rec10/trunk/src/auto_rec.py

   1 #!/usr/bin/python
   2 # coding: UTF-8
   3 # Rec10 TS Recording Tools
   4 # Copyright (C) 2009 Yukikaze
   5 import n_gram
   6 import rec10d
   7 def get_db_key(key,chtxt):
   8     return rec10d.rec10db.select_by_key_bayeskey(key, chtxt)
   9 def change_db_epg(chtxt,beforenum,newnum):
  10     rec10d.rec10db.change_ratio_all_reduce_bayeskey(chtxt, beforenum, newnum)
  11 def change_db_num(chtxt,recaddnum,alladdnum):
  12     """
  13     add num.(Not change)
  14     """
  15     rec10d.rec10db.add_num_bayeskey(chtxt, recaddnum,alladdnum)
  16 def change_db_ratio_rec(key,chtxt,beforenum,addnum):
  17     rec10d.rec10db.change_ratio_rec_bayeskey(key, chtxt, beforenum, addnum)
  18 def change_db_ratio_rec_many(chtxt,beforenum,list):
  19     rec10d.rec10db.change_multi_ratio_rec_bayeskey(chtxt, beforenum, list)
  20 def change_db_reduce_ratio_rec(chtxt,beforenum,addnum):
  21     rec10d.rec10db.change_ratio_rec_reduce_bayeskey(chtxt, beforenum, addnum)
  22 def change_db_ratio_all(key,chtxt,beforenum,addnum):
  23     rec10d.rec10db.change_ratio_all_bayeskey(key, chtxt, beforenum, addnum)
  24 def change_db_ratio_all_many(chtxt,beforenum,list):
  25     rec10d.rec10db.change_multi_ratio_all_bayeskey(chtxt,beforenum, list)
  26 def get_db_num(chtxt):
  27     ret=get_db_key("NUM", chtxt)
  28     if len(ret)<3:
  29         rec10d.rec10db.add_bayeskey("NUM", chtxt, 1,1)
  30         ret=get_db_key("NUM", chtxt)
  31     return ret
  32 def update_recall(chtxt,titles,descs,newnum):
  33     """
  34     update recall
  35     """
  36     titles=" "+titles+" "
  37     descs=" "+descs+" "
  38     str1=titles*2+descs+titles*2
  39     noun=n_gram.get_noun_quad_gram(str1)
  40     bnum=get_db_num(chtxt)[3]
  41     if bnum<1:
  42         bnum=100
  43     change_db_ratio_all_many(chtxt, bnum,noun)
  44     change_db_epg(chtxt, bnum, newnum)
  45     change_db_num(chtxt, 0, newnum)
  46 def add_key(chtxt,title,desc):
  47     title=u" "+title+u" "
  48     desc=u" "+desc+u" "
  49     str1=title*2+desc+title*2##タイトルは重視したいので幾度か足す。
  50     print str1
  51     noun=n_gram.get_noun_quad_gram(str1)#nounは辞書のキーに文字を、中身に出現回数を書いたもの。
  52     bnum=get_db_num(chtxt)[2]
  53     print noun
  54     if bnum<1:
  55         bnum=1
  56     change_db_ratio_rec_many(chtxt, bnum, noun)
  57     change_db_num(chtxt,1,0)
  58     change_db_reduce_ratio_rec(chtxt, bnum,1)
  59 def calc_key(chtxt,title,desc):
  60     """
  61
  62     """
  63     kyoukai=250
  64     title=" "+title+" "
  65     desc=" "+desc+" "
  66     str1=title*2+desc+title*2##タイトルは重視したいので幾度か足す。
  67     noun=n_gram.get_noun_quad_gram(str1)#nounは辞書のキーに文字を、中身に出現回数を書いたもの。
  68     tnum=1
  69     nump=0
  70     for key,num in noun.items():
  71         pp=get_db_key(key,chtxt)
  72         pp2=get_db_key(key,"ALL")
  73         if len(pp)>3:
  74             trec=pp[2]
  75             tall=pp[3]
  76             tarec=pp2[2]
  77             taall=pp2[3]
  78             if tall<=trec or taall<=tarec:
  79                 tnum=tnum+1
  80                 p=(1000*trec+5)*100/(1000*tall+5)
  81                 p2=(1000*tarec+5)*100/(1000*taall+5)
  82                 nump=nump+p+p2
  83     if (nump/tnum)>kyoukai:
  84         return 1
  85     else:
  86         return 0