A script to analyze a TMX file and try to figure out how much time has been spent translating.

TMXTimekeeper.py

'''
TMXTimekeeper.py

Analyze a TMX file and try to figure out how much time
has been spent translating. Assume a minimum of 5 minutes
for translating "sessions".

Created on 2013/02/17

@author: Aaron Madlon-Kay
'''

import sys
import os
from xml.etree import ElementTree
from datetime import datetime, timedelta, tzinfo
import time

BIN_DELTA = timedelta(hours=1)
MIN_WORK_TIME = timedelta(minutes=5)


class TimeZone(tzinfo):

    def __init__(self, offset, name=None):
        self.__offset = offset
        self.__name = name if name != None else "UTC+{0}".format(offset)

    def utcoffset(self, dt):
        return timedelta(seconds=self.__offset)

    def dst(self, dt):
        return timedelta(0)

    def tzname(self):
        return self.__name


# From StackExchange: http://stackoverflow.com/a/13406277/448068
def local_time_offset(t=None):
    """Return offset of local zone from GMT, either at present or at time t."""
    # python2.3 localtime() can't take None
    if t is None:
        t = time.time()

    if time.localtime(t).tm_isdst and time.daylight:
        return -time.altzone
    else:
        return -time.timezone

UTC = TimeZone(0)
USER_TZ = TimeZone(local_time_offset())


def get_dates(tree):
    '''
    Extract translation timestamps as //tuv@changedate or //tuv@creationdate
    strings, and return them as a sorted list of datetimes.
    '''
    result = []
    for tuv in tree.findall(".//tuv"):
        creation_date = tuv.get("creationdate")
        change_date = tuv.get("changedate")
        try:
            if creation_date != None:
                dt = datetime.strptime(creation_date, "%Y%m%dT%H%M%SZ")
                result.append(dt.replace(tzinfo=UTC))
            if change_date != None and change_date != creation_date:
                dt = datetime.strptime(change_date, "%Y%m%dT%H%M%SZ")
                result.append(dt.replace(tzinfo=UTC))
        except ValueError, e:
            pass
    return sorted(result)


def bin_dates(dates):
    '''
    Given a sorted list of datetimes, reduce to a set of "bins"
    (start and end times), defined as a period of time where no two
    datetimes are separated by more than BIN_DELTA.
    '''
    result = {}
    if len(dates) < 1:
        return result
    start = dates[0]
    result[start] = None
    for date in dates:
        end = result[start]
        if end == None:
            end = start
        if date - end < BIN_DELTA:
            result[start] = date
        else:
            start = date
            result[start] = None
    return result


def print_bins(bins):
    '''
    Given a dictionary of bins, print the date, start and end
    times, and duration. Finally, print the total time covered
    by all bins.
    '''
    total = timedelta(0)
    master_adjusted = ""
    for start, end in sorted(bins.items()):
        local_start = start.astimezone(USER_TZ)
        print "{0}: {1} -".format(local_start.strftime("%Y/%m/%d"),
                                  local_start.strftime("%H:%M")),
        adjusted = ""
        # A bin may have end == None if just one TUV was changed.
        if end == None:
            end = start
        local_end = end.astimezone(USER_TZ)
        delta = local_end - local_start
        # Ensure minimum work time
        if delta < MIN_WORK_TIME:
            delta = MIN_WORK_TIME
            local_end = local_start + MIN_WORK_TIME
            adjusted = "*"
            master_adjusted = "*"
        total += delta
        print local_end.strftime("%H:%M"), " ({0}{1})".format(str(delta),
                                                              adjusted)

    print "Total: {0}{1}".format(str(total), master_adjusted)

if __name__ == "__main__":

    if len(sys.argv) < 2 or not os.path.isfile(sys.argv[1]) \
    or not sys.argv[1].endswith(".tmx"):
        print "Please specify a valid input TMX"
        sys.exit(1)

    try:
        tree = ElementTree.parse(sys.argv[1])
        dates = get_dates(tree)
        bins = bin_dates(dates)
        print_bins(bins)
    except ElementTree.ParseError:
        print "Invalid input TMX"
        sys.exit(1)