1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2004, 2008, 2009, 2011, 2012 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
19 #include "math/histogram.h"
21 #include <gsl/gsl_histogram.h>
24 #include "data/settings.h"
25 #include "libpspp/message.h"
26 #include "libpspp/assertion.h"
27 #include "libpspp/cast.h"
28 #include "math/chart-geometry.h"
31 #define _(msgid) gettext (msgid)
32 #define N_(msgid) msgid
35 #include "gl/xalloc.h"
38 histogram_add (struct histogram *h, double y, double c)
40 struct statistic *stat = &h->parent;
41 stat->accumulate (stat, NULL, c, 0, y);
45 acc (struct statistic *s, const struct ccase *cx UNUSED, double c, double cc UNUSED, double y)
47 struct histogram *hist = UP_CAST (s, struct histogram, parent);
48 gsl_histogram *gslh = hist->gsl_hist;
50 /* Include cases which are just on the boundary */
51 if (y == gsl_histogram_max (gslh))
54 gsl_histogram_get_range (gslh, gsl_histogram_bins (gslh)-1, &lower, &upper);
55 gsl_histogram_accumulate (gslh, lower + (upper - lower)/2.0, c);
58 gsl_histogram_accumulate (hist->gsl_hist, y, c);
62 destroy (struct statistic *s)
64 struct histogram *h = UP_CAST (s, struct histogram, parent);
65 gsl_histogram_free (h->gsl_hist);
70 /* Find a bin width which is adapted to the scaling of the x axis
71 In the example here, the binwidth is half of the tick interval.
75 |....+....+....+. .+....|
79 This only works, when the min and max value for the histogram are adapted
80 such that (max-min) is a multiple of the binwidth. Then the location of the
81 first bin has to be aligned to the ticks.
84 hist_find_pretty_no_of_bins(double bin_width_in, double min, double max,
85 double *adjusted_min, double *adjusted_max)
87 double lower, interval;
92 chart_get_scale (max, min, &lower, &interval, &n_ticks);
94 if (bin_width_in >= 2 * interval)
96 binwidth = floor(bin_width_in/interval) * interval;
97 *adjusted_min = lower;
99 else if (bin_width_in >= 1.5 * interval)
101 binwidth = 1.5 * interval;
102 if (min < (lower + 0.5 * interval))
103 *adjusted_min = lower;
105 *adjusted_min = lower + 0.5 * interval;
107 else if (bin_width_in >= interval)
110 *adjusted_min = lower;
112 else if (bin_width_in >= (2.0/3.0 * interval))
114 binwidth = (2.0/3.0 * interval);
115 if (min >= lower + binwidth)
116 *adjusted_min = lower + binwidth;
118 *adjusted_min = lower;
123 for(i = 2; bin_width_in < interval/i; i++);
124 binwidth = interval/i;
125 *adjusted_min = floor((min - lower)/binwidth)*binwidth + lower;
128 nbins = ceil((max-*adjusted_min)/binwidth);
129 *adjusted_max = nbins*binwidth + *adjusted_min;
136 histogram_create (double bin_width_in, double min, double max)
139 struct statistic *stat;
141 double adjusted_min, adjusted_max;
145 msg (MW, _("Not creating histogram because the data contains less than 2 distinct values"));
149 assert (bin_width_in > 0);
151 bins = hist_find_pretty_no_of_bins(bin_width_in, min, max, &adjusted_min, &adjusted_max);
153 h = xmalloc (sizeof *h);
155 h->gsl_hist = gsl_histogram_alloc (bins);
157 gsl_histogram_set_ranges_uniform (h->gsl_hist, adjusted_min, adjusted_max);
160 stat->accumulate = acc;
161 stat->destroy = destroy;