convert google location data from google takeout to owntracks native format

Slightly modified from https://www.technowizardry.net/2024/01/migrating-from-google-location-history-to-owntracks/

Script to convert google location takeout to owntracks native format. No need for a MQTT connection, just dump the created files into the owntracks-recorder data dir.

#!/usr/bin/python
import pandas as pd
import json
import time

# https://owntracks.org/booklet/features/tid/
tracker_id = 'ex' # A two-character identifier
df_gps = pd.read_json('Records.json', typ='frame', orient='records')
print('There are {:,} rows in the location history dataset'.format(len(df_gps)))

df_gps = df_gps.apply(lambda x: x['locations'], axis=1, result_type='expand')
df_gps['latitudeE7'] = df_gps['latitudeE7'] / 10.**7
df_gps['longitudeE7'] = df_gps['longitudeE7'] / 10.**7

# In some cases, Google appears to generate timestamps in two different formats.
# Here we convert both types, then normalize it all into the timestamp Pandas dtype.
try:
    df_gps.loc[df_gps['timestamp'].str.len() == len('2013-12-16T05:42:25.711Z'), 'timestamp'] = pd.to_datetime(df_gps['timestamp'], format='mixed')
except ValueError:
    df_gps.loc[df_gps['timestamp'].str.len() == len('2013-12-16T05:42:25Z'), 'timestamp'] = pd.to_datetime(df_gps['timestamp'], format='%Y-%m-%dT%H:%M:%S%Z', utc=True)

df_gps['timestamp'] = pd.to_datetime(df_gps['timestamp'])

owntracks = df_gps.rename(columns={'latitudeE7': 'lat', 'longitudeE7': 'lon', 'accuracy': 'acc', 'altitude': 'alt', 'verticalAccuracy': 'vac'})
owntracks['tst'] = (owntracks['timestamp'].astype(int) / 10**9)

files = {}

years = df_gps['timestamp'].dt.year.agg(['min', 'max'])

for year in range(years['min'], years['max'] + 1):
    for month in range(1, 13):
        files[f"{year}-{month}"] = open(f"location/{year}-{str(month).rjust(2, '0')}.rec", 'w')

try:
    for index, row in owntracks.iterrows():
        d = row.to_dict()
        record = {
            '_type': 'location',
            'tid': tracker_id
        }
        record['tst'] = int(time.mktime(d['timestamp'].timetuple()))

        for key in ['lat', 'lon']:
            if key in row and not pd.isnull(row[key]):
                record[key] = row[key]
        for key in ['acc', 'alt', 'vac']:
            if key in row and not pd.isnull(row[key]):
                record[key] = int(row[key])

        timestamp = row['timestamp'].strftime("%Y-%m-%dT%H:%M:%SZ")
        line = f"{timestamp}\t*                 \t{json.dumps(record, separators=(',', ':'))}\n"
        files[f"{d['timestamp'].year}-{d['timestamp'].month}"].write(line)
finally:
    for key, file in files.items():
        file.flush()
        file.close()