iSki tracker exporter

iski-convert.py

#!/usr/bin/python
# coding=UTF-8

"""
Convert html saved from iSki tracker website to gpx format to import
any application or service such QLandkarteGT, Endomondo, etc.

Need to install the following packages:

$ pip install pytz
$ pip install tzlocal

Usage:

$ iski-convert.py <input-file>

Background:

There is no export function on iSki tracker website, but the html source
contains the date, coordinates, profile in different format. This script 
converts it to gpx that is a common GPS data format for software 
applications.

The script use the locale timezone and convert the dates to UTC, because
the GPS format defines date in Coordinated Universal Time (UTC) using 
ISO 8601 format. The Z at the end of the dates is the zone designator for 
the zero UTC offset. 
"""

import sys
import re
from datetime import datetime, timedelta, date, time

import pytz # $ pip install pytz
from tzlocal import get_localzone # $ pip install tzlocal

import locale

def parsecoordinates(line):
	# new google.maps.LatLng(46.68102, 13.89978)
	coordinates = []
	pattern = re.compile(r"LatLng\((\d+\.\d+),\ (\d+\.\d+)\)")
	for m in re.finditer(pattern, line):	
		pair = (float(m.group(1)), float(m.group(2)))
		coordinates.append(pair)
	return coordinates

def parseprofile(line, basedate):
	# {"x":46211799.99995232,"y":1435.0,"chartDataIndex":0}
	profile = []
	pattern = re.compile(r"\"x\"\:(\d+\.\d+)\,\"y\"\:(\d+\.\d+)")
	for m in re.finditer(pattern, line):	
		pair = (toisoformat(float(m.group(1)), basedate), float(m.group(2)))
		profile.append(pair)
	return profile

def toisoformat(f, basedate):
	local_tz = get_localzone()
	d = datetime(basedate.year, basedate.month, basedate.day, tzinfo = local_tz)
	d = d + timedelta(milliseconds = f)
	d = d.astimezone(pytz.utc)
	return d.strftime("%Y-%m-%dT%H:%M:%SZ")

def parsebasedate(line):
	d = re.search(r"\d{2}\s\w+\s", line).group(0)
	d = str(d) + " " + str(datetime.now().year)
	loc = locale.getlocale()
	locale.setlocale(locale.LC_TIME, "en_GB.utf8")
	basedate = datetime.strptime(d, "%d %b %Y")
	locale.setlocale(locale.LC_TIME, loc)
	return basedate

if len(sys.argv) != 2:
	print "Usage: iski-convert.py <input-file>"
	exit()

f = open(sys.argv[1], 'r')
state = None
for line in f:
	if state == "DAY":
		basedate = parsebasedate(line)		
		state = None		
	if """<span class="caption">DAY</span>""" in line:
		state = "DAY"
	if line.strip().find("new google.maps.LatLng(") >= 0:
		coordinates = parsecoordinates(line.strip())
	if line.strip().find("data: ") >= 0:
		profile = parseprofile(line.strip(), basedate)

print """<?xml version="1.0" encoding="UTF-8" standalone="no" ?>
 
<gpx xmlns="http://www.topografix.com/GPX/1/1" creator="byHand" version="1.1"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.topografix.com/GPX/1/1 http://www.topografix.com/GPX/1/1/gpx.xsd">
	<trk>
		<trkseg>"""

for coordinate,elevation in zip(coordinates, profile):
	print """
			<trkpt lon="%f" lat="%f">
				<ele>%f</ele>
				<time>%s</time>
			</trkpt>""" % (coordinate[1], coordinate[0], elevation[1], elevation[0])

print """
		</trkseg>
	</trk>
</gpx>"""

@vicziani the closest I can find is when you click share on a specific track it references a json file geopath of the format:

{
  "path": [
    {
      "lat": 45.090715,
      "lng": 6.075981,
      "time": 40792000,
      "elevation": 1806
    },
    {
      "lat": 45.092651,
      "lng": 6.075863,
      "time": 41032592,
      "elevation": 1831
    },
    {
      "lat": 45.09284,
      "lng": 6.075512,
      "time": 41033592,
      "elevation": 1829
    },
    ...
}

I guess I'll need to write my own parser for that :(

I have now written the script but the time format is a little crazy. Can't for the life of me figure out how

51603000 -> 13:30:37.722
54575086 -> 14:20:10.811
…
56794776 -> 14:57:08.491
56795766 -> 14:57:10.488

Any experience with that?

I didn't know, but the ChatGPT helped me. It is milliseconds from midnight. So it does not contain the date! Sample code:

def milliseconds_to_time(milliseconds):
    seconds, milliseconds = divmod(milliseconds, 1000)
    minutes, seconds = divmod(seconds, 60)
    hours, minutes = divmod(minutes, 60)
    return hours, minutes, seconds, milliseconds

time_in_milliseconds = 35340000
hours, minutes, seconds, milliseconds = milliseconds_to_time(time_in_milliseconds)

print(f"{hours} hours, {minutes} minutes, {seconds} seconds, {milliseconds} milliseconds")

The graph on Share page shows always Jan 1. :)

ChatGPTs code seems to return for the examples given above:

51603000 -> 14:20:03 not 13:30:37.722
54575086 -> 15:09:35.86 not 14:20:10.811
…
56794776 -> 15:46:34.776 not 14:57:08.491
56795766 -> 15:46:35.766 not 14:57:10.488

I wrote my own parser for the details file, which is at CatsLover2006/iski2GPX. However, I've noticed that the details file has a very limited number of points to work with.