1. Project Clover database Fri Dec 6 2024 13:47:14 GMT
  2. Package jalview.math

File MatrixTest.java

 

Code metrics

44
227
19
1
605
376
42
0.19
11.95
19
2.21

Classes

Class
Line #
Actions
MatrixTest 36 227 42
0.8379310483.8%
 

Contributing tests

This file is covered by 13 tests. .

Source view

1    /*
2    * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3    * Copyright (C) $$Year-Rel$$ The Jalview Authors
4    *
5    * This file is part of Jalview.
6    *
7    * Jalview is free software: you can redistribute it and/or
8    * modify it under the terms of the GNU General Public License
9    * as published by the Free Software Foundation, either version 3
10    * of the License, or (at your option) any later version.
11    *
12    * Jalview is distributed in the hope that it will be useful, but
13    * WITHOUT ANY WARRANTY; without even the implied warranty
14    * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15    * PURPOSE. See the GNU General Public License for more details.
16    *
17    * You should have received a copy of the GNU General Public License
18    * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19    * The Jalview Authors are detailed in the 'AUTHORS' file.
20    */
21    package jalview.math;
22   
23    import static org.testng.Assert.assertEquals;
24    import static org.testng.Assert.assertFalse;
25    import static org.testng.Assert.assertNotSame;
26    import static org.testng.Assert.assertNull;
27    import static org.testng.Assert.assertTrue;
28    import static org.testng.Assert.fail;
29   
30    import java.util.Arrays;
31    import java.util.Random;
32   
33    import org.testng.annotations.Test;
34    import org.testng.internal.junit.ArrayAsserts;
35   
 
36    public class MatrixTest
37    {
38    final static double DELTA = 0.000001d;
39   
 
40  0 toggle @Test(groups = "Timing")
41    public void testPreMultiply_timing()
42    {
43  0 int rows = 50; // increase to stress test timing
44  0 int cols = 100;
45  0 double[][] d1 = new double[rows][cols];
46  0 double[][] d2 = new double[cols][rows];
47  0 Matrix m1 = new Matrix(d1);
48  0 Matrix m2 = new Matrix(d2);
49  0 long start = System.currentTimeMillis();
50  0 m1.preMultiply(m2);
51  0 long elapsed = System.currentTimeMillis() - start;
52  0 System.out.println(rows + "x" + cols
53    + " multiplications of double took " + elapsed + "ms");
54    }
55   
 
56  1 toggle @Test(groups = "Functional")
57    public void testPreMultiply()
58    {
59  1 Matrix m1 = new Matrix(new double[][] { { 2, 3, 4 } }); // 1x3
60  1 Matrix m2 = new Matrix(new double[][] { { 5 }, { 6 }, { 7 } }); // 3x1
61   
62    /*
63    * 1x3 times 3x1 is 1x1
64    * 2x5 + 3x6 + 4*7 = 56
65    */
66  1 MatrixI m3 = m2.preMultiply(m1);
67  1 assertEquals(m3.height(), 1);
68  1 assertEquals(m3.width(), 1);
69  1 assertEquals(m3.getValue(0, 0), 56d);
70   
71    /*
72    * 3x1 times 1x3 is 3x3
73    */
74  1 m3 = m1.preMultiply(m2);
75  1 assertEquals(m3.height(), 3);
76  1 assertEquals(m3.width(), 3);
77  1 assertEquals(Arrays.toString(m3.getRow(0)), "[10.0, 15.0, 20.0]");
78  1 assertEquals(Arrays.toString(m3.getRow(1)), "[12.0, 18.0, 24.0]");
79  1 assertEquals(Arrays.toString(m3.getRow(2)), "[14.0, 21.0, 28.0]");
80    }
81   
 
82  1 toggle @Test(
83    groups = "Functional",
84    expectedExceptions =
85    { IllegalArgumentException.class })
86    public void testPreMultiply_tooManyColumns()
87    {
88  1 Matrix m1 = new Matrix(new double[][] { { 2, 3, 4 }, { 3, 4, 5 } }); // 2x3
89   
90    /*
91    * 2x3 times 2x3 invalid operation -
92    * multiplier has more columns than multiplicand has rows
93    */
94  1 m1.preMultiply(m1);
95  0 fail("Expected exception");
96    }
97   
 
