1 # churn.py - create a graph showing who changed the most lines

     2 #

     3 # Copyright 2006 Josef "Jeff" Sipek <jeffpc@josefsipek.net>

     4 #

     5 # This software may be used and distributed according to the terms

     6 # of the GNU General Public License, incorporated herein by reference.

     7 #

     8 #

     9 # Aliases map file format is simple one alias per line in the following

    10 # format:

    11 #

    12 # <alias email> <actual email>

    13 

    14 from mercurial.demandload import *

    15 from mercurial.i18n import gettext as _

    16 demandload(globals(), 'time sys signal os')

    17 demandload(globals(), 'mercurial:hg,mdiff,fancyopts,commands,ui,util,templater,node')

    18 

    19 def __gather(ui, repo, node1, node2):

    20     def dirtywork(f, mmap1, mmap2):

    21         lines = 0

    22 

    23         to = mmap1 and repo.file(f).read(mmap1[f]) or None

    24         tn = mmap2 and repo.file(f).read(mmap2[f]) or None

    25 

    26         diff = mdiff.unidiff(to, "", tn, "", f).split("\n")

    27 

    28         for line in diff:

    29             if not line:

    30                 continue # skip EOF

    31             if line.startswith(" "):

    32                 continue # context line

    33             if line.startswith("--- ") or line.startswith("+++ "):

    34                 continue # begining of diff

    35             if line.startswith("@@ "):

    36                 continue # info line

    37 

    38             # changed lines

    39             lines += 1

    40 

    41         return lines

    42 

    43     ##

    44 

    45     lines = 0

    46 

    47     changes = repo.status(node1, node2, None, util.always)[:5]

    48 

    49     modified, added, removed, deleted, unknown = changes

    50 

    51     who = repo.changelog.read(node2)[1]

    52     who = templater.email(who) # get the email of the person

    53 

    54     mmap1 = repo.manifest.read(repo.changelog.read(node1)[0])

    55     mmap2 = repo.manifest.read(repo.changelog.read(node2)[0])

    56     for f in modified:

    57         lines += dirtywork(f, mmap1, mmap2)

    58 

    59     for f in added:

    60         lines += dirtywork(f, None, mmap2)

    61         

    62     for f in removed:

    63         lines += dirtywork(f, mmap1, None)

    64 

    65     for f in deleted:

    66         lines += dirtywork(f, mmap1, mmap2)

    67 

    68     for f in unknown:

    69         lines += dirtywork(f, mmap1, mmap2)

    70 

    71     return (who, lines)

    72 

    73 def gather_stats(ui, repo, amap, revs=None, progress=False):

    74     stats = {}

    75     

    76     cl    = repo.changelog

    77 

    78     if not revs:

    79         revs = range(0, cl.count())

    80 

    81     nr_revs = len(revs)

    82     cur_rev = 0

    83 

    84     for rev in revs:

    85         cur_rev += 1 # next revision

    86 

    87         node2    = cl.node(rev)

    88         node1    = cl.parents(node2)[0]

    89 

    90         if cl.parents(node2)[1] != node.nullid:

    91             ui.note(_('Revision %d is a merge, ignoring...\n') % (rev,))

    92             continue

    93 

    94         who, lines = __gather(ui, repo, node1, node2)

    95 

    96         # remap the owner if possible

    97         if amap.has_key(who):

    98             ui.note("using '%s' alias for '%s'\n" % (amap[who], who))

    99             who = amap[who]

   100 

   101         if not stats.has_key(who):

   102             stats[who] = 0

   103         stats[who] += lines

   104 

   105         ui.note("rev %d: %d lines by %s\n" % (rev, lines, who))

   106 

   107         if progress:

   108             if int(100.0*(cur_rev - 1)/nr_revs) < int(100.0*cur_rev/nr_revs):

   109                 ui.write("%d%%.." % (int(100.0*cur_rev/nr_revs),))

   110                 sys.stdout.flush()

   111 

   112     if progress:

   113         ui.write("done\n")

   114         sys.stdout.flush()

   115 

   116     return stats

   117 

   118 def churn(ui, repo, **opts):

   119     "Graphs the number of lines changed"

   120     

   121     def pad(s, l):

   122         if len(s) < l:

   123             return s + " " * (l-len(s))

   124         return s[0:l]

   125 

   126     def graph(n, maximum, width, char):

   127         n = int(n * width / float(maximum))

   128         

   129         return char * (n)

   130 

   131     def get_aliases(f):

   132         aliases = {}

   133 

   134         for l in f.readlines():

   135             l = l.strip()

   136             alias, actual = l.split(" ")

   137             aliases[alias] = actual

   138 

   139         return aliases

   140     

   141     amap = {}

   142     aliases = opts.get('aliases')

   143     if aliases:

   144         try:

   145             f = open(aliases,"r")

   146         except OSError, e:

   147             print "Error: " + e

   148             return

   149 

   150         amap = get_aliases(f)

   151         f.close()

   152 

   153     revs = [int(r) for r in commands.revrange(ui, repo, opts['rev'])]

   154     revs.sort()

   155     stats = gather_stats(ui, repo, amap, revs, opts.get('progress'))

   156 

   157     # make a list of tuples (name, lines) and sort it in descending order

   158     ordered = stats.items()

   159     ordered.sort(cmp=lambda x, y: cmp(y[1], x[1]))

   160 

   161     maximum = ordered[0][1]

   162 

   163     ui.note("Assuming 80 character terminal\n")

   164     width = 80 - 1

   165 

   166     for i in ordered:

   167         person = i[0]

   168         lines = i[1]

   169         print "%s %6d %s" % (pad(person, 20), lines,

   170                 graph(lines, maximum, width - 20 - 1 - 6 - 2 - 2, '*'))

   171 

   172 cmdtable = {

   173     "churn":

   174     (churn,

   175      [('r', 'rev', [], _('limit statistics to the specified revisions')),

   176       ('', 'aliases', '', _('file with email aliases')),

   177       ('', 'progress', None, _('show progress'))],

   178     'hg churn [-r revision range] [-a file] [--progress]'),

   179 }