导出微信聊天记录为 txt

wechat2txt.py

#!/usr/bin/python2.7
# -*- coding: UTF-8 -*-
# vim:set shiftwidth=4 tabstop=4 expandtab textwidth=79:
'''导出微信聊天记录为 txt

@author: Shellexy Wang <shellexyone@gmail.com>
@license: LGPLv3+
@see: 
'''

import os
import sys
import re
import hashlib
import csv
import time
import locale
import getopt

WECHAT_DB_KEY = ''

def get_db_path():
    '''get wechat EnMicroMsg.db path

    like /data/data/com.tencent.mm/MicroMsg/********************************/EnMicroMsg.db
    '''
    os.popen('adb root').close()
    txt = os.popen("adb shell su -c 'ls /data/data/com.tencent.mm/MicroMsg/*/EnMicroMsg.db' ").read()
    return txt.splitlines()[-1] if txt else ''

def get_uin():
    '''get wechat uin
    '''
    os.popen('adb root').close()
    txt = os.popen("adb shell su -c 'cat /data/data/com.tencent.mm/shared_prefs/system_config_prefs.xml' ").read()
    uin = re.findall('name="default_uin" value="([0-9]+)"', txt)
    return uin[0] if uin else 0

def get_imei():
    '''get IMEI
    '''
    txt = os.popen('adb shell dumpsys iphonesubinfo').read()
    imei = re.findall('Device ID = ([0-9]+)', txt)
    return imei[0] if imei else 0

def get_key():
    '''get key
    '''
    global WECHAT_DB_KEY
    if WECHAT_DB_KEY:
        return WECHAT_DB_KEY
    uin = get_uin()
    imei = get_imei()
    if uin and imei:
        return hashlib.md5(imei + uin).hexdigest()[0:7]
    return ''


def messagecsv2chat(msgcsv):
    '''message csv to chatlog

    header: msgId,msgSvrId,type,status,isSend,isShowTimer,createTime,talker,content,imgPath,reserved,lvbuffer
    '''
    locale.setlocale(locale.LC_ALL, '')
    if hasattr(msgcsv, 'title'):
        msgcsv = [line + '\n' for line in msgcsv.splitlines()]
        pass
    lines = csv.reader(msgcsv)
    lines.next()
    # 聊天记录用 createTime 来重新排序，因为数据库里的序号可能有乱的，目前 csv 格式 创建时间 是在 [6]
    lines = sorted(list(lines), key = lambda line: line[6])
    for line in lines:
        try:
            msgId,msgSvrId,type_,status,isSend,isShowTimer,createTime,talker,content,imgPath,reserved,lvbuffer = line[:12]
            pass
        except:
            continue
        sender = 'me' if (isSend == '1') else talker
        sendtime = time.localtime(int(createTime)/1000)
        sendat = time.strftime("%Y-%m-%d 周%a %H:%M:%S", sendtime)
        yield [talker, sendat, sender, content, imgPath]
        pass
    pass

def chat2talkers(chat):
    '''get talker list from chat
    '''
    names = {}
    for talker, sendat, sender, content, imgPath in chat:
        names[talker] = 1
        pass
    return names.keys()

def chat2txt(chat, name = ''):
    '''chatlog to txt
    '''
    txt = []
    name = name.lower()
    for talker, sendat, sender, content, imgPath in chat:
        # 去掉群聊记录里 content 第一行 name: 后额外的换行
        if '@chatroom' in name: content = content.replace(':\n', ': ', 1)
        # 其他换行在行首加空格
        content = content.replace('\n', '\n  ')
        imgPath = ('\t' + imgPath) if imgPath else ''
        if not name:
            txt.append('%s: %s %s: %s %s' % (talker, sendat, sender, content, imgPath))
            pass
        elif talker.lower() == name:
            txt.append('%s %s: %s %s' % (sendat, sender, content, imgPath))
            pass
        pass
    return '\n'.join(txt) + '\n'

def chat2txts(chat, names = []):
    '''chat log to txts for names
    '''
    txts = {}
    names = [name.lower() for name in names]
    for talker, sendat, sender, content, imgPath in chat:
        name = talker.lower()
        if name in names:
            # 去掉群聊记录里 content 第一行 name: 后额外的换行
            if '@chatroom' in name: content = content.replace(':\n', ': ', 1)
            # 其他换行在行首加空格
            content = content.replace('\n', '\n  ')
            imgPath = ('\t' + imgPath) if imgPath else ''
            txt = txts.get(name, [])
            if not txt: txts[name] = txt
            txt.append('%s %s: %s %s' % (sendat, sender, content, imgPath))
            pass
        pass
    txts = { name: '\n'.join(txt) + '\n' for name, txt in txts.items() }
    return txts

