parse_google_search_history.py

"""
Reads Google search history JSON files from the given directory.
Expected JSON format :
{"event":[
 {"query":{"id":[{"timestamp_usec":"1135905619017279"}],"query_text":"XYZ"}},
 {"query":{"id":[{"timestamp_usec":"1135903586447380"}],"query_text":"ABC"}},
]}


The folder containing the JSON files is stored in a config.ini file with the section
[google.search.history]
searchesFolder = /path/to/folder

"""
__author__  = 'satish'
__version__ = 0.1

import os
import glob
import json
from datetime import datetime
import ConfigParser
import csv
path='/Users/<Location>/*.json'
print glob.glob(path)

try:
    os.remove('searchlog.csv')
except OSError:
    pass

outfile = open('searchlog.csv','wt')

try:
    fieldnames=('date','weekday','month','hour','min','search_text')
    writer = csv.DictWriter(outfile,fieldnames=fieldnames)
    headers =dict((n,n) for n in fieldnames)
    writer.writerow(headers)
    print
    for infile in glob.glob(path):
        print(infile)
        json_data = open(infile)
        data = json.load(json_data)
        for i in range(0,len(data['event'])):
            search = data['event'][i]['query']['query_text']
            #print search
            when  = datetime.fromtimestamp(int(data['event'][i]['query']['id'][0]['timestamp_usec'])/1000000)
            #print when
            day = when.strftime(("%Y-%m-%d"))
            #print day
            weekday = when.strftime('%A')
            #print weekday
            month = when.strftime('%B')
            #print month
            hour = when.strftime('%H')
            min = when.strftime('%M')
            writer.writerow({
                'date':when,
                'weekday':weekday,
                'month':month,
                'hour':hour,
                'min':min,
                'search_text':search.encode('utf-8')
            })
finally:
    outfile.close()