98  1 toggle @Test(
99    groups = "Functional",
100    expectedExceptions =
101    { IllegalArgumentException.class })
102    public void testPreMultiply_tooFewColumns()
103    {
104  1 Matrix m1 = new Matrix(new double[][] { { 2, 3, 4 }, { 3, 4, 5 } }); // 2x3
105   
106    /*
107    * 3x2 times 3x2 invalid operation -
108    * multiplier has more columns than multiplicand has row
109    */
110  1 m1.preMultiply(m1);
111  0 fail("Expected exception");
112    }
113   
 
114  1 toggle private boolean matrixEquals(Matrix m1, Matrix m2)
115    {
116  1 if (m1.width() != m2.width() || m1.height() != m2.height())
117    {
118  0 return false;
119    }
120  6 for (int i = 0; i < m1.height(); i++)
121    {
122  5 if (!Arrays.equals(m1.getRow(i), m2.getRow(i)))
123    {
124  0 return false;
125    }
126    }
127  1 return true;
128    }
129   
 
130  1 toggle @Test(groups = "Functional")
131    public void testPostMultiply()
132    {
133    /*
134    * Square matrices
135    * (2 3) . (10 100)
136    * (4 5) (1000 10000)
137    * =
138    * (3020 30200)
139    * (5040 50400)
140    */
141  1 MatrixI m1 = new Matrix(new double[][] { { 2, 3 }, { 4, 5 } });
142  1 MatrixI m2 = new Matrix(
143    new double[][]
144    { { 10, 100 }, { 1000, 10000 } });
145  1 MatrixI m3 = m1.postMultiply(m2);
146  1 assertEquals(Arrays.toString(m3.getRow(0)), "[3020.0, 30200.0]");
147  1 assertEquals(Arrays.toString(m3.getRow(1)), "[5040.0, 50400.0]");
148   
149    /*
150    * also check m2.preMultiply(m1) - should be same as m1.postMultiply(m2)
151    */
152  1 m3 = m2.preMultiply(m1);
153  1 assertEquals(Arrays.toString(m3.getRow(0)), "[3020.0, 30200.0]");
154  1 assertEquals(Arrays.toString(m3.getRow(1)), "[5040.0, 50400.0]");
155   
156    /*
157    * m1 has more rows than columns
158    * (2).(10 100 1000) = (20 200 2000)
159    * (3) (30 300 3000)
160    */
161  1 m1 = new Matrix(new double[][] { { 2 }, { 3 } });
162  1 m2 = new Matrix(new double[][] { { 10, 100, 1000 } });
163  1 m3 = m1.postMultiply(m2);
164  1 assertEquals(m3.height(), 2);
165  1 assertEquals(m3.width(), 3);
166  1 assertEquals(Arrays.toString(m3.getRow(0)), "[20.0, 200.0, 2000.0]");
167  1 assertEquals(Arrays.toString(m3.getRow(1)), "[30.0, 300.0, 3000.0]");
168  1 m3 = m2.preMultiply(m1);
169  1 assertEquals(m3.height(), 2);
170  1 assertEquals(m3.width(), 3);
171  1 assertEquals(Arrays.toString(m3.getRow(0)), "[20.0, 200.0, 2000.0]");
172  1 assertEquals(Arrays.toString(m3.getRow(1)), "[30.0, 300.0, 3000.0]");
173   
174    /*
175    * m1 has more columns than rows
176    * (2 3 4) . (5 4) = (56 25)
177    * (6 3)
178    * (7 2)
179    * [0, 0] = 2*5 + 3*6 + 4*7 = 56
180    * [0, 1] = 2*4 + 3*3 + 4*2 = 25
181    */
182  1 m1 = new Matrix(new double[][] { { 2, 3, 4 } });
183  1 m2 = new Matrix(new double[][] { { 5, 4 }, { 6, 3 }, { 7, 2 } });
184  1 m3 = m1.postMultiply(m2);
185  1 assertEquals(m3.height(), 1);
186  1 assertEquals(m3.width(), 2);
187  1 assertEquals(m3.getRow(0)[0], 56d);
188  1 assertEquals(m3.getRow(0)[1], 25d);
189   
190    /*
191    * and check premultiply equivalent
192    */
193  1 m3 = m2.preMultiply(m1);
194  1 assertEquals(m3.height(), 1);
195  1 assertEquals(m3.width(), 2);
196  1 assertEquals(m3.getRow(0)[0], 56d);
197  1 assertEquals(m3.getRow(0)[1], 25d);
198    }
199   
 
