rec10/trunk/src/xml2db_dom.py

   1 #!/usr/bin/python
   2 # coding: UTF-8
   3 # Rec10 TS Recording Tools
   4 # Copyright (C) 2009 Yukikaze
   5
   6 import xml.dom.minidom
   7 import datetime
   8
   9 import zenhan
  10 import chdb
  11 import auto_rec
  12 import recdb
  13 import rec10d
  14 import n_gram
  15 import recdblist
  16 def getText(elm):
  17     nodelist = elm.childNodes
  18     rc = ""
  19     # 全てのノードに対して
  20     for node in nodelist:
  21         # テキストノードなら値を取得
  22         if node.nodeType == node.TEXT_NODE:
  23             rc = rc + node.data
  24     return rc
  25 def getText_item(obj):
  26     rc=""
  27     for o in obj:
  28         rc=rc+getText(o)
  29     return rc
  30 def writemultitvDB(bctype,tvlists):
  31     rec10d.rec10db.new_epg_timeline(bctype)
  32     for channel,start,stop,title,desc,longdesc,category in tvlists:
  33         rec10d.rec10db.add_epg_timeline(bctype, channel, start, stop, title, desc, longdesc, category)
  34 def writemultichDB(chlists):
  35     for ontv,dn in chlists:
  36         rec10d.rec10db.update_chname_by_ontv_epg_ch(ontv, dn)
  37 def xml2db_dom(xmlpath, bctype):#bctypeは放送種別で'TE'(地デジ)BS,CSがある。地デジの場合は te数字 が入る
  38     dtb=datetime.datetime.now()
  39     dom=xml.dom.minidom.parse(file(xmlpath))
  40     chlist=[]
  41     tvlist=[]
  42     bayes={}
  43     for ch in dom.getElementsByTagName('channel'):
  44         ontv=ch.getAttribute("id")
  45         chname=ch.getElementsByTagName("display-name").item(0).childNodes[0].data
  46         chlist.append([ontv,chname])
  47     for tv in dom.getElementsByTagName('programme'):
  48         channel=zenhan.toHankaku_ABC123(tv.getAttribute("channel"))
  49         start=tv.getAttribute("start").replace(" +0900", "")
  50         stop=tv.getAttribute("stop").replace(" +0900", "")
  51         title=zenhan.toHankaku_ABC123(getText_item(tv.getElementsByTagName("title"))).replace(",", "_")
  52         desc=zenhan.toHankaku_ABC123(getText_item(tv.getElementsByTagName("desc")))
  53         longdesc=zenhan.toHankaku_ABC123(getText_item(tv.getElementsByTagName("longdesc")))
  54         category=zenhan.toHankaku_ABC123(getText_item(tv.getElementsByTagName("category")))
  55         tch=chdb.ontvsearch(channel)
  56         if tch!=None:
  57             if tch.get('bctype',"")==bctype:
  58                 bt=bayes.get(tch['chtxt'],["","",0])
  59                 bt2=[bt[0]+title+" ",bt[1]+desc+" "+longdesc+" ",bt[2]+1]
  60                 bayes[tch['chtxt']]=bt2
  61                 tvlist.append([channel,start,stop,title,desc,longdesc,category])
  62                 p=auto_rec.calc_key(tch['chtxt'], title,desc+" "+longdesc)
  63                 if p>2000:
  64                     bttt=datetime.datetime.strptime(start,"%Y%m%d%H%M%S")
  65                     bttt=bttt-datetime.timedelta(seconds=1200)
  66                     bttime=bttt.strftime("%Y-%m-%d %H:%M:%S")
  67                     ettt=datetime.datetime.strptime(stop,"%Y%m%d%H%M%S")
  68                     ettt=ettt+datetime.timedelta(seconds=1200)
  69                     ettime=ettt.strftime("%Y-%m-%d %H:%M:%S")
  70                     chs=rec10d.rec10db.select_bytime_all_timeline(bttime,ettime,tch['chtxt'])
  71                     p2=0
  72                     if len(chs)>0:
  73                         for ch in chs:
  74                             p2t=n_gram.trigram(ch['title'],title)
  75                             if p2t>p2:
  76                                 p2=p2t
  77                     if p2<400:
  78                         recdb.add_auto_bayes(tch['chtxt'],title,start,stop,p)
  79     dom.unlink()
  80     if len(chlist) > 0:
  81         writemultichDB(chlist)
  82         chlist = []
  83         if len(tvlist) > 0:
  84             writemultitvDB(bctype,tvlist)
  85         tvlist = []
  86         for ct,list in bayes.items():
  87             auto_rec.update_recall(ct,list[0],list[1],list[2])
  88             auto_rec.update_recall("ALL",list[0],list[1],list[2])
  89     dtb=datetime.datetime.now()-dtb
  90     recdblist.printutf8(bctype + u" epg取り出し終了")
  91     recdblist.printutf8(str(dtb.days * 24 * 60 * 60 + dtb.seconds)+u"seconds taken for updating bayes-auto DB .")
  92 def getChdata(xmlpath):
  93     dom=xml.dom.minidom.parse(file(xmlpath))
  94     chlist=[]
  95     for ch in dom.getElementsByTagName('channel'):
  96         ontv=ch.getAttribute("id")
  97         chname=ch.getElementsByTagName("display-name").item(0).childNodes[0].data
  98         chlist.append([ontv,chname])
  99     dom.unlink()
 100     if len(chlist)>0:
 101         return chlist[0]
 102     else:
 103         return ["",""]