Load a svn xml logfile into a sqlite database for easier analysis.

svnlog2sqlite.py

import xml.etree.ElementTree as ET
import datetime
import sqlite3
import sys

def convert(logfile_filename, database_filename):
    cnt = 0

    # open file handles
    with open(logfile_filename, "rb") as logfile:
        with sqlite3.connect(database_filename) as db:
            cursor = db.cursor()

            # create schema
            cursor.execute("""
                create table commits (
                    revision integer primary key,
                    author text not null,
                    date integer not null,
                    message text null
                )
            """)
            db.commit()

            # parse logfile
            tree = ET.parse(logfile)
            log = tree.getroot()
            assert log.tag == "log"

            # iterate over commits
            for logentry in log:
                assert logentry.tag == "logentry"

                # extract information
                revision = logentry.get("revision")
                assert revision is not None
                revision = int(revision)

                author = logentry.find("author")
                assert author is not None
                author = author.text

                date = logentry.find("date")
                assert date is not None
                date = datetime.datetime.strptime(
                    date.text, "%Y-%m-%dT%H:%M:%S.%fZ")

                msg = logentry.find("msg")
                assert msg is not None
                msg = msg.text
                if msg is not None and len(msg) == 0:
                    msg = None

                # insert database record
                cursor.execute("""
                    insert into commits (revision, author, date, message)
                    values (?, ?, ?, ?)
                """, (revision, author, date, msg))
                cnt += 1

            # commit inserted records
            db.commit()

    return cnt

def main():
    if len(sys.argv) != 3:
        print "usage %s <svn xml logfile> <sqlite database>" % (sys.argv[0],)
        sys.exit(1)

    cnt = convert(sys.argv[1], sys.argv[2])
    print "converted %d entries" % (cnt,)

if __name__ == "__main__":
    main()