Extracts duration from gpx and tcx files for further processing

activity_extractor.py

#!/usr/local/bin/python
# encoding: utf-8
"""
activity_extractor.py
Created by Christian Stade-Schuldt on 2013-10-29.
"""

import os
from lxml import etree as et
from datetime import timedelta, datetime

PATH = "/Users/stade/Documents/Dropbox/Eigene Dateien/GPS-Tracks/Processed"

def extract_period(activity_file):
    """Takes .gpx or .tcx file and returns list
    of tuples withdatetime object and activity type"""

    # files are organized activity_type/year/file.{gpx,tcx}
    activity_type = activity_file.split("/")[-3]
    root = et.parse(activity_file).getroot()

    extension = activity_file.split(".")[-1]
    if extension == "gpx":
        times = root.findall(".//{http://www.topografix.com/GPX/1/1}time")
        start = datetime.strptime(times[0].text, "%Y-%m-%dT%H:%M:%SZ").replace(second=0, microsecond=0)
        end = datetime.strptime(times[-1].text, "%Y-%m-%dT%H:%M:%SZ").replace(second=0, microsecond=0)

    if extension == "tcx":
        #tcx files use milliseconds
        times = root.findall(".//{http://www.garmin.com/xmlschemas/TrainingCenterDatabase/v2}Time")
        start = datetime.strptime(times[0].text, "%Y-%m-%dT%H:%M:%S.%fZ").replace(second=0, microsecond=0)
        end = datetime.strptime(times[-1].text, "%Y-%m-%dT%H:%M:%S.%fZ").replace(second=0, microsecond=0)

    #get longitude for time-offset
    lon = root.find(".//{http://www.topografix.com/GPX/1/1}trkpt")
    if lon != None:
        offset = int(round(float(lon.attrib.get('lon'))  * 24 / 360))
    else:
        offset = 1 # assume Central European Time if no longitude is present

    #apply daylight saving time, roughly
    if start.month in range(4, 11):
        offset += 1

    start = start + timedelta(hours=offset)
    end = end + timedelta(hours=offset)

    timeframe = []
    while end-start > timedelta(0):
        timeframe.append((start, activity_type))
        start = start - timedelta(minutes=-1)
    return timeframe

def main():
    total_times = []
    for root, files in os.walk(PATH):
        for file_name in files:
            if file_name.endswith("x"):
                total_times.extend(extract_period(os.path.join(root, file_name)))

    total_times = [(str(d[0].strftime("%Y-%m-%d %H:%M:%S")), d[1]) for d in total_times]

    w = open("output.csv", "w")
    w.write("Timestamp, Activity\n")
    for t in total_times:
        w.write("%s, %s\n" %(t[0], t[1]))
    w.close()

if __name__ == '__main__':
    main()