From e37ed32d9b861581942ec5cfb8948f0602c0a481 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bj=C3=B6rn=20B=2E=20Brandenburg?= <bbb@cs.unc.edu>
Date: Wed, 24 Nov 2010 14:59:25 -0500
Subject: add --hist option to vplot.py to plot histograms

---
 vplot.py | 106 +++++++++++++++++++++++++++++++++++++++++++++++----------------
 1 file changed, 80 insertions(+), 26 deletions(-)

(limited to 'vplot.py')

diff --git a/vplot.py b/vplot.py
index 71344c6..81ec2e8 100755
--- a/vplot.py
+++ b/vplot.py
@@ -6,7 +6,7 @@ from util import load_csv_file, write_csv_file
 
 from math import ceil
 
-from numpy import amin, amax, mean, median, std, histogram
+from numpy import amin, amax, mean, median, std, histogram, zeros
 
 from os.path  import splitext, basename
 from optparse import make_option as o
@@ -23,6 +23,7 @@ options = [
     # formatting options
     # These may or may not be supported by a particular experiment plotter.
     o(None, '--smooth', action='store_true', dest='smooth'),
+    o(None, '--hist', action='store_true', dest='histogram'),
     ]
 
 defaults = {
@@ -32,9 +33,20 @@ defaults = {
     'prefix' : '',
 
     # formatting options
-    'alternate' : False,
+    'histogram' : False,
+    'smooth' : False,
+
+    'binsize' : 0.1,
     }
 
+def get_stats_label(samples):
+    avg = mean(samples)
+    med = median(samples)
+    dev = std(samples)
+    max = amax(samples)
+    min = amin(samples)
+    return  "min=%.2fms max=%.2fms avg=%.2fms  median=%.2fms  std=%.2fms" \
+        % (min, max, avg, med, dev)
 
 class VideoPlotter(defapp.App):
     def __init__(self):
@@ -59,18 +71,44 @@ class VideoPlotter(defapp.App):
         else:
             return False
 
-    def load(self, datafile, name):
-        data = load_csv_file(datafile)
-        # make a copy without comments, etc.
+    def write(self, data, name, ext='data'):
         if self.options.save_script:
-            fname = "%s.data" % name
+            fname = "%s.%s" % (name, ext)
             write_csv_file(fname, data)
-            return (data, fname)
+            return fname
         else:
             tmp = write_csv_file(None, data)
             # keep a reference so that it isn't deleted
             self.tmpfiles.append(tmp)
-            return (data, tmp.name)
+            return tmp.name
+
+    def load(self, datafile, name):
+        data = load_csv_file(datafile)
+        # make a copy without comments, etc.
+        return (data, self.write(data, name))
+
+    def write_histogram(self, samples, name, labels=10):
+        max = ceil(amax(samples))
+        bin_size = self.options.binsize
+        num_bins  = int(max / bin_size)
+        (bins, edges) = histogram(samples, bins=num_bins,
+                                  range=(self.options.binsize / 2,
+                                         max + self.options.binsize / 2))
+        data = zeros((num_bins, 3))
+        cumulative = 0
+        for i in xrange(len(bins)):
+            data[i, 0]  = (edges[i] + edges[i + 1]) / 2.0
+            data[i, 1]  = bins[i]
+            cumulative += bins[i]
+            data[i, 2]  = cumulative
+
+        label_rate = len(bins) / labels
+        for_file = []
+        for i, row in enumerate(data):
+            label = '%.2f' % row[0] if i % label_rate == 0 else ''
+            for_file.append([row[0], row[1], row[2], label])
+
+        return (data, self.write(for_file, name, ext='hist'))
 
     def render(self, p):
         if self.options.save_script:
@@ -81,31 +119,47 @@ class VideoPlotter(defapp.App):
     def plot_vdecode(self, datafile, name, conf):
         (data, fname)  = self.load(datafile, name)
 
-        decode_times = data[:,0]
+        max_cost = amax(data[:,0])
 
-        avg = mean(decode_times)
-        med = median(decode_times)
-        dev = std(decode_times)
-        max = amax(decode_times)
-        min = amin(decode_times)
-
-        stats_label = "min=%.2fms max=%.2fms avg=%.2fms  median=%.2fms  std=%.2fms" \
-            % (min, max, avg, med, dev)
+        if self.options.histogram:
+            name += '_hist'
 
         p = self.make_plot(name)
 
-        p.title = "raw decoding cost; input=%s; host=%s" \
-            % (conf['file'], conf['host'])
+        p.labels = [label(0.5, 0.9, get_stats_label(data[:,0]),
+                          coord=['graph', 'screen'], align='center')]
+
+        if self.options.histogram:
+            (data, fname) = self.write_histogram(data[:,0], name)
+
+            p.setup_histogram(gap=1, boxwidth=1.0)
+
+            p.title = "decoding costs; input=%s; host=%s" \
+                % (conf['file'], conf['host'])
 
-        p.ylabel = "decoding cost (ms)"
-        p.xlabel = "frame number"
-        p.xrange = (0, len(data))
+            p.ylabel = "number of frames"
+            p.xlabel = "decoding cost in milliseconds (bin size = %.2fms)" \
+                % self.options.binsize
+#            p.xrange = (0, ceil(max_cost))
+            p.xticks = (0, 10)
+#            p.yticks = (0, 1)
+            p.yrange = (0, (ceil(amax(data[:,1]) / 100) * 100))
+            p.curves = [curve(histogram=fname, col=2, labels_col=4)]
+
+        else:
+            # plot raw samples
+            p.title = "raw decoding cost; input=%s; host=%s" \
+                % (conf['file'], conf['host'])
+
+            p.ylabel = "decoding cost (ms)"
+            p.xlabel = "frame number"
+            p.xrange = (0, len(data))
         #p.xticks = (0, 100)
-        p.yticks = (0, 1)
-        p.yrange = (0, ceil(max))
+            p.yticks = (0, 1)
+            p.yrange = (0, ceil(max_cost))
+
+            p.curves = [curve(fname=fname, xcol=2, ycol=1, title="decoding cost")]
 
-        p.curves = [curve(fname=fname, xcol=2, ycol=1, title="decoding cost")]
-        p.labels = [label(0.5, 0.9, stats_label, coord=['graph', 'screen'], align='center')]
 
         #### Styling.
 
-- 
cgit v1.2.2