rec10/trunk/src/guess.py

   1 #!/usr/bin/python
   2 # coding: UTF-8
   3 # Rec10 TS Recording Tools
   4 # Copyright (C) 2009-2010 Yukikaze
   5
   6 import os
   7 import os.path
   8 import re
   9 import time
  10 import os
  11 import datetime
  12 import n_gram
  13 import recdblist
  14 import zenhan
  15 import auto_move
  16
  17 def detName(title,path):
  18     tt=detNameType(title, path)
  19     if tt['type']=="C"or tt['type']=="D":
  20         tt['num']=detNum(tt['title'],tt['folder'])
  21 def detSeriesNum(execpath):
  22     files=os.listdir(execpath)
  23     ss={}
  24     for file in files:
  25         ftitle=os.path.splitext(file)
  26         fname=ftitle[0]
  27         fname=zenhan.toHankaku(unicode(name))
  28         st=detNameType(fname, execpath)['title']
  29         if not len(ss[st])>0:
  30             ss[st]=[]
  31         ss[detNum(fname,execpath)]=ftitle
  32     return ss
  33 def detNum(title,movepath):
  34     """
  35         #番組の話数を推測する。
  36     """
  37     files=os.listdir(movepath)
  38     ff=[]
  39     maxnum=0
  40     for file in files:
  41         recdblist.printutf8(file)
  42         file=os.path.join(movepath,file)
  43         if os.path.isfile(file):
  44             name=os.path.splitext(os.path.split(file)[1])
  45             name=name[0]
  46             name=zenhan.toHankaku(unicode(name))
  47             p1=detNameType(name,movepath)
  48             recdblist.printutf8(p1['title']+" "+str(p1['num']))
  49             time1=time.localtime(os.path.getmtime(file))
  50             time1=datetime.datetime.fromtimestamp(os.path.getmtime(file))
  51             if p1['num']!=-1:
  52                 ff.append([p1['num'],p1['title'],time1])
  53                 if maxnum<p1['num']+1:
  54                     maxnum=p1['num']+1
  55     return detMultiDeltaDays(ff)
  56 def detMultiDeltaDays(num_with_title_with_dates):
  57     maxnum=0
  58     for ft in num_with_title_with_dates:
  59         if maxnum<ft[0]+1:
  60             maxnum=ft[0]+1
  61     ff=num_with_title_with_dates
  62     f3=[None]*maxnum
  63     fret={}
  64     for i in range(0, maxnum, 1):
  65         f3[i]=[]
  66     for f2 in ff:
  67         f3[f2[0]].append(f2)
  68     for i in range(maxnum):
  69         for j in range(i+1,maxnum,1):
  70             for ft3 in f3[i]:
  71                 for ft4 in f3[j]:
  72                     for gdd in detSingleDeltaDay(i, ft3[2], j, ft4[2]):
  73                         #print gdd
  74                         if fret.get(gdd[0])!=None:
  75                             #print fret
  76                             fret[gdd[0]]=fret[gdd[0]]+gdd[1]
  77                         else:
  78                             fret[gdd[0]]=gdd[1]
  79     #print fret
  80     maxk=0
  81     maxp=0
  82     for i in range(maxnum):
  83         if fret.get(i)!=None:
  84             if maxp<fret[i]:
  85                 maxk=i
  86                 maxp=fret[i]
  87     #print maxk
  88     #print maxp
  89     return maxk
  90 def detSingleDeltaDay(num1,date1,num2,date2):
  91     dd=date1-date2
  92     if date1<date2:
  93         dd=date2-date1
  94     dn=dd.days
  95     d=num2-num1
  96     if d<0:
  97         d=-1*d
  98     dp=dd+datetime.timedelta(hours=6)
  99     dp=dp.days
 100     dm=dd-datetime.timedelta(hours=6)
 101     dm=dm.days
 102     ret=[]
 103     if dn%d*2>d:
 104         dn=dn+d
 105     if dm%d*2>d:
 106         dm=dm+d
 107     if dp%d*2>d:
 108         dp=dp+d
 109     recdblist.printutf8(str(d)+":"+str(dn))
 110     if dp!=dn:
 111         ret.append([dn/d,60])
 112         ret.append([dp/d,40])
 113     elif dm!=dn:
 114         ret.append([dn/d,60])
 115         ret.append([dm/d,40])
 116     else:
 117         ret.append([dn/d,100])
 118     return ret
 119 def detNameType(title,path):
 120     """
 121     type A ---title#<number>
 122     type B ---title#<number>subtitle
 123     type C ---title subtitle
 124     type D ---title(without number)
 125     type Aj ---title第<number>話
 126     path --search reflexively
