
+        bc_value[i][0] = sol[bc_index[i]]
+
+    return [sol, bc_value]
+
+
+# right-hand side
+def RighthandBatch(xy):
+    """
+    RighthandBatch
+    """
+    return [
+        8.0 * 3.1415926 * 3.1415926 * paddle.sin(2.0 * np.pi * xy[:, 0]) *
+        paddle.cos(2.0 * np.pi * xy[:, 1])
+    ]
+
+
+# Geometry
+geo = psci.geometry.Rectangular(
+    space_origin=(0.0, 0.0), space_extent=(1.0, 1.0))
+
+# PDE Laplace
+pdes = psci.pde.Laplace2D()
+
+# Discretization
+pdes, geo = psci.discretize(pdes, geo, space_nsteps=(4, 4))
+
+# bc value
+golden, bc_value = GenSolution(geo.space_domain, geo.bc_index)
+pdes.set_bc_value(bc_value=bc_value)
+
+# Network
+net = psci.network.FCNet(
+    num_ins=2,
+    num_outs=1,
+    num_layers=2,
+    hidden_size=1,
+    dtype="float32",
+    activation="tanh")
+
+net._parameters["w_0"].set_value(
+    paddle.to_tensor([[1], [1]]).astype("float32"))
+net._parameters["w_1"].set_value(paddle.to_tensor([[1]]).astype("float32"))
+
+# Loss, TO rename
+loss = psci.loss.L2(pdes=pdes, geo=geo, aux_func=RighthandBatch)
+
+# Algorithm
+algo = psci.algorithm.PINNs(net=net, loss=loss)
+
+# Optimizer
+opt = psci.optimizer.Adam(learning_rate=0.001, parameters=net.parameters())
+
+# Solver
+solver = psci.solver.Solver(algo=algo, opt=opt)
+solution = solver.solve(num_epoch=10)
+
+# Use solution
+rslt = solution(geo)
+
+# Calculate diff and l2 relative error
+diff = rslt - golden
+root_square_error = np.linalg.norm(diff, ord=2)
+mean_square_error = root_square_error * root_square_error / geo.get_domain_size(
+)
+
+
+@pytest.mark.ldc2d
+def test_darcy2D():
+    """
+    test darcy2d
+    """
+    golden = np.load("./golden/darcy2D.npz")
+    expect = golden['expect']
+    diff_expect = golden['diff_expect']
+    assert np.allclose(rslt, expect), "the rslt was changed"
+    assert np.isclose(diff_expect,
+                      mean_square_error), "the mean_square_error was changed"

tests/test_examples/test_laplace2d.py

@@ -0,0 +1,110 @@
+"""
+# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+
+import pytest
+import paddlescience as psci
+import numpy as np
+import paddle
+
+
+# Analytical solution
+def LaplaceRecSolution(x, y, k=1.0):
+    """
+    LaplaceRecSolution
+    """
+    if k == 0.0:
+        return x * y
+    else:
+        return np.cos(k * x) * np.cosh(k * y)
+
+
+# Generate analytical Solution using Geometry points
+def GenSolution(xy, bc_index):
+    """
+    GenSolution
+    """
+    sol = np.zeros((len(xy), 1)).astype(np.float32)
+    bc_value = np.zeros((len(bc_index), 1)).astype(np.float32)
+    length1 = len(xy)
+    for i in range(length1):
+        sol[i] = LaplaceRecSolution(xy[i][0], xy[i][1])
+
+    length2 = len(bc_index)
+    for i in range(length2):
+        bc_value[i][0] = sol[bc_index[i]]
+    return [sol, bc_value]
+
+
+# Geometry
+geo = psci.geometry.Rectangular(
+    space_origin=(0.0, 0.0), space_extent=(1.0, 1.0))
+
+# PDE Laplace
+pdes = psci.pde.Laplace2D()
+
+# Discretization
+pdes, geo = psci.discretize(pdes, geo, space_nsteps=(4, 4))
+
+# bc value
+golden, bc_value = GenSolution(geo.space_domain, geo.bc_index)
+pdes.set_bc_value(bc_value=bc_value)
+
+# Network
+net = psci.network.FCNet(
+    num_ins=2,
+    num_outs=1,
+    num_layers=2,
+    hidden_size=1,
+    dtype="float32",
+    activation="tanh")
+net._parameters["w_0"].set_value(
+    paddle.to_tensor([[1], [1]]).astype("float32"))
+net._parameters["w_1"].set_value(paddle.to_tensor([[1]]).astype("float32"))
+
+# Loss, TO rename
+loss = psci.loss.L2(pdes=pdes, geo=geo)
+
+# Algorithm
+algo = psci.algorithm.PINNs(net=net, loss=loss)
+
+# Optimizer
+opt = psci.optimizer.Adam(learning_rate=0.001, parameters=net.parameters())
+
+# Solver
+solver = psci.solver.Solver(algo=algo, opt=opt)
+solution = solver.solve(num_epoch=30)
+
+# Use solution
+rslt = solution(geo)
+
+# Calculate diff and l2 relative error
+diff = rslt - golden
+root_square_error = np.linalg.norm(diff, ord=2)
+mean_square_error = root_square_error * root_square_error / geo.get_domain_size(
+)
+
+
+@pytest.mark.laplace2d
+def test_Laplace2D():
+    """
+    test Laplace2D
+    """
+    golden = np.load("./golden/laplace2d.npz")
+    expect = golden['expect']
+    diff_expect = golden['diff_expect']
+    assert np.allclose(rslt, expect), "the rslt was changed"
+    assert np.isclose(diff_expect,
+                      mean_square_error), "the mean_square_error was changed"