200  1 toggle @Test(groups = "Functional")
201    public void testCopy()
202    {
203  1 Random r = new Random();
204  1 int rows = 5;
205  1 int cols = 11;
206  1 double[][] in = new double[rows][cols];
207   
208  6 for (int i = 0; i < rows; i++)
209    {
210  60 for (int j = 0; j < cols; j++)
211    {
212  55 in[i][j] = r.nextDouble();
213    }
214    }
215  1 Matrix m1 = new Matrix(in);
216   
217  1 Matrix m2 = (Matrix) m1.copy();
218  1 assertNotSame(m1, m2);
219  1 assertTrue(matrixEquals(m1, m2));
220  1 assertNull(m2.d);
221  1 assertNull(m2.e);
222   
223    /*
224    * now add d and e vectors and recopy
225    */
226  1 m1.d = Arrays.copyOf(in[2], in[2].length);
227  1 m1.e = Arrays.copyOf(in[4], in[4].length);
228  1 m2 = (Matrix) m1.copy();
229  1 assertNotSame(m2.d, m1.d);
230  1 assertNotSame(m2.e, m1.e);
231  1 assertEquals(m2.d, m1.d);
232  1 assertEquals(m2.e, m1.e);
233    }
234   
235    /**
236    * main method extracted from Matrix
237    *
238    * @param args
239    */
 
240  0 toggle public static void main(String[] args) throws Exception
241    {
242  0 int n = Integer.parseInt(args[0]);
243  0 double[][] in = new double[n][n];
244   
245  0 for (int i = 0; i < n; i++)
246    {
247  0 for (int j = 0; j < n; j++)
248    {
249  0 in[i][j] = Math.random();
250    }
251    }
252   
253  0 Matrix origmat = new Matrix(in);
254   
255    // System.out.println(" --- Original matrix ---- ");
256    // / origmat.print(System.out);
257    // System.out.println();
258    // System.out.println(" --- transpose matrix ---- ");
259  0 MatrixI trans = origmat.transpose();
260   
261    // trans.print(System.out);
262    // System.out.println();
263    // System.out.println(" --- OrigT * Orig ---- ");
264  0 MatrixI symm = trans.postMultiply(origmat);
265   
266    // symm.print(System.out);
267    // System.out.println();
268    // Copy the symmetric matrix for later
269    // Matrix origsymm = symm.copy();
270   
271    // This produces the tridiagonal transformation matrix
272    // long tstart = System.currentTimeMillis();
273  0 symm.tred();
274   
275    // long tend = System.currentTimeMillis();
276   
277    // System.out.println("Time take for tred = " + (tend-tstart) + "ms");
278    // System.out.println(" ---Tridiag transform matrix ---");
279    // symm.print(System.out);
280    // System.out.println();
281    // System.out.println(" --- D vector ---");
282    // symm.printD(System.out);
283    // System.out.println();
284    // System.out.println(" --- E vector ---");
285    // symm.printE(System.out);
286    // System.out.println();
287    // Now produce the diagonalization matrix
288    // tstart = System.currentTimeMillis();
289  0 symm.tqli();
290    // tend = System.currentTimeMillis();
291   
292    // System.out.println("Time take for tqli = " + (tend-tstart) + " ms");
293    // System.out.println(" --- New diagonalization matrix ---");
294    // symm.print(System.out);
295    // System.out.println();
296    // System.out.println(" --- D vector ---");
297    // symm.printD(System.out);
298    // System.out.println();
299    // System.out.println(" --- E vector ---");
300    // symm.printE(System.out);
301    // System.out.println();
302    // System.out.println(" --- First eigenvector --- ");
303    // double[] eigenv = symm.getColumn(0);
304    // for (int i=0; i < eigenv.length;i++) {
305    // Format.print(System.out,"%15.4f",eigenv[i]);
306    // }
307    // System.out.println();
308    // double[] neigenv = origsymm.vectorPostMultiply(eigenv);
309    // for (int i=0; i < neigenv.length;i++) {
310    // Format.print(System.out,"%15.4f",neigenv[i]/symm.d[0]);
311    // }
312    // System.out.println();
313    }
314   
 
