rec10/trunk/src/guess.py

   1 #!/usr/bin/python
   2 # coding: UTF-8
   3 # Rec10 TS Recording Tools
   4 # Copyright (C) 2009-2011 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 detSeriesNum(execpath):
  18     files=os.listdir(execpath)
  19     ss={}
  20     for file in files:
  21         ftitle=os.path.splitext(file)
  22         fname=ftitle[0]
  23         fname=zenhan.toHankaku(unicode(fname,"utf-8"))
  24         nt=detNameType(fname, execpath)
  25         st=nt['title']
  26         if not ss.has_key(st):
  27             ss[st]={}
  28         ss[st][nt['num']]=ftitle
  29         print ss
  30     return ss
  31 def detDeltaNum(title,movepath):
  32     """
  33         #番組の話数を推測する。
  34     """
  35     files=os.listdir(movepath)
  36     ff=[]
  37     maxnum=0
  38     for file in files:
  39         recdblist.printutf8(unicode(file,'utf-8'),verbose_level=800)
  40         file=os.path.join(movepath,file)
  41         if os.path.isfile(file):
  42             name=os.path.splitext(os.path.split(file)[1])
  43             name=name[0]
  44             name=zenhan.toHankaku(unicode(name,'utf-8'))
  45             p1=detNameType(name,movepath)
  46             #recdblist.printutf8(p1['title']+" "+str(p1['num']))
  47             time1=time.localtime(os.path.getmtime(file))
  48             time1=datetime.datetime.fromtimestamp(os.path.getmtime(file))
  49             if p1['num']!=-1:
  50                 ff.append([p1['num'],p1['title'],time1])
  51                 if maxnum<p1['num']+1:
  52                     maxnum=p1['num']+1
  53     return detMultiDeltaDays(ff)
  54 def detMultiDeltaDays(num_with_title_with_dates):
  55     maxnum=0
  56     for ft in num_with_title_with_dates:
  57         if maxnum<ft[0]+1:
  58             maxnum=ft[0]+1
  59     ff=num_with_title_with_dates
  60     f3=[None]*maxnum
  61     fret={}
  62     for i in range(0, maxnum, 1):
  63         f3[i]=[]
  64     for f2 in ff:
  65         f3[f2[0]].append(f2)
  66     for i in range(maxnum):
  67         for j in range(i+1,maxnum,1):
  68             for ft3 in f3[i]:
  69                 for ft4 in f3[j]:
  70                     for gdd in detSingleDeltaDay(i, ft3[2], j, ft4[2]):
  71                         #print gdd
  72                         if fret.get(gdd[0])!=None:
  73                             #print fret
  74                             fret[gdd[0]]=fret[gdd[0]]+gdd[1]
  75                         else:
  76                             fret[gdd[0]]=gdd[1]
  77     #print fret
  78     maxk=0
  79     maxp=0
  80     for i in range(maxnum):
  81         if fret.get(i)!=None:
  82             if maxp<fret[i]:
  83                 maxk=i
  84                 maxp=fret[i]
  85     #print maxk
  86     #print maxp
  87     return maxk
  88 def detSingleDeltaDay(num1,date1,num2,date2):
  89     dd=date1-date2
  90     if date1<date2:
  91         dd=date2-date1
  92     dn=dd.days
  93     d=num2-num1
  94     if d<0:
  95         d=-1*d
  96     dp=dd+datetime.timedelta(hours=6)
  97     dp=dp.days
  98     dm=dd-datetime.timedelta(hours=6)
  99     dm=dm.days
 100     ret=[]
 101     if dn%d*2>d:
 102         dn=dn+d
 103     if dm%d*2>d:
 104         dm=dm+d
 105     if dp%d*2>d:
 106         dp=dp+d
 107     #recdblist.printutf8(str(d)+":"+str(dn))
 108     if dp!=dn:
 109         ret.append([dn/d,60])
 110         ret.append([dp/d,40])
 111     elif dm!=dn:
 112         ret.append([dn/d,60])
 113         ret.append([dm/d,40])
 114     else:
 115         ret.append([dn/d,100])
 116     return ret
 117 def detNameType(title,path):
 118     """
 119     type A ---title#<number>
 120     type B ---title#<number>subtitle
 121     type C ---title subtitle
 122     type D ---title(without number)
 123     type Aj ---title第<number>話
 124     path --search reflexively
 125     """
 126     new=0
 127     if re.search(u"[新]",title) or re.search(u" 新",title):
 128         title=title.replace(u"[新]","")
 129         title=title.replace(u" 新","")
