/*******************************************************************************
* Copyright 2014 Analog Devices, Inc.
*
* 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.
********************************************************************************/
package com.analog.lyric.dimple.test.FactorFunctions;
import java.util.Arrays;
import java.util.Random;
import org.junit.Test;
import com.analog.lyric.dimple.factorfunctions.MatrixProduct;
import com.analog.lyric.dimple.model.domains.Domain;
import com.analog.lyric.dimple.model.domains.RealDomain;
import com.analog.lyric.dimple.model.values.Value;
public class TestMatrixProduct extends FactorFunctionTester
{
private final Random _rand = new Random(123123);
@Test
public void test()
{
final int minSize = 2, maxSize = 4;
for (int nr = minSize; nr <= maxSize; ++nr)
for (int nx = minSize; nx <= maxSize; ++nx)
for (int nc = minSize; nc <= maxSize; ++nc)
testDeterministic(nr, nx, nc);
}
private void testDeterministic(final int nr, final int nx, final int nc)
{
final int nCases = 5;
final int outLength = nr * nc;
final int in1Row = nr;
final int in1Col = nx;
final int in2Row = nx;
final int in2Col = nc;
final int in1Length = in1Row * in1Col;
final int in2Length = in2Row * in2Col;
final MatrixProduct function = new MatrixProduct(nr, nx, nc);
final int[] outputIndices = new int[outLength];
for (int i = 0; i < outLength; ++i)
{
outputIndices[i] = i;
}
final double[][][] in1Matrices = new double[nCases][][];
final double[][][] in2Matrices = new double[nCases][][];
final double[][][] outMatrices = new double[nCases][][];
for (int i = 0; i < nCases; ++i)
{
double[][] in1Matrix = in1Matrices[i] = new double[in1Row][];
for (int row = 0; row < in1Row; ++row)
{
double[] rowValues = in1Matrix[row] = new double[in1Col];
for (int col = 0; col < in1Col; ++col)
{
rowValues[col] = 1 + (_rand.nextDouble() * 10);
}
}
double[][] in2Matrix = in2Matrices[i] = new double[in2Row][];
for (int row = 0; row < in2Row; ++row)
{
double[] rowValues = in2Matrix[row] = new double[in2Col];
for (int col = 0; col < in2Col; ++col)
{
rowValues[col] = 1 + (_rand.nextDouble() * 10);
}
}
double[][] outMatrix = outMatrices[i] = new double[nr][];
for (int row = 0; row < nr; ++row)
{
outMatrix[row] = new double[nc];
for (int col = 0; col < nc; ++col)
{
double out = 0.0;
for (int x = 0; x < nx; ++x)
out += in1Matrix[row][x] * in2Matrix[x][col];
outMatrix[row][col] = out;
}
}
}
final RealDomain rd = RealDomain.unbounded();
Domain[] domains = null;
Value[][] testCases = new Value[nCases][];
//
// Constant matrix inputs
//
domains = new Domain[2 + outLength];
Arrays.fill(domains, rd);
domains[outLength] = null; // matrix
domains[outLength + 1] = null;
for (int i = 0; i < nCases; ++i)
{
double[][] in1Matrix = in1Matrices[i];
double[][] in2Matrix = in2Matrices[i];
double[][] outMatrix = outMatrices[i];
Value[] testCase = testCases[i] = new Value[2 + outLength];
int j = 0;
for (int col = 0; col < nc; ++col)
for (int row = 0; row < nr; ++row)
testCase[j++] = Value.create(outMatrix[row][col]);
testCase[outLength] = Value.create(in1Matrix);
testCase[outLength + 1] = Value.create(in2Matrix);
}
testEvalDeterministic(function, domains, outputIndices, testCases);
//
// Second matrix constant, first flattened variables
//
domains = new Domain[1 + outLength + in1Length];
Arrays.fill(domains, rd);
domains[domains.length - 1] = null; // second input matrix
for (int i = 0; i < nCases; ++i)
{
double[][] in1Matrix = in1Matrices[i];
double[][] in2Matrix = in2Matrices[i];
double[][] outMatrix = outMatrices[i];
Value[] testCase = testCases[i] = new Value[domains.length];
int j = 0;
for (int col = 0; col < nc; ++col)
for (int row = 0; row < nr; ++row)
testCase[j++] = Value.create(outMatrix[row][col]);
for (int col = 0; col < in1Col; ++col)
for (int row = 0; row < in1Row; ++row)
testCase[j++] = Value.create(in1Matrix[row][col]);
testCase[j++] = Value.create(in2Matrix);
}
testEvalDeterministic(function, domains, outputIndices, testCases);
//
// First matrix constant, second flattened variables
//
domains = new Domain[1 + outLength + in2Length];
Arrays.fill(domains, rd);
domains[outLength] = null; // first input matrix
for (int i = 0; i < nCases; ++i)
{
double[][] in1Matrix = in1Matrices[i];
double[][] in2Matrix = in2Matrices[i];
double[][] outMatrix = outMatrices[i];
Value[] testCase = testCases[i] = new Value[domains.length];
int j = 0;
for (int col = 0; col < nc; ++col)
for (int row = 0; row < nr; ++row)
testCase[j++] = Value.create(outMatrix[row][col]);
testCase[j++] = Value.create(in1Matrix);
for (int col = 0; col < in2Col; ++col)
for (int row = 0; row < in2Row; ++row)
testCase[j++] = Value.create(in2Matrix[row][col]);
}
testEvalDeterministic(function, domains, outputIndices, testCases);
//
// Flattened matrix variable inputs
//
domains = new Domain[] { rd };
for (int i = 0; i < nCases; ++i)
{
double[][] in1Matrix = in1Matrices[i];
double[][] in2Matrix = in2Matrices[i];
double[][] outMatrix = outMatrices[i];
Value[] testCase = testCases[i] = new Value[outLength + in1Length + in2Length];
int j = 0;
for (int col = 0; col < nc; ++col)
for (int row = 0; row < nr; ++row)
testCase[j++] = Value.create(outMatrix[row][col]);
for (int col = 0; col < in1Col; ++col)
for (int row = 0; row < in1Row; ++row)
testCase[j++] = Value.create(in1Matrix[row][col]);
for (int col = 0; col < in2Col; ++col)
for (int row = 0; row < in2Row; ++row)
testCase[j++] = Value.create(in2Matrix[row][col]);
}
testEvalDeterministic(function, domains, outputIndices, testCases);
// TODO: test evalEnergy for non-deterministic cases w/ smoothing
}
}