315  0 toggle @Test(groups = "Timing")
316    public void testSign()
317    {
318  0 assertEquals(Matrix.sign(-1, -2), -1d);
319  0 assertEquals(Matrix.sign(-1, 2), 1d);
320  0 assertEquals(Matrix.sign(-1, 0), 1d);
321  0 assertEquals(Matrix.sign(1, -2), -1d);
322  0 assertEquals(Matrix.sign(1, 2), 1d);
323  0 assertEquals(Matrix.sign(1, 0), 1d);
324    }
325   
326    /**
327    * Helper method to make values for a sparse, pseudo-random symmetric matrix
328    *
329    * @param rows
330    * @param cols
331    * @param occupancy
332    * one in 'occupancy' entries will be non-zero
333    * @return
334    */
 
335  1 toggle public double[][] getSparseValues(int rows, int cols, int occupancy)
336    {
337  1 Random r = new Random(1729);
338   
339    /*
340    * generate whole number values between -12 and +12
341    * (to mimic score matrices used in Jalview)
342    */
343  1 double[][] d = new double[rows][cols];
344  1 int m = 0;
345  11 for (int i = 0; i < rows; i++)
346    {
347  10 if (++m % occupancy == 0)
348    {
349  0 d[i][i] = r.nextInt() % 13; // diagonal
350    }
351  55 for (int j = 0; j < i; j++)
352    {
353  45 if (++m % occupancy == 0)
354    {
355  18 d[i][j] = r.nextInt() % 13;
356  18 d[j][i] = d[i][j];
357    }
358    }
359    }
360  1 return d;
361   
362    }
363   
364    /**
365    * Verify that the results of method tred() are the same if the calculation is
366    * redone
367    */
 
368  1 toggle @Test(groups = "Functional")
369    public void testTred_reproducible()
370    {
371    /*
372    * make a pseudo-random symmetric matrix as required for tred/tqli
373    */
374  1 int rows = 10;
375  1 int cols = rows;
376  1 double[][] d = getSparseValues(rows, cols, 3);
377   
378    /*
379    * make a copy of the values so m1, m2 are not
380    * sharing arrays!
381    */
382  1 double[][] d1 = new double[rows][cols];
383  11 for (int row = 0; row < rows; row++)
384    {
385  110 for (int col = 0; col < cols; col++)
386    {
387  100 d1[row][col] = d[row][col];
388    }
389    }
390  1 Matrix m1 = new Matrix(d);
391  1 Matrix m2 = new Matrix(d1);
392  1 assertMatricesMatch(m1, m2); // sanity check
393  1 m1.tred();
394  1 m2.tred();
395  1 assertMatricesMatch(m1, m2);
396    }
397   
 
398  2 toggle public static void assertMatricesMatch(MatrixI m1, MatrixI m2)
399    {
400  2 if (m1.height() != m2.height())
401    {
402  0 fail("height mismatch");
403    }
404  2 if (m1.width() != m2.width())
405    {
406  0 fail("width mismatch");
407    }
408  22 for (int row = 0; row < m1.height(); row++)
409    {
410  220 for (int col = 0; col < m1.width(); col++)
411    {
412  200 double v2 = m2.getValue(row, col);
413  200 double v1 = m1.getValue(row, col);
414  200 if (Math.abs(v1 - v2) > DELTA)
415    {
416  0 fail(String.format("At [%d, %d] %f != %f", row, col, v1, v2));
417    }
418    }
419    }
420  2 ArrayAsserts.assertArrayEquals("D vector", m1.getD(), m2.getD(),
421    0.00001d);
422  2 ArrayAsserts.assertArrayEquals("E vector", m1.getE(), m2.getE(),
423    0.00001d);
424    }
425   
 
