Esempi in Python per PTT

Esempio n. 1

File: PTT_LOAN.py Progetto: ntuaha/NewsInsight2

import re
import sys
import os
import datetime
import json
#處理掉unicode 和 str 在ascii上的問題
reload(sys)
sys.setdefaultencoding('utf8')

#aha's library
from PTT import PTT_DB,PTT



if __name__ =="__main__":
  ptt = PTT('https://www.ptt.cc/bbs/Loan/index.html')
  db = PTT_DB(os.path.dirname(__file__)+"/mongodb.inf","loan")
  now = datetime.datetime.now()
  t = datetime.datetime(now.year,now.month,now.day) - datetime.timedelta(days=1)
  print t
  if len(sys.argv)>1:
    limit = int(sys.argv[1])
  else:
    limit = 0
  posts = ptt.fetchData(t,limit)

  #Append News
  print db.bulkInsertNews(posts)

Esempio n. 2

File: PTT2.py Progetto: yenchih/NewsInsight

import re
import sys
import os
import datetime
import json
#處理掉unicode 和 str 在ascii上的問題
reload(sys)
sys.setdefaultencoding('utf8')

#aha's library
from PTT import PTT_DB,PTT



if __name__ =="__main__":
  ptt = PTT('https://www.ptt.cc/bbs/Bank_Service/index.html')
  db = PTT_DB(os.path.dirname(__file__)+"/mongodb.inf","bank_service")
  now = datetime.datetime.now()
  #t = datetime.datetime(now.year,now.month,now.day) - datetime.timedelta(days=1)
  t = datetime.datetime(2014,6,1)
  print t
  if len(sys.argv)>1:
    limit = int(sys.argv[1])
  else:
    limit = 0
  posts = ptt.fetchData(t,limit)

  #Append News
  print db.bulkInsertNews(posts)

Esempio n. 3

#FUND
#boy-girl
#gay
#finance
#creditcard
#Foreign_Inv/lP
#IC-Card
#Lifeismoney
#tax
#CFP
#e-coupon
#food

if __name__ == "__main__":
    board = sys.argv[1]
    ptt = PTT('https://www.ptt.cc/bbs/%s/index.html' % board)
    #db = PTT_DB(os.path.dirname(__file__)+"/mongodb.inf",board)
    #now = datetime.datetime.now()
    #t = datetime.datetime(now.year,now.month,now.day) - datetime.timedelta(days=1)
    t = datetime.datetime(2013, 1, 1)
    #print t
    #if len(sys.argv)>1:
    #  limit = int(sys.argv[1])
    #else:
    limit = 0
    #posts = ptt.fetchData(t,limit)
    reRun = False
    url = 'https://www.ptt.cc/bbs/%s/index.html' % board
    while 1:
        reRun, url = ptt.crawlData(t, limit, board, url)
        if reRun == False:

Esempio n. 4

File: PTT_CRAWLER.py Progetto: ntuaha/NewsInsight2

#FUND
#boy-girl
#gay
#finance
#creditcard
#Foreign_Inv/lP
#IC-Card
#Lifeismoney
#tax
#CFP
#e-coupon
#food

if __name__ =="__main__":
  board = sys.argv[1]
  ptt = PTT('https://www.ptt.cc/bbs/%s/index.html'%board)
  #db = PTT_DB(os.path.dirname(__file__)+"/mongodb.inf",board)
  #now = datetime.datetime.now()
  #t = datetime.datetime(now.year,now.month,now.day) - datetime.timedelta(days=1)
  t = datetime.datetime(2013,1,1)
  #print t
  #if len(sys.argv)>1:
  #  limit = int(sys.argv[1])
  #else:
  limit = 0
  #posts = ptt.fetchData(t,limit)
  reRun = False
  url = 'https://www.ptt.cc/bbs/%s/index.html'%board
  while 1:
    reRun,url = ptt.crawlData(t,limit,board,url)
    if reRun == False:

Esempio n. 5

File: PTT_ForeignEX.py Progetto: yenchih/NewsInsight

# -*- coding: utf-8 -*-

import re
import sys
import os
import datetime
import json
#處理掉unicode 和 str 在ascii上的問題
reload(sys)
sys.setdefaultencoding('utf8')

#aha's library
from PTT import PTT_DB, PTT

if __name__ == "__main__":
    ptt = PTT('https://www.ptt.cc/bbs/ForeignEX/index.html')
    db = PTT_DB(os.path.dirname(__file__) + "/mongodb.inf", "foreign_ex")
    now = datetime.datetime.now()
    t = datetime.datetime(now.year, now.month,
                          now.day) - datetime.timedelta(days=1)
    print t
    if len(sys.argv) > 1:
        limit = int(sys.argv[1])
    else:
        limit = 0
    posts = ptt.fetchData(t, limit)

    #Append News
    print db.bulkInsertNews(posts)

Esempio n. 6

File: PTT_LOAN.py Progetto: yenchih/NewsInsight

# -*- coding: utf-8 -*-

import re
import sys
import os
import datetime
import json
#處理掉unicode 和 str 在ascii上的問題
reload(sys)
sys.setdefaultencoding('utf8')

#aha's library
from PTT import PTT_DB, PTT

if __name__ == "__main__":
    ptt = PTT('https://www.ptt.cc/bbs/Loan/index.html')
    db = PTT_DB(os.path.dirname(__file__) + "/mongodb.inf", "loan")
    now = datetime.datetime.now()
    t = datetime.datetime(now.year, now.month,
                          now.day) - datetime.timedelta(days=1)
    print t
    if len(sys.argv) > 1:
        limit = int(sys.argv[1])
    else:
        limit = 0
    posts = ptt.fetchData(t, limit)

    #Append News
    print db.bulkInsertNews(posts)