X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Fmath%2Fhistogram.c;h=9158590dd75c3a140244292004087a773eddec3c;hb=50c7b5b389eb39accf65c8716f6a3f73b5b30aa6;hp=1b55d204b866fa2e15dc4ff3f84d3b3a509a9915;hpb=fe8dc2171009e90d2335f159d05f7e6660e24780;p=pspp

diff --git a/src/math/histogram.c b/src/math/histogram.c
index 1b55d204b8..9158590dd7 100644
--- a/src/math/histogram.c
+++ b/src/math/histogram.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 2004, 2008, 2009, 2011 Free Software Foundation, Inc.
+   Copyright (C) 2004, 2008, 2009, 2011, 2012 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -21,10 +21,17 @@
 #include <gsl/gsl_histogram.h>
 #include <math.h>
 
+#include "data/settings.h"
+#include "libpspp/message.h"
 #include "libpspp/assertion.h"
 #include "libpspp/cast.h"
 #include "math/chart-geometry.h"
 
+#include "gettext.h"
+#define _(msgid) gettext (msgid)
+#define N_(msgid) msgid
+
+
 #include "gl/xalloc.h"
 
 void
@@ -34,8 +41,6 @@ histogram_add (struct histogram *h, double y, double c)
   stat->accumulate (stat, NULL, c, 0, y);
 }
 
-
-
 static void
 acc (struct statistic *s, const struct ccase *cx UNUSED, double c, double cc UNUSED, double y)
 {
@@ -44,7 +49,6 @@ acc (struct statistic *s, const struct ccase *cx UNUSED, double c, double cc UNU
   gsl_histogram_accumulate (hist->gsl_hist, y, c);
 }
 
-
 static void
 destroy (struct statistic *s)
 {
@@ -54,30 +58,96 @@ destroy (struct statistic *s)
 }
 
 
-struct histogram *
-histogram_create (int bins, double min, double max)
+/* Find a bin width which is adapted to the scaling of the x axis
+In the example here, the binwidth is half of the tick interval.
+
+        binwidth
+         >   <
+     |....+....+....+.      .+....|
+   LOWER  1    2    3     N_TICKS
+        ^LOWDBL                 ^HIGHDBL
+
+This only works, when the min and max value for the histogram are adapted
+such that (max-min) is a multiple of the binwidth. Then the location of the
+first bin has to be aligned to the ticks.
+*/
+static int
+hist_find_pretty_no_of_bins(double bin_width_in, double min, double max,
+			    double *adjusted_min, double *adjusted_max)
 {
-  struct histogram *h = xmalloc (sizeof *h);
-  struct statistic *stat = &h->parent;
-  double upper_limit, lower_limit;
+  double lower, interval;
+  int n_ticks;
+  double binwidth;
+  int nbins;
+
+  chart_get_scale (max, min, &lower, &interval, &n_ticks);
+
+  if (bin_width_in >= 2 * interval)
+    {
+      binwidth = floor(bin_width_in/interval) * interval;
+      *adjusted_min = lower;
+    }
+  else if (bin_width_in >= 1.5 * interval)
+    {
+      binwidth = 1.5 * interval;
+      if (min < (lower + 0.5 * interval))
+	*adjusted_min = lower;
+      else
+	*adjusted_min = lower + 0.5 * interval;
+    }
+  else if (bin_width_in >= interval)
+    {
+      binwidth = interval;
+      *adjusted_min = lower;
+    }
+  else if (bin_width_in >= (2.0/3.0 * interval))
+    {
+      binwidth = (2.0/3.0 * interval);
+      if (min >= lower + binwidth)
+	*adjusted_min = lower + binwidth;
+      else
+	*adjusted_min = lower;
+    }
+  else
+    {
+      int i;
+      for(i = 2; bin_width_in < interval/i; i++);
+      binwidth = interval/i;
+      *adjusted_min = floor((min - lower)/binwidth)*binwidth + lower;
+    }
+
+  nbins = ceil((max-*adjusted_min)/binwidth);
+  *adjusted_max = nbins*binwidth + *adjusted_min;
+
+  return nbins;
+}
 
-  double bin_width = chart_rounded_tick ((max - min) / (double) bins);
-  double bin_width_2 = bin_width / 2.0;
 
-  int n =  ceil (max / (bin_width_2) ) ;
+struct histogram *
+histogram_create (double bin_width_in, double min, double max)
+{
+  struct histogram *h;
+  struct statistic *stat;
+  int bins;
+  double adjusted_min, adjusted_max;
+
+  if (max == min)
+    {
+      msg (MW, _("Not creating histogram because the data contains less than 2 distinct values"));
+      return NULL;
+    }
 
-  assert (max >= min);
+  assert (bin_width_in > 0);
 
-  if ( ! (n % 2 ) ) n++;
-  upper_limit = n * bin_width_2;
+  bins = hist_find_pretty_no_of_bins(bin_width_in, min, max, &adjusted_min, &adjusted_max);
 
-  n =  floor (min / (bin_width_2) ) ;
-  if ( ! (n % 2 ) ) n--;
-  lower_limit = n * bin_width_2;
+  h = xmalloc (sizeof *h);
 
   h->gsl_hist = gsl_histogram_alloc (bins);
-  gsl_histogram_set_ranges_uniform (h->gsl_hist, lower_limit, upper_limit);
 
+  gsl_histogram_set_ranges_uniform (h->gsl_hist, adjusted_min, adjusted_max);
+
+  stat = &h->parent;
   stat->accumulate = acc;
   stat->destroy = destroy;