Simple way to create a clean changelog between two commits. Allows changelog filtering to only include commits containing a substring or regular expression.

gen-changelog.py

#!/usr/bin/env python

import sys
import re
from optparse import OptionParser
from subprocess import Popen, PIPE

class GitChangelog(object):
    def __init__(self, from_ref=None, to_ref='HEAD', log_filter=None):
        self.from_ref = from_ref
        self.to_ref = to_ref
        self.log_filter = log_filter

    def get_log(self):
        raw_log = self._generate_log()
        log_entries = self._parse_log(raw_log)
        pretty_log = []
        for author in sorted(log_entries.iterkeys()):
            if log_entries[author]:
                pretty_log.append("%s:" % author)
                pretty_log += log_entries[author]
                pretty_log.append("")
        return "\n".join(pretty_log)
        
    def _parse_log(self, raw_log):
        log_entries = {}
        log_lines = raw_log.splitlines()

        current_author = None
        for line in log_lines:
            if re.search('^\w+', line):
                current_author = re.search('^([\w\s-]+)\s', line).groups()[0]
                log_entries[current_author] = []
            else:
                if current_author not in log_entries:
                    raise ValueError("author '%s' does not exist" % current_author)
                if self.log_filter:
                    if re.search(self.log_filter, line):
                        log_entries[current_author].append(line)
                else:
                    log_entries[current_author].append(line)
        return log_entries

    def _generate_log(self):
        delta = None
        if self.from_ref is None:
            delta = self.to_ref
        else:
            delta = "%s..%s" % (self.from_ref, self.to_ref)
        p = Popen(["git", "shortlog", delta], stdin=PIPE, stdout=PIPE)
        raw_log = p.communicate()[0]
        if p.returncode != 0:
            raise Exception("Failed to execute git")
        return raw_log

    def _strip_line(self, data, length=80):
        data = data.lstrip()
        return (data[:length] + '..') if len(data) > length else data


if __name__ == "__main__":
    parser = OptionParser()
    parser.add_option("-f", "--from", dest="from_ref",
                      help="changelog starts at this reference")
    parser.add_option("-t", "--to", dest="to_ref", default="HEAD",
                      help="changelog ends at this reference")
    parser.add_option("-l", "--log-filter", dest="log_filter",
                      help="take changes that include this string")

    (opts, args) = parser.parse_args()

    changelog = GitChangelog(from_ref=opts.from_ref, to_ref=opts.to_ref,
                             log_filter=opts.log_filter)

    sys.stdout.write(changelog.get_log())
    sys.stdout.flush()

Example usage:

./gen-changelog.py -f 3.3-m5 -t 3.3-m6 -l EUCA-\\d+ > changelog.txt