authordensity.py

"""
Most useful as:

    git ls-tree --name-only -r HEAD | xargs python /path/to/authordensity.py
    
When run from the root of your git repository. If people show up under
multiple names, use the synonyms dict to map their aliases to a canonical
name.

Needs no external libs.
"""

from __future__ import division, print_function

import collections
import re
import subprocess
import sys

synonyms = {
}

match_files = re.compile(r"\.(cc|h|cpp|hpp|c|py|pxi|pyx)$")
author_line_re = re.compile("^author ")
count = 30

file_densities = dict()
authors_found = set()
total_freq = collections.defaultdict(lambda: 0)

for f in sys.argv[1:]:
    if not match_files.search(f):
        continue
    try:
        blame = subprocess.check_output(
            ["git", "blame", "--line-porcelain", f], stderr=subprocess.PIPE)
    except subprocess.CalledProcessError as e:
        if e.returncode == 128:
            continue
        raise
    blame_lines = blame.splitlines()
    author_lines = filter(lambda l: author_line_re.match(l), blame_lines)
    authors_freq = collections.defaultdict(lambda: 0)
    for line in author_lines:
        author = line.split(" ", 1)[1]
        if author in synonyms:
            author = synonyms[author]
        authors_found.add(author)
        authors_freq[author] += 1
        total_freq[author] += 1
    total_lines = len(author_lines)
    authors = {author: author_lines / total_lines
               for author, author_lines in authors_freq.iteritems()}
    file_densities[f] = authors

for author in authors_found:
    author_densities = [(file_density.get(author, 0), fname)
                        for fname, file_density in file_densities.iteritems()
                        if file_density.get(author)]
    author_densities.sort(reverse=True)
    print("\n%s" % author)
    print("\n".join("%3.0f%%\t%s" % (100. * s[0], s[1])
                    for s in author_densities[:count]))
    print()

total_lines = sum(total_freq.values())
records = total_freq.items()
records.sort(key=lambda p: (p[1], p[0]), reverse=True)
for author, freq in records:
    print("%s\t%.3f%%\t\t%s" % (freq, 100. * freq / total_lines, author))