prefix/parseDat.py

   1 #!/usr/bin/env python2
   2
   3 import os, subprocess
   4
   5 def listtofile(listing, file):
   6         fd = open(file, 'r')
   7         for i in listing:
   8                 fd.write(i+"\n")
   9         fd.close()
  10
  11 def makeGraphs(graphs, graphtitle):
  12         gendir = "stats/generated/"
  13         subprocess.call(["mkdir", "-p", gendir])
  14         p = subprocess.Popen(['gnuplot'],stdout=subprocess.PIPE,stdin=subprocess.PIPE)
  15         # now do the plots
  16         for graphfile in graphs:
  17                 # set output 'filename'
  18                 p.stdin.write("set title 'Timing of each "+graphtitle+"'\n")
  19                 p.stdin.write("set output '"+gendir+graphfile+"'\n")
  20                 # plot 'filename' using 1:2 title 'algorithm' with lines, 'filename2' using 1:2 title 'algo2' with lines
  21                 plots = []
  22                 for title, file, cols in graphs[graphfile]:
  23                         plots.append("'"+gendir+file+"' using "+cols+" title '"+title+"' with lines")
  24                 plotcommand = ", ".join(plots)
  25                 #print gendir+graphfile
  26                 p.stdin.write("plot "+plotcommand+"\n")
  27         p.communicate()[0]
  28         p.stdin.close()
  29
  30 statsdir = 'stats/raw/'
  31 runs = os.listdir(statsdir)
  32 #print runs
  33
  34 bigrunlist = []
  35 for i in runs:
  36         runfiles = os.listdir(statsdir+i)
  37 #       print runfiles
  38         for file in runfiles:
  39 #               print "File: %s%s/%s" % (statsdir, i, file)
  40                 splitfilename = file.replace('hillis_', 'hillis-').split('_')
  41                 splitfilename[1] = splitfilename[1].replace('hillis-', 'hillis_')
  42                 dict = {}
  43                 dict['scheduler'] = splitfilename[0]
  44                 dict['algo'] = splitfilename[1]
  45                 dict['n'] = int(splitfilename[2].lstrip('n'))
  46                 fd = open(statsdir+i+'/'+file, "r")
  47                 fd_text = fd.read()
  48                 fd.close()
  49                 fd_lines = fd_text.split("\n")
  50                 for line in fd_lines:
  51                         if line == "":
  52                                 continue
  53                         elements = line.split(" ")
  54                         for elem in elements:
  55                                 (key, value) = elem.split("=", 1)
  56                                 if key == 'binname':
  57                                         (_, prog) = value.split("/", 1)
  58                                         if prog != dict['algo']:
  59                                                 print "Failed: %s is not %s" % (prog, dict['algo'])
  60                                 elif key == 'size':
  61                                         size = int(value)
  62                                         if size != dict['n']:
  63                                                 print "Failed: %s is not %s" % (size, dict['n'])
  64                                 elif key == 'time':
  65                                         time = float(value)
  66                         dict['time'] = time
  67                         bigrunlist.append(dict)
  68
  69 algos = {}
  70 schedulers = {}
  71 ns = {}
  72 foolist = {}
  73 for i in bigrunlist:
  74         if not (i['algo'], i['scheduler'], i['n']) in foolist:
  75                 foolist[(i['algo'], i['scheduler'], i['n'])] = [i['time']]
  76         else:
  77                 foolist[(i['algo'], i['scheduler'], i['n'])].append(i['time'])
  78         if not i['algo'] in algos:
  79                 algos[i['algo']] = {}
  80         if not i['scheduler'] in schedulers:
  81                 schedulers[i['scheduler']] = {}
  82         if not i['n'] in ns:
  83                 ns[i['n']] = {}
  84 #print "Algos: %s" % algos
  85 #print "Schedulers: %s" % schedulers
  86 #print "Ns: %s" % ns
  87 uniquelist = []
  88 for key in foolist:
  89         minval = min(foolist[key])
  90         uniquelist.append((key[0], key[1], key[2], minval))
  91 uniquelist.sort()
  92 alg_dict = {}
  93 sched_dict = {}
  94 for i in uniquelist:
  95         if i[0] not in alg_dict:
  96                 alg_dict[i[0]] = {}
  97         if i[1] not in alg_dict[i[0]]:
  98                 alg_dict[i[0]][i[1]] = []
  99         alg_dict[i[0]][i[1]].append(str(i[2])+";"+str(i[3]))
 100         if i[1] not in sched_dict:
 101                 sched_dict[i[1]] = {}
 102         if i[0] not in sched_dict[i[1]]:
 103                 sched_dict[i[1]][i[0]] = []
 104         sched_dict[i[1]][i[0]].append(str(i[2])+";"+str(i[3]))
 105         #print "od["+i[0]+"]["+i[1]+"] = "+str(i[2])+":"+str(i[3])
 106
 107 graphs_alg = {}
 108 graphs_sched = {}
 109 gendir = 'stats/generated'
 110 for alg in alg_dict:
 111         pngfile = 'algorithm.'+alg+'.png'
 112         graphs_alg[pngfile] = []
 113         for sched in alg_dict[alg]:
 114                 datfile = 'data.'+alg+'.'+sched+'.dat'
 115                 fh = open(gendir+'/'+datfile, 'a+')
 116                 graphs_alg[pngfile].append((sched, datfile, '1:2'))
 117                 for dot in alg_dict[alg][sched]:
 118                         fh.write(dot+"\n")
 119                 fh.close()
 120 for sched in sched_dict:
 121         pngfile = 'scheduler.'+sched+'.png'
 122         graphs_sched[pngfile] = []
 123         for alg in sched_dict[sched]:
 124                 datfile = 'data.'+alg+'.'+sched+'.dat'
 125                 graphs_sched[pngfile].append((alg, datfile, '1:2'))
 126
 127 #graphs['test1.png'] = [('linetitle1', 'prefix.hillis.dat', '1:2')]
 128
 129 makeGraphs(graphs_alg, "algorithm")
 130 makeGraphs(graphs_sched, "scheduler")