support/scripts/size-stats

   1 #!/usr/bin/env python
   2
   3 # Copyright (C) 2014 by Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
   4
   5 # This program is free software; you can redistribute it and/or modify
   6 # it under the terms of the GNU General Public License as published by
   7 # the Free Software Foundation; either version 2 of the License, or
   8 # (at your option) any later version.
   9 #
  10 # This program is distributed in the hope that it will be useful,
  11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  13 # General Public License for more details.
  14 #
  15 # You should have received a copy of the GNU General Public License
  16 # along with this program; if not, write to the Free Software
  17 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
  18
  19 import sys
  20 import os
  21 import os.path
  22 import argparse
  23 import csv
  24 import collections
  25
  26 try:
  27     import matplotlib
  28     matplotlib.use('Agg')
  29     import matplotlib.font_manager as fm
  30     import matplotlib.pyplot as plt
  31 except ImportError:
  32     sys.stderr.write("You need python-matplotlib to generate the size graph\n")
  33     exit(1)
  34
  35 colors = ['#e60004', '#009836', '#2e1d86', '#ffed00',
  36           '#0068b5', '#f28e00', '#940084', '#97c000']
  37
  38 #
  39 # This function adds a new file to 'filesdict', after checking its
  40 # size. The 'filesdict' contain the relative path of the file as the
  41 # key, and as the value a tuple containing the name of the package to
  42 # which the file belongs and the size of the file.
  43 #
  44 # filesdict: the dict to which  the file is added
  45 # relpath: relative path of the file
  46 # fullpath: absolute path to the file
  47 # pkg: package to which the file belongs
  48 #
  49 def add_file(filesdict, relpath, abspath, pkg):
  50     if not os.path.exists(abspath):
  51         return
  52     if os.path.islink(abspath):
  53         return
  54     sz = os.stat(abspath).st_size
  55     filesdict[relpath] = (pkg, sz)
  56
  57 #
  58 # This function returns a dict where each key is the path of a file in
  59 # the root filesystem, and the value is a tuple containing two
  60 # elements: the name of the package to which this file belongs and the
  61 # size of the file.
  62 #
  63 # builddir: path to the Buildroot output directory
  64 #
  65 def build_package_dict(builddir):
  66     filesdict = {}
  67     with open(os.path.join(builddir, "build", "packages-file-list.txt")) as filelistf:
  68         for l in filelistf.readlines():
  69             pkg, fpath = l.split(",", 1)
  70             # remove the initial './' in each file path
  71             fpath = fpath.strip()[2:]
  72             fullpath = os.path.join(builddir, "target", fpath)
  73             add_file(filesdict, fpath, fullpath, pkg)
  74     return filesdict
  75
  76 #
  77 # This function builds a dictionary that contains the name of a
  78 # package as key, and the size of the files installed by this package
  79 # as the value.
  80 #
  81 # filesdict: dictionary with the name of the files as key, and as
  82 # value a tuple containing the name of the package to which the files
  83 # belongs, and the size of the file. As returned by
  84 # build_package_dict.
  85 #
  86 # builddir: path to the Buildroot output directory
  87 #
  88 def build_package_size(filesdict, builddir):
  89     pkgsize = collections.defaultdict(int)
  90
  91     seeninodes = set()
  92     for root, _, files in os.walk(os.path.join(builddir, "target")):
  93         for f in files:
  94             fpath = os.path.join(root, f)
  95             if os.path.islink(fpath):
  96                 continue
  97
  98             st = os.stat(fpath)
  99             if st.st_ino in seeninodes:
 100                 # hard link
 101                 continue
 102             else:
 103                 seeninodes.add(st.st_ino)
 104
 105             frelpath = os.path.relpath(fpath, os.path.join(builddir, "target"))
 106             if not frelpath in filesdict:
 107                 print("WARNING: %s is not part of any package" % frelpath)
 108                 pkg = "unknown"
 109             else:
 110                 pkg = filesdict[frelpath][0]
 111
 112             pkgsize[pkg] += st.st_size
 113
 114     return pkgsize
 115
 116 #
 117 # Given a dict returned by build_package_size(), this function
 118 # generates a pie chart of the size installed by each package.
 119 #
 120 # pkgsize: dictionary with the name of the package as a key, and the
 121 # size as the value, as returned by build_package_size.
