wilieNillieMilageTripCount

import haversine as hs
from haversine import haversine, haversine_vector, Unit
import pandas as pd
import numpy as np
import json

print('hello')

lat=30.393000
lon=-98.070050
loc1=(lat,lon)
Takeout_google_location_history='C:/Users/PC/Downloads/Records.json'
output='C:/Users/PC/Downloads/2022_WillieNillie_geo_location.json'

with open(Takeout_google_location_history) as f:
    data = json.loads(f.read())

df = pd.json_normalize(data['locations'])

print( len(df.index) )
df = df.drop('activity', axis=1)
df = df.drop('batteryCharging', axis=1)
df = df.drop('formFactor', axis=1)
df = df.drop('deviceTimestamp', axis=1)
df = df.drop('serverTimestamp', axis=1)
df = df.drop('osLevel', axis=1)
df = df.drop('velocity', axis=1)
df = df.drop('platformType', axis=1)
df = df.drop('locationMetadata', axis=1)
df = df.drop('verticalAccuracy', axis=1)
df = df.drop('heading', axis=1)
df = df.drop('inferredLocation', axis=1)
df = df.drop('activeWifiScan.accessPoints', axis=1)

df['tstamp'] = pd.to_datetime(df['timestamp'])
df = df[df['tstamp'].dt.strftime('%Y') == '2022']
print( len(df.index) )

df['latitudeE7'] = df['latitudeE7'].div(10000000.0)
df['longitudeE7'] = df['longitudeE7'].div(10000000.0)
print( df.head() )


#df['diff'] = hs.haversine(loc1,(df['latitudeE7'],df['longitudeE7']),unit=Unit.MILES)
df['diff'] = df.apply(lambda row: hs.haversine(loc1, (row['latitudeE7'], row['longitudeE7']), unit=Unit.MILES), axis=1)
print( df.head() )
print( len(df.index) )
df=df[df['diff']<0.5]
print( df.head() )
print( len(df.index) )
print( df )

print( "Create day column" )
df['day'] = df['tstamp'].dt.strftime('%Y-%m-%d')
print( len(df.index) )
print( df )


df['tstamp'] = df['tstamp'].astype(str)

print( "Set index to day column and convert to list" )
dct = df.set_index('day').T.to_dict('list')
print( dct )
print( len(dct) )

with open(output, 'w') as convert_file:
         convert_file.write(json.dumps(dct))