Clover icon

Coverage Report

  1. Project Clover database Mon Nov 11 2024 16:01:40 GMT
  2. Package jalview.math

File SparseMatrix.java

 

Coverage histogram

../../img/srcFileCovDistChart9.png
12% of files have more coverage

Code metrics

32
56
10
1
236
139
27
0.48
5.6
10
2.7

Classes

Class Line # Actions
SparseMatrix 33 56 27
0.8163265681.6%
 

Contributing tests

This file is covered by 11 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 jalview.ext.android.SparseDoubleArray;
24   
25    /**
26    * A variant of Matrix intended for use for sparse (mostly zero) matrices. This
27    * class uses a SparseDoubleArray to hold each row of the matrix. The sparse
28    * array only stores non-zero values. This gives a smaller memory footprint, and
29    * fewer matrix calculation operations, for mostly zero matrices.
30    *
31    * @author gmcarstairs
32    */
 
33    public class SparseMatrix extends Matrix
34    {
35    /*
36    * we choose columns for the sparse arrays as this allows
37    * optimisation of the preMultiply() method used in PCA.run()
38    */
39    SparseDoubleArray[] sparseColumns;
40   
41    /**
42    * Constructor given data in [row][column] order
43    *
44    * @param v
45    */
 
46  22 toggle public SparseMatrix(double[][] v)
47    {
48  22 super(v.length, v.length > 0 ? v[0].length : 0);
49   
50  22 sparseColumns = new SparseDoubleArray[cols];
51   
52    /*
53    * transpose v[row][col] into [col][row] order
54    */
55  107 for (int col = 0; col < cols; col++)
56    {
57  85 SparseDoubleArray sparseColumn = new SparseDoubleArray();
58  85 sparseColumns[col] = sparseColumn;
59  522 for (int row = 0; row < rows; row++)
60    {
61  437 double value = v[row][col];
62  437 if (value != 0d)
63    {
64  184 sparseColumn.put(row, value);
65    }
66    }
67    }
68    }
69   
70    /**
71    * Answers the value at row i, column j
72    */
 
73  7019 toggle @Override
74    public double getValue(int i, int j)
75    {
76  7019 return sparseColumns[j].get(i);
77    }
78   
79    /**
80    * Sets the value at row i, column j to val
81    */
 
82  870 toggle @Override
83    public void setValue(int i, int j, double val)
84    {
85  870 if (val == 0d)
86    {
87  385 sparseColumns[j].delete(i);
88    }
89    else
90    {
91  485 sparseColumns[j].put(i, val);
92    }
93    }
94   
 
95  0 toggle @Override
96    public double[] getColumn(int i)
97    {
98  0 double[] col = new double[height()];
99   
100  0 SparseDoubleArray vals = sparseColumns[i];
101  0 for (int nonZero = 0; nonZero < vals.size(); nonZero++)
102    {
103  0 col[vals.keyAt(nonZero)] = vals.valueAt(nonZero);
104    }
105  0 return col;
106    }
107   
 
108  0 toggle @Override
109    public MatrixI copy()
110    {
111  0 double[][] vals = new double[height()][width()];
112  0 for (int i = 0; i < height(); i++)
113    {
114  0 vals[i] = getRow(i);
115    }
116  0 return new SparseMatrix(vals);
117    }
118   
 
119  1 toggle @Override
120    public MatrixI transpose()
121    {
122  1 double[][] out = new double[cols][rows];
123   
124    /*
125    * for each column...
126    */
127  4 for (int i = 0; i < cols; i++)
128    {
129    /*
130    * put non-zero values into the corresponding row
131    * of the transposed matrix
132    */
133  3 SparseDoubleArray vals = sparseColumns[i];
134  7 for (int nonZero = 0; nonZero < vals.size(); nonZero++)
135    {
136  4 out[i][vals.keyAt(nonZero)] = vals.valueAt(nonZero);
137    }
138    }
139   
140  1 return new SparseMatrix(out);
141    }
142   
143    /**
144    * Answers a new matrix which is the product in.this. If the product contains
145    * less than 20% non-zero values, it is returned as a SparseMatrix, else as a
146    * Matrix.
147    * <p>
148    * This method is optimised for the sparse arrays which store column values
149    * for a SparseMatrix. Note that postMultiply is not so optimised. That would
150    * require redundantly also storing sparse arrays for the rows, which has not
151    * been done. Currently only preMultiply is used in Jalview.
152    */
 
153  12 toggle @Override
154    public MatrixI preMultiply(MatrixI in)
155    {
156  12 if (in.width() != rows)
157    {
158  2 throw new IllegalArgumentException("Can't pre-multiply " + this.rows
159    + " rows by " + in.width() + " columns");
160    }
161  10 double[][] tmp = new double[in.height()][this.cols];
162   
163  10 long count = 0L;
164  34 for (int i = 0; i < in.height(); i++)
165    {
166  98 for (int j = 0; j < this.cols; j++)
167    {
168    /*
169    * result[i][j] is the vector product of
170    * in.row[i] and this.column[j]
171    * we only need to use non-zero values from the column
172    */
173  74 SparseDoubleArray vals = sparseColumns[j];
174  74 boolean added = false;
175  161 for (int nonZero = 0; nonZero < vals.size(); nonZero++)
176    {
177  87 int myRow = vals.keyAt(nonZero);
178  87 double myValue = vals.valueAt(nonZero);
179  87 tmp[i][j] += (in.getValue(i, myRow) * myValue);
180  87 added = true;
181    }
182  74 if (added && tmp[i][j] != 0d)
183    {
184  41 count++; // non-zero entry in product
185    }
186    }
187    }
188   
189    /*
190    * heuristic rule - if product is more than 80% zero
191    * then construct a SparseMatrix, else a Matrix
192    */
193  10 if (count * 5 < in.height() * cols)
194    {
195  1 return new SparseMatrix(tmp);
196    }
197    else
198    {
199  9 return new Matrix(tmp);
200    }
201    }
202   
 
203  146 toggle @Override
204    protected double divideValue(int i, int j, double divisor)
205    {
206  146 if (divisor == 0d)
207    {
208  0 return getValue(i, j);
209    }
210  146 double v = sparseColumns[j].divide(i, divisor);
211  146 return v;
212    }
213   
 
214  1432 toggle @Override
215    protected double addValue(int i, int j, double addend)
216    {
217  1432 double v = sparseColumns[j].add(i, addend);
218  1432 return v;
219    }
220   
221    /**
222    * Returns the fraction of the whole matrix size that is actually modelled in
223    * sparse arrays (normally, the non-zero values)
224    *
225    * @return
226    */
 
227  1 toggle public float getFillRatio()
228    {
229  1 long count = 0L;
230  1 for (SparseDoubleArray col : sparseColumns)
231    {
232  5 count += col.size();
233    }
234  1 return count / (float) (height() * width());
235    }
236    }