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);
49 gsl_histogram_accumulate (hist->gsl_hist, y, c);
53 destroy (struct statistic *s)
55 struct histogram *h = UP_CAST (s, struct histogram, parent);
56 gsl_histogram_free (h->gsl_hist);
61 /* Find a bin width which is adapted to the scaling of the x axis
62 In the example here, the binwidth is half of the tick interval.
66 |....+....+....+. .+....|
70 This only works, when the min and max value for the histogram are adapted
71 such that (max-min) is a multiple of the binwidth. Then the location of the
72 first bin has to be aligned to the ticks.
75 hist_find_pretty_no_of_bins(double bin_width_in, double min, double max,
76 double *adjusted_min, double *adjusted_max)
78 double lower, interval;
83 chart_get_scale (max, min, &lower, &interval, &n_ticks);
85 if (bin_width_in >= 2 * interval)
87 binwidth = floor(bin_width_in/interval) * interval;
88 *adjusted_min = lower;
90 else if (bin_width_in >= 1.5 * interval)
92 binwidth = 1.5 * interval;
93 if (min < (lower + 0.5 * interval))
94 *adjusted_min = lower;
96 *adjusted_min = lower + 0.5 * interval;
98 else if (bin_width_in >= interval)
101 *adjusted_min = lower;
103 else if (bin_width_in >= (2.0/3.0 * interval))
105 binwidth = (2.0/3.0 * interval);
106 if (min >= lower + binwidth)
107 *adjusted_min = lower + binwidth;
109 *adjusted_min = lower;
114 for(i = 2; bin_width_in < interval/i; i++);
115 binwidth = interval/i;
116 *adjusted_min = floor((min - lower)/binwidth)*binwidth + lower;
119 nbins = ceil((max-*adjusted_min)/binwidth);
120 *adjusted_max = nbins*binwidth + *adjusted_min;
127 histogram_create (double bin_width_in, double min, double max)
130 struct statistic *stat;
132 double adjusted_min, adjusted_max;
136 msg (MW, _("Not creating histogram because the data contains less than 2 distinct values"));
140 assert (bin_width_in > 0);
142 bins = hist_find_pretty_no_of_bins(bin_width_in, min, max, &adjusted_min, &adjusted_max);
144 h = xmalloc (sizeof *h);
146 h->gsl_hist = gsl_histogram_alloc (bins);
148 gsl_histogram_set_ranges_uniform (h->gsl_hist, adjusted_min, adjusted_max);
151 stat->accumulate = acc;
152 stat->destroy = destroy;