 127     """
 128     recdblist.printutf8(title)
 129     title=auto_move.getTitle(title)##titleから日時を除く
 130     #rA=re.compile(".+(?P<title>)#\d(?P<num>)\s[0,10]\z")
 131     rA=re.compile("(.+)#(\d*)\s*\Z")
 132     tA=rA.match(title)
 133     rB=re.compile("(.+)#(\d*)\s*(\D*)")
 134     tB=rB.match(title)
 135     rAj=re.compile("(.+)第(\d*)話\s*\Z")
 136     tAj=rAj.match(title)
 137     ret={'title':"",'type':"",'num':0,'subtitle':"",'folder':""}
 138     if tA:
 139         recdblist.printutf8("typeA")
 140         recdblist.printutf8("title="+tA.group(1))
 141         recdblist.printutf8("num="+tA.group(2))
 142         ret['type']="A"
 143         ret['title']=tA.group(1)
 144         ret['num']=int(tA.group(2))
 145         ret['folder']=searchFolder(tA.group(1),path)
 146     if tAj:
 147         recdblist.printutf8("typeA")
 148         recdblist.printutf8("title="+tAj.group(1))
 149         recdblist.printutf8("num="+tAj.group(2))
 150         ret['type']="Aj"
 151         ret['title']=tAj.group(1)
 152         ret['num']=int(tAj.group(2))
 153         ret['folder']=searchFolder(tAj.group(1),path)
 154     elif tB:
 155         recdblist.printutf8("typeB")
 156         recdblist.printutf8("title="+tB.group(1))
 157         recdblist.printutf8("num="+tB.group(2))
 158         recdblist.printutf8("subtitle="+tB.group(3))
 159         ret['type']="B"
 160         ret['title']=tB.group(1)
 161         ret['num']=int(tB.group(2))
 162         ret['folder']=searchFolder(tB.group(1),path)
 163         ret['subtitle']=tB.group(3)
 164     else:#type C or type D
 165         #fold=searchFolder(title, path)
 166         ts=title.split(" ")
 167         tt=""
 168         rt=["",0,""]
 169         for t in ts:
 170             tt=tt+" "+t
 171             ft1=searchFolder(tt,path)
 172             recdblist.printutf8(tt)
 173             #print ft1
 174             if ft1!="":
 175                 #recdblist.printutf8(rt)
 176                 #recdblist.printutf8(ft1[0]+" : "+str(ft1[1]))
 177                 if ft1[1]>rt[1]:
 178                     rt[0]=tt
 179                     rt[1]=ft1[1]
 180                     rt[2]=ft1[0]
 181                     #recdblist.printutf8(rt)
 182         #recdblist.printutf8("title="+rt[0][1:]+"/")
 183         #recdblist.printutf8("subtitle = "+title.replace(rt[0][1:],"")[1:])
 184         ret['title']=rt[0][1:]
 185         ret['num']=-1
 186         ret['folder']=rt[2]
 187         ret['subtitle']=title.replace(rt[0][1:],"")[1:]
 188         if ret['subtitle'].replace(" ","")=="":
 189             ret['type']="D"
 190         else:
 191             ret['type']="C"
 192     return ret
 193 def searchFolder(title,path,threshold=500):
 194     """
 195     titleからフォルダーを探す
 196     """
 197     folderpath=os.listdir(path)
 198     lfpath=[]
 199     ngram=[]
 200     for ft in folderpath:
 201         fullpath=os.path.join(path, ft)
 202         if os.path.isdir(fullpath):
 203             lfpath.append(fullpath)
 204             ftt=os.listdir(fullpath)
 205             if len(ftt)>0:
 206                 for ft2 in ftt:
 207                     folderpath.append(os.path.join(fullpath, ft2))
 208         else:
 209             lfpath.append(fullpath)
 210     for dirp in lfpath:
 211         cmpp=""
 212         appp=""
 213         if os.path.isdir(dirp):
 214             cmpp=os.path.dirname(dirp)
 215             appp=dirp
 216         else:
 217             cmpp=os.path.basename(dirp)
 218             appp=os.path.dirname(dirp)
 219         ntitle=auto_move.getTitle(title)
 220         ncmpp=auto_move.getTitle(cmpp)
 221         p=n_gram.trigram(ntitle,ncmpp)
 222         if p>0:
 223             ngram.append((p,appp))
 224     ngram=list(set(ngram))
 225     ngram.sort()
 226     ngram.reverse()
 227     if len(ngram)>0:
 228         #recdblist.printutf8(title + ngram[0][1] + " : "+str(ngram[0][0]))
 229         if ngram[0][0]>threshold:
 230             return ngram[0][1]
 231         else:
 232             return ""
 233     else:
 234         return ""