426  1 toggle @Test(groups = "Functional")
427    public void testFindMinMax()
428    {
429    /*
430    * empty matrix case
431    */
432  1 Matrix m = new Matrix(new double[][] { {} });
433  1 assertNull(m.findMinMax());
434   
435    /*
436    * normal case
437    */
438  1 double[][] vals = new double[2][];
439  1 vals[0] = new double[] { 7d, 1d, -2.3d };
440  1 vals[1] = new double[] { -12d, 94.3d, -102.34d };
441  1 m = new Matrix(vals);
442  1 double[] minMax = m.findMinMax();
443  1 assertEquals(minMax[0], -102.34d);
444  1 assertEquals(minMax[1], 94.3d);
445    }
446   
 
447  1 toggle @Test(groups = { "Functional", "Timing" })
448    public void testFindMinMax_timing()
449    {
450  1 Random r = new Random();
451  1 int size = 1000; // increase to stress test timing
452  1 double[][] vals = new double[size][size];
453  1 double max = -Double.MAX_VALUE;
454  1 double min = Double.MAX_VALUE;
455  1001 for (int i = 0; i < size; i++)
456    {
457  1000 vals[i] = new double[size];
458  1001000 for (int j = 0; j < size; j++)
459    {
460    // use nextLong rather than nextDouble to include negative values
461  1000000 double d = r.nextLong();
462  1000000 if (d > max)
463    {
464  13 max = d;
465    }
466  1000000 if (d < min)
467    {
468  12 min = d;
469    }
470  1000000 vals[i][j] = d;
471    }
472    }
473  1 Matrix m = new Matrix(vals);
474  1 long now = System.currentTimeMillis();
475  1 double[] minMax = m.findMinMax();
476  1 System.out.println(String.format("findMinMax for %d x %d took %dms",
477    size, size, (System.currentTimeMillis() - now)));
478  1 assertEquals(minMax[0], min);
479  1 assertEquals(minMax[1], max);
480    }
481   
482    /**
483    * Test range reversal with maximum value becoming zero
484    */
 
485  1 toggle @Test(groups = "Functional")
486    public void testReverseRange_maxToZero()
487    {
488  1 Matrix m1 = new Matrix(
489    new double[][]
490    { { 2, 3.5, 4 }, { -3.4, 4, 15 } });
491   
492    /*
493    * subtract all from max: range -3.4 to 15 becomes 18.4 to 0
494    */
495  1 m1.reverseRange(true);
496  1 assertEquals(m1.getValue(0, 0), 13d, DELTA);
497  1 assertEquals(m1.getValue(0, 1), 11.5d, DELTA);
498  1 assertEquals(m1.getValue(0, 2), 11d, DELTA);
499  1 assertEquals(m1.getValue(1, 0), 18.4d, DELTA);
500  1 assertEquals(m1.getValue(1, 1), 11d, DELTA);
501  1 assertEquals(m1.getValue(1, 2), 0d, DELTA);
502   
503    /*
504    * repeat operation - range is now 0 to 18.4
505    */
506  1 m1.reverseRange(true);
507  1 assertEquals(m1.getValue(0, 0), 5.4d, DELTA);
508  1 assertEquals(m1.getValue(0, 1), 6.9d, DELTA);
509  1 assertEquals(m1.getValue(0, 2), 7.4d, DELTA);
510  1 assertEquals(m1.getValue(1, 0), 0d, DELTA);
511  1 assertEquals(m1.getValue(1, 1), 7.4d, DELTA);
512  1 assertEquals(m1.getValue(1, 2), 18.4d, DELTA);
513    }
514   
515    /**
516    * Test range reversal with minimum and maximum values swapped
517    */
 