def get_sqlc_tables(dbn, key = ''):
    '''list tables of db
    '''
    i, o = os.popen2(['sqlcipher', dbn])
    if key:
        i.write('PRAGMA key=%s;\n' % `key`)
        i.write('PRAGMA cipher_migrate;\n')
        pass
    i.write('.tables\n')
    i.close()
    ## 由于加入 PRAGMA cipher_migrate; 后输出开头会多了 "cipher_migrate\r\n0\r\n"，所以需要去掉头两行
    return o.read().split()[2:]

def sqlc2csv(dbn, key = '', table = 'message'):
    '''export csv of db
    '''
    table = table or 'message'
    i, o = os.popen2(['sqlcipher', dbn])
    i.write('.header on\n')
    i.write('.mode csv\n')
    if key:
        i.write('PRAGMA key=%s;\n' % `key`)
        i.write('PRAGMA cipher_migrate;\n')
        pass
    i.write('select * from %s;\n' % `table`)
    i.close()
    ## 由于加入 PRAGMA cipher_migrate; 后输出开头会多了 "cipher_migrate\r\n0\r\n"，所以需要去掉头两行
    o.readline()
    o.readline()
    return o.read()

def wechat2csv(tables = []):
    '''export db to csv
    '''
    ldbn = 'EnMicroMsg.db'
    dbn = get_db_path()
    key = get_key()
    os.popen('adb wait-for-device')
    #os.popen('adb pull %s %s' % (`dbn`, `ldbn`)).close()
    '''
    dbbin = os.popen("adb shell su -c 'cat %s' " % `dbn`).read().replace("\r\n", "\n")
    if len(dbbin) < 10:
        return {}
    file(ldbn, 'w').write(dbbin)
    '''
    os.popen('adb shell su -c "cp %s /storage/sdcard1/EnMicroMsg.db" ; adb pull /storage/sdcard1/EnMicroMsg.db %s && adb shell rm /storage/sdcard1/EnMicroMsg.db ' % (`dbn`, `ldbn`)).close()
    if not tables:
        tables = get_sqlc_tables(ldbn, key)
        pass
    for table in tables:
        csvfn = '%s.csv' % table
        csvtxt = sqlc2csv(ldbn, key, table)
        if csvtxt.find('\n') > 0:
            file(csvfn, 'w').write(csvtxt)
            pass
        pass
    return

def wechat2csvtxt(tables = []):
    '''export db to csv
    '''
    ldbn = 'EnMicroMsg.db'
    dbn = get_db_path()
    key = get_key()
    csvtxts = {}
    os.popen('adb wait-for-device')
    '''
    dbbin = os.popen("adb shell su -c 'cat %s' " % `dbn`).read().replace("\r\n", "\n")
    if len(dbbin) < 10:
        return {}
    file(ldbn, 'w').write(dbbin)
    '''
    os.popen('adb shell su -c "cp %s /storage/sdcard1/EnMicroMsg.db" ; adb pull /storage/sdcard1/EnMicroMsg.db %s && adb shell rm /storage/sdcard1/EnMicroMsg.db ' % (`dbn`, `ldbn`)).close()
    if not tables:
        tables = get_sqlc_tables(ldbn, key)
        pass
    for table in tables:
        csvtxt = sqlc2csv(ldbn, key, table)
        if csvtxt.find('\n') > 0:
            csvtxts[table] = csvtxt
            pass
        pass
    return csvtxts


def wechat2txt(names = []):
    '''export chatlog to txt
    '''
    csvtxt = wechat2csvtxt(tables=['message']).get('message')
    file('message.csv', 'w').write(csvtxt)
    chat = list(messagecsv2chat(csvtxt))
    if not chat:
        return 1
    if not names:
        names = chat2talkers(chat)
        pass
    txts = chat2txts(chat, names)
    for name in names:
        fn = 'message.%s.txt' % name
        txt = txts.get(name.lower())
        if len(txt) > 4:
            file(fn, 'w').write(txt)
            pass
        pass
    pass

USAGE = '''Usage: wechat2txt.py [OPTIONS] [NAME]...

OPTIONS:
    -h        display this help and exit
    -t        export csv of database tables
    -k <KEY>  set the wechat db sqlcipher key
'''

def main():
    try:
        opts, args = getopt.getopt(sys.argv[1:], 'htk:')
    except getopt.error, msg:
        print USAGE
        return 1
    global WECHAT_DB_KEY
    tables = []
    for opt, arg in opts:
        if opt == '-h':
            print USAGE
            return 1
        elif opt == '-t':
            tables = args
            pass
        elif opt == '-k':
            key = arg
            WECHAT_DB_KEY = arg
            pass
        pass
    if tables:
        wechat2csv(tables)
        return
    names = args
    txt = wechat2txt(names)
    return not txt

if __name__=="__main__":
    sys.exit(main())