comparison hgext/churn.py @ 3037:f74077473b36

Churn extension
author Josef "Jeff" Sipek <jeff@josefsipek.net>
date Wed, 26 Jul 2006 19:42:56 -0400
parents
children 45942bb49194
comparison
equal deleted inserted replaced
2682:4e2dc5c16e61 3037:f74077473b36
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 import time, sys, signal, os
15 from mercurial import hg, mdiff, fancyopts, commands, ui, util, templater
16
17 def __gather(ui, repo, node1, node2):
18 def dirtywork(f, mmap1, mmap2):
19 lines = 0
20
21 to = None
22 if mmap1:
23 to = repo.file(f).read(mmap1[f])
24 tn = None
25 if mmap2:
26 tn = repo.file(f).read(mmap2[f])
27
28 diff = mdiff.unidiff(to, "", tn, "", f).split("\n")
29
30 for line in diff:
31 if len(line) <= 0:
32 continue # skip EOF
33 if line[0] == " ":
34 continue # context line
35 if line[0:4] == "--- " or line[0:4] == "+++ ":
36 continue # begining of diff
37 if line[0:3] == "@@ ":
38 continue # info line
39
40 # changed lines
41 lines += 1
42
43 return lines
44
45 ##
46
47 lines = 0
48
49 changes = repo.changes(node1, node2, None, util.always)
50
51 modified, added, removed, deleted, unknown = changes
52
53 who = repo.changelog.read(node2)[1]
54 who = templater.email(who) # get the email of the person
55
56 mmap1 = repo.manifest.read(repo.changelog.read(node1)[0])
57 mmap2 = repo.manifest.read(repo.changelog.read(node2)[0])
58 for f in modified:
59 lines += dirtywork(f, mmap1, mmap2)
60
61 for f in added:
62 lines += dirtywork(f, None, mmap2)
63
64 for f in removed:
65 lines += dirtywork(f, mmap1, None)
66
67 for f in deleted:
68 lines += dirtywork(f, mmap1, mmap2)
69
70 for f in unknown:
71 lines += dirtywork(f, mmap1, mmap2)
72
73 return (who, lines)
74
75 def gather_stats(ui, repo, amap):
76 stats = {}
77
78 cl = repo.changelog
79
80 for rev in range(1,cl.count()):
81 node2 = cl.node(rev)
82 node1 = cl.parents(node2)[0]
83
84 who, lines = __gather(ui, repo, node1, node2)
85
86 # remap the owner if possible
87 if amap.has_key(who):
88 ui.note("using '%s' alias for '%s'\n" % (amap[who], who))
89 who = amap[who]
90
91 if not stats.has_key(who):
92 stats[who] = 0
93 stats[who] += lines
94
95 ui.note("rev %d: %d lines by %s\n" % (rev, lines, who))
96
97 return stats
98
99 def churn(ui, repo, aliases):
100 "Graphs the number of lines changed"
101
102 def pad(s, l):
103 if len(s) < l:
104 return s + " " * (l-len(s))
105 return s[0:l]
106
107 def graph(n, maximum, width, char):
108 n = int(n * width / float(maximum))
109
110 return char * (n)
111
112 def get_aliases(f):
113 aliases = {}
114
115 for l in f.readlines():
116 l = l.strip()
117 alias, actual = l.split(" ")
118 aliases[alias] = actual
119
120 return aliases
121
122 amap = {}
123 if aliases:
124 try:
125 f = open(aliases,"r")
126 except OSError, e:
127 print "Error: " + e
128 return
129
130 amap = get_aliases(f)
131 f.close()
132
133 os.chdir(repo.root)
134 stats = gather_stats(ui, repo, amap)
135
136 # make a list of tuples (name, lines) and sort it in descending order
137 ordered = stats.items()
138 ordered.sort(cmp=lambda x,y:cmp(x[1], y[1]))
139 ordered.reverse()
140
141 maximum = ordered[0][1]
142
143 ui.note("Assuming 80 character terminal\n")
144 width = 80 - 1
145
146 for i in ordered:
147 person = i[0]
148 lines = i[1]
149 print "%s %6d %s" % (pad(person, 20), lines,
150 graph(lines, maximum, width - 20 - 1 - 6 - 2 - 2, '*'))
151
152 cmdtable = {
153 "churn":
154 (churn,
155 [('', 'aliases', '', 'file with email aliases')],
156 'hg churn [-a file]'),
157 }
158
159 def reposetup(ui, repo):
160 pass
161