Analyze git-shadow repos

gistfile1.py

#!/usr/bin/env python
#
# A quick hack to calculate time spent coding on a project that is tracked 
# with `git-shadow`
# 
# Jonathan Fote
# http://foote.pub
# 2014-02-12
#

import subprocess, os, sys
from datetime import datetime, timedelta

if __name__ == "__main__":


    # Gather timestamps for all shadow repos
    # Post: timestamps is a list of all commits that match the user-supplied
    # git-log incantation

    timestamps = []
    dirpath = ".shadow"
    git_args = sys.argv[1:]
    print "git_args", git_args
    for d in os.listdir(dirpath):
        d = os.path.join(dirpath, d)
        if not os.path.isdir(d):
            continue
        try: 
            for line in  subprocess.check_output(["git", "log", "--format=%ct"] + git_args, cwd=d).splitlines():
                if not line:
                    continue
                import re
                m = re.match("^[\d]+$", line)
                if not m:
                    continue
                timestamp = datetime.fromtimestamp(int(line.strip()))
                timestamps.append(timestamp)
        except subprocess.CalledProcessError:
            sys.stderr.write("git log failed for:" + d)
        

    # Calculate work intervals -- assumes more than 5 minutes w/o coding is 
    # a break
    # Post: intervals is a list of non-overlapping coding intervals (start
    # and end timestamps) 

    intervals = []
    timestamps = sorted(timestamps)
    start = timestamps[0]
    for i in range(1, len(timestamps)):
        if ((timestamps[i] - timestamps[i-1]) > timedelta(minutes=5)):
            if start == timestamps[i-1]:
                # If this was a single-line edit, count it as 5 seconds
                intervals.append([start, start+timedelta(seconds=5)])
            else:
                intervals.append([start, timestamps[i-1]])
            start = timestamps[i]

    total = timedelta(0)
    by_date = {}
    for interval in intervals:
        diff = interval[1] - interval[0]
        #print "Coded from " + interval[0].strftime("%a, %d %b %Y %H:%M:%S") + \
        #    " to " + interval[1].strftime("%a, %d %b %Y %H:%M:%S") + "(%s)" % diff 
        total += diff

        by_date[interval[0].strftime("%Y-%m-%d %A")] = \
            by_date.get(interval[0].strftime("%Y-%m-%d %A"), timedelta(0)) + diff 

    for datestr, td in sorted(by_date.items()):
        print datestr, "\t:", td

    print "*"*8
    print "Total time coding (days, H:M:S):", total
    print "Total commits:", len(timestamps)


    import json
    for_tinychart = {"labels":[], "datasets":[{"data":[]}]}
    for datestr, td in sorted(by_date.items()):
        for_tinychart["labels"].append(datestr)
        for_tinychart["datasets"][0]["data"].append(td.total_seconds())
    print "for tinychart:"
    print json.dumps(for_tinychart)