 122 #
 123 # outputf: output file for the graph
 124 #
 125 def draw_graph(pkgsize, outputf):
 126     total = sum(pkgsize.values())
 127     labels = []
 128     values = []
 129     other_value = 0
 130     for (p, sz) in pkgsize.items():
 131         if sz < (total * 0.01):
 132             other_value += sz
 133         else:
 134             labels.append("%s (%d kB)" % (p, sz / 1000.))
 135             values.append(sz)
 136     labels.append("Other (%d kB)" % (other_value / 1000.))
 137     values.append(other_value)
 138
 139     plt.figure()
 140     patches, texts, autotexts = plt.pie(values, labels=labels,
 141                                         autopct='%1.1f%%', shadow=True,
 142                                         colors=colors)
 143     # Reduce text size
 144     proptease = fm.FontProperties()
 145     proptease.set_size('xx-small')
 146     plt.setp(autotexts, fontproperties=proptease)
 147     plt.setp(texts, fontproperties=proptease)
 148
 149     plt.suptitle("Filesystem size per package", fontsize=18, y=.97)
 150     plt.title("Total filesystem size: %d kB" % (total / 1000.), fontsize=10, y=.96)
 151     plt.savefig(outputf)
 152
 153 #
 154 # Generate a CSV file with statistics about the size of each file, its
 155 # size contribution to the package and to the overall system.
 156 #
 157 # filesdict: dictionary with the name of the files as key, and as
 158 # value a tuple containing the name of the package to which the files
 159 # belongs, and the size of the file. As returned by
 160 # build_package_dict.
 161 #
 162 # pkgsize: dictionary with the name of the package as a key, and the
 163 # size as the value, as returned by build_package_size.
 164 #
 165 # outputf: output CSV file
 166 #
 167 def gen_files_csv(filesdict, pkgsizes, outputf):
 168     total = 0
 169     for (p, sz) in pkgsizes.items():
 170         total += sz
 171     with open(outputf, 'w') as csvfile:
 172         wr = csv.writer(csvfile, delimiter=',', quoting=csv.QUOTE_MINIMAL)
 173         wr.writerow(["File name",
 174                      "Package name",
 175                      "File size",
 176                      "Package size",
 177                      "File size in package (%)",
 178                      "File size in system (%)"])
 179         for f, (pkgname, filesize) in filesdict.items():
 180             pkgsize = pkgsizes[pkgname]
 181             wr.writerow([f, pkgname, filesize, pkgsize,
 182                          "%.1f" % (float(filesize) / pkgsize * 100),
 183                          "%.1f" % (float(filesize) / total * 100)])
 184
 185
 186 #
 187 # Generate a CSV file with statistics about the size of each package,
 188 # and their size contribution to the overall system.
 189 #
 190 # pkgsize: dictionary with the name of the package as a key, and the
 191 # size as the value, as returned by build_package_size.
 192 #
 193 # outputf: output CSV file
 194 #
 195 def gen_packages_csv(pkgsizes, outputf):
 196     total = sum(pkgsizes.values())
 197     with open(outputf, 'w') as csvfile:
 198         wr = csv.writer(csvfile, delimiter=',', quoting=csv.QUOTE_MINIMAL)
 199         wr.writerow(["Package name", "Package size", "Package size in system (%)"])
 200         for (pkg, size) in pkgsizes.items():
 201             wr.writerow([pkg, size, "%.1f" % (float(size) / total * 100)])
 202
 203 parser = argparse.ArgumentParser(description='Draw size statistics graphs')
 204
 205 parser.add_argument("--builddir", '-i', metavar="BUILDDIR", required=True,
 206                     help="Buildroot output directory")
 207 parser.add_argument("--graph", '-g', metavar="GRAPH",
 208                     help="Graph output file (.pdf or .png extension)")
 209 parser.add_argument("--file-size-csv", '-f', metavar="FILE_SIZE_CSV",
 210                     help="CSV output file with file size statistics")
 211 parser.add_argument("--package-size-csv", '-p', metavar="PKG_SIZE_CSV",
 212                     help="CSV output file with package size statistics")
 213 args = parser.parse_args()
 214
 215 # Find out which package installed what files
 216 pkgdict = build_package_dict(args.builddir)
 217
 218 # Collect the size installed by each package
 219 pkgsize = build_package_size(pkgdict, args.builddir)
 220
 221 if args.graph:
 222     draw_graph(pkgsize, args.graph)
 223 if args.file_size_csv:
 224     gen_files_csv(pkgdict, pkgsize, args.file_size_csv)
 225 if args.package_size_csv:
 226     gen_packages_csv(pkgsize, args.package_size_csv)