518  1 toggle @Test(groups = "Functional")
519    public void testReverseRange_swapMinMax()
520    {
521  1 Matrix m1 = new Matrix(
522    new double[][]
523    { { 2, 3.5, 4 }, { -3.4, 4, 15 } });
524   
525    /*
526    * swap all values in min-max range
527    * = subtract from (min + max = 11.6)
528    * range -3.4 to 15 becomes 18.4 to -3.4
529    */
530  1 m1.reverseRange(false);
531  1 assertEquals(m1.getValue(0, 0), 9.6d, DELTA);
532  1 assertEquals(m1.getValue(0, 1), 8.1d, DELTA);
533  1 assertEquals(m1.getValue(0, 2), 7.6d, DELTA);
534  1 assertEquals(m1.getValue(1, 0), 15d, DELTA);
535  1 assertEquals(m1.getValue(1, 1), 7.6d, DELTA);
536  1 assertEquals(m1.getValue(1, 2), -3.4d, DELTA);
537   
538    /*
539    * repeat operation - original values restored
540    */
541  1 m1.reverseRange(false);
542  1 assertEquals(m1.getValue(0, 0), 2d, DELTA);
543  1 assertEquals(m1.getValue(0, 1), 3.5d, DELTA);
544  1 assertEquals(m1.getValue(0, 2), 4d, DELTA);
545  1 assertEquals(m1.getValue(1, 0), -3.4d, DELTA);
546  1 assertEquals(m1.getValue(1, 1), 4d, DELTA);
547  1 assertEquals(m1.getValue(1, 2), 15d, DELTA);
548    }
549   
 
550  1 toggle @Test(groups = "Functional")
551    public void testMultiply()
552    {
553  1 Matrix m = new Matrix(
554    new double[][]
555    { { 2, 3.5, 4 }, { -3.4, 4, 15 } });
556  1 m.multiply(2d);
557  1 assertEquals(m.getValue(0, 0), 4d, DELTA);
558  1 assertEquals(m.getValue(0, 1), 7d, DELTA);
559  1 assertEquals(m.getValue(0, 2), 8d, DELTA);
560  1 assertEquals(m.getValue(1, 0), -6.8d, DELTA);
561  1 assertEquals(m.getValue(1, 1), 8d, DELTA);
562  1 assertEquals(m.getValue(1, 2), 30d, DELTA);
563    }
564   
 
565  1 toggle @Test(groups = "Functional")
566    public void testConstructor()
567    {
568  1 double[][] values = new double[][] { { 1, 2, 3 }, { 4, 5, 6 } };
569  1 Matrix m = new Matrix(values);
570  1 assertEquals(m.getValue(0, 0), 1d, DELTA);
571   
572    /*
573    * verify the matrix has a copy of the original array
574    */
575  1 assertNotSame(values[0], m.getRow(0));
576  1 values[0][0] = -1d;
577  1 assertEquals(m.getValue(0, 0), 1d, DELTA); // unchanged
578    }
579   
 
580  1 toggle @Test(groups = "Functional")
581    public void testEquals()
582    {
583  1 double[][] values = new double[][] { { 1, 2, 3 }, { 4, 5, 6 } };
584  1 Matrix m1 = new Matrix(values);
585  1 double[][] values2 = new double[][] { { 1, 2, 3 }, { 4, 5, 6 } };
586  1 Matrix m2 = new Matrix(values2);
587   
588  1 double delta = 0.0001d;
589  1 assertTrue(m1.equals(m1, delta));
590  1 assertTrue(m1.equals(m2, delta));
591  1 assertTrue(m2.equals(m1, delta));
592   
593  1 double[][] values3 = new double[][] { { 1, 2, 3 }, { 4, 5, 7 } };
594  1 m2 = new Matrix(values3);
595  1 assertFalse(m1.equals(m2, delta));
596  1 assertFalse(m2.equals(m1, delta));
597   
598    // must be same shape
599  1 values2 = new double[][] { { 1, 2, 3 } };
600  1 m2 = new Matrix(values2);
601  1 assertFalse(m2.equals(m1, delta));
602   
603  1 assertFalse(m1.equals(null, delta));
604    }
605    }