可以将文章内容翻译成中文,广告屏蔽插件可能会导致该功能失效(如失效，请关闭广告屏蔽插件后再试):
问题:
Working on a script to collect users browser history with time stamps ( educational setting). Firefox 3 history is kept in a sqlite file, and stamps are in UNIX epoch time... getting them and converting to readable format via a SQL command in python is pretty straightforward:
sql_select = """ SELECT datetime(moz_historyvisits.visit_date/1000000,'unixepoch','localtime'), moz_places.url FROM moz_places, moz_historyvisits WHERE moz_places.id = moz_historyvisits.place_id """ get_hist = list(cursor.execute (sql_select))
Chrome also stores history in a sqlite file.. but it's history time stamp is apparently formatted as the number of microseconds since midnight UTC of 1 January 1601....
How can this timestamp be converted to a readable format as in the Firefox example (like 2010-01-23 11:22:09)? I am writing the script with python 2.5.x ( the version on OS X 10.5 ), and importing sqlite3 module....
回答1:
Try this:
sql_select = """ SELECT datetime(last_visit_time/1000000-11644473600,'unixepoch','localtime'), url FROM urls ORDER BY last_visit_time DESC """ get_hist = list(cursor.execute (sql_select))
Or something along those lines
seems to be working for me.
回答2:
This is a more pythonic and memory-friendly way to do what you described (by the way, thanks for the initial code!):
#!/usr/bin/env python import os import datetime import sqlite3 import opster from itertools import izip SQL_TIME = 'SELECT time FROM info' SQL_URL = 'SELECT c0url FROM pages_content' def date_from_webkit(webkit_timestamp): epoch_start = datetime.datetime(1601,1,1) delta = datetime.timedelta(microseconds=int(webkit_timestamp)) return epoch_start + delta @opster.command() def import_history(*paths): for path in paths: assert os.path.exists(path) c = sqlite3.connect(path) times = (row[0] for row in c.execute(SQL_TIME)) urls = (row[0] for row in c.execute(SQL_URL)) for timestamp, url in izip(times, urls): date_time = date_from_webkit(timestamp) print date_time, url c.close() if __name__=='__main__': opster.dispatch()
The script can be used this way:
$ ./chrome-tools.py import-history ~/.config/chromium/Default/History* > history.txt
Of course Opster can be thrown out but seems handy to me :-)
回答3:
The sqlite module returns datetime objects for datetime fields, which have a format method for printing readable strings called strftime.
You can do something like this once you have the recordset:
for record in get_hist: date_string = record[0].strftime("%Y-%m-%d %H:%M:%S") url = record[1]
回答4:
This may not be the most Pythonic code in the world, but here's a solution: Cheated by adjusting for time zone (EST here) by doing this:
utctime = datetime.datetime(1601,1,1) + datetime.timedelta(microseconds = ms, hours =-5)
Here's the function : It assumes that the Chrome history file has been copied from another account into /Users/someuser/Documents/tmp/Chrome/History
def getcr(): connection = sqlite3.connect('/Users/someuser/Documents/tmp/Chrome/History') cursor = connection.cursor() get_time = list(cursor.execute("""SELECT last_visit_time FROM urls""")) get_url = list(cursor.execute("""SELECT url from urls""")) stripped_time = [] crf = open ('/Users/someuser/Documents/tmp/cr/cr_hist.txt','w' ) itr = iter(get_time) itr2 = iter(get_url) while True: try: newdate = str(itr.next()) stripped1 = newdate.strip(' (),L') ms = int(stripped1) utctime = datetime.datetime(1601,1,1) + datetime.timedelta(microseconds = ms, hours =-5) stripped_time.append(str(utctime)) newurl = str(itr2.next()) stripped_url = newurl.strip(' ()') stripped_time.append(str(stripped_url)) crf.write('\n') crf.write(str(utctime)) crf.write('\n') crf.write(str(newurl)) crf.write('\n') crf.write('\n') crf.write('********* Next Entry *********') crf.write('\n') except StopIteration: break crf.close() shutil.copy('/Users/someuser/Documents/tmp/cr/cr_hist.txt' , '/Users/parent/Documents/Chrome_History_Logs') os.rename('/Users/someuser/Documents/Chrome_History_Logs/cr_hist.txt','/Users/someuser/Documents/Chrome_History_Logs/%s.txt' % formatdate)