 130         new=1
 131     recdblist.printutf8(title,verbose_level=800)
 132     title=auto_move.getTitle(title)##titleから日時を除く
 133     title=title.replace(u"無料≫","")
 134     #rA=re.compile(".+(?P<title>)#\d(?P<num>)\s[0,10]\z")
 135     rA=re.compile("(.+)#(\d*)\s*\Z")
 136     tA=rA.match(title)
 137     rB=re.compile("(.+)#(\d*)\s*(\D*)")
 138     tB=rB.match(title)
 139     rAj=re.compile("(.+)第(\d*)話\s*\Z")
 140     tAj=rAj.match(title)
 141     ret={'title':"",'type':"",'num':0,'subtitle':"",'folder':""}
 142     if tA:
 143         #recdblist.printutf8("typeA")
 144         #recdblist.printutf8("title="+tA.group(1))
 145         #recdblist.printutf8("num="+tA.group(2))
 146         ret['type']="A"
 147         ret['title']=tA.group(1).replace(" ","")
 148         ret['num']=int(tA.group(2))
 149         ret['folder']=searchFolder(tA.group(1),unicode(path,'utf-8'))
 150     if tAj:
 151         #recdblist.printutf8("typeA")
 152         #recdblist.printutf8("title="+tAj.group(1))
 153         #recdblist.printutf8("num="+tAj.group(2))
 154         ret['type']="Aj"
 155         ret['title']=tAj.group(1).replace(" ","")
 156         ret['num']=int(tAj.group(2))
 157         ret['folder']=searchFolder(tAj.group(1),unicode(path,'utf-8'))
 158     elif tB:
 159         #recdblist.printutf8("typeB")
 160         #recdblist.printutf8("title="+tB.group(1))
 161         #recdblist.printutf8("num="+tB.group(2))
 162         #recdblist.printutf8("subtitle="+tB.group(3))
 163         ret['type']="B"
 164         ret['title']=tB.group(1).replace(" ","")
 165         ret['num']=int(tB.group(2))
 166         ret['folder']=searchFolder(tB.group(1),unicode(path,'utf-8'))
 167         ret['subtitle']=tB.group(3)
 168     else:#type C or type D
 169         #fold=searchFolder(title, path)
 170         ts=title.split(" ")
 171         tt=""
 172         rt=["",0,""]
 173         for t in ts:
 174             tt=tt+" "+t
 175             ft1=searchFolder(tt,unicode(path,'utf-8'))
 176             #recdblist.printutf8(tt)
 177             #print ft1
 178             if ft1!="":
 179                 #recdblist.printutf8(rt)
 180                 #recdblist.printutf8(ft1[0]+" : "+str(ft1[1]))
 181                 if ft1[1]>rt[1]:
 182                     rt[0]=tt
 183                     rt[1]=ft1[1]
 184                     rt[2]=ft1[0]
 185                     #recdblist.printutf8(rt)
 186         #recdblist.printutf8("title="+rt[0][1:]+"/")
 187         #recdblist.printutf8("subtitle = "+title.replace(rt[0][1:],"")[1:])
 188         ret['title']=rt[0][1:].replace(" ","")
 189         ret['num']=-1
 190         ret['folder']=rt[2]
 191         ret['subtitle']=title.replace(rt[0][1:],"")[1:]
 192         if ret['subtitle'].replace(" ","")=="":
 193             ret['type']="D"
 194         else:
 195             ret['type']="C"
 196     if new==1:
 197         ret['num']=1
 198     return ret
 199 def searchFolder(title,path,threshold=500):
 200     """
 201     titleからフォルダーを探す
 202     """
 203     folderpath=os.listdir(path)
 204     lfpath=[]
 205     ngram=[]
 206     for ft in folderpath:
 207         fullpath=os.path.join(path, ft)
 208         if os.path.isdir(fullpath):
 209             lfpath.append(fullpath)
 210             ftt=os.listdir(fullpath)
 211             if len(ftt)>0:
 212                 for ft2 in ftt:
 213                     folderpath.append(os.path.join(fullpath, ft2))
 214         else:
 215             lfpath.append(fullpath)
 216     for dirp in lfpath:
 217         cmpp=""
 218         appp=""
 219         if os.path.isdir(dirp):
 220             cmpp=os.path.dirname(dirp)
 221             appp=dirp
 222         else:
 223             cmpp=os.path.basename(dirp)
 224             appp=os.path.dirname(dirp)
 225         ntitle=auto_move.getTitle(title)
 226         ncmpp=auto_move.getTitle(cmpp)
 227         p=n_gram.trigram(ntitle,ncmpp)
 228         if p>0:
 229             ngram.append((p,appp))
 230     ngram=list(set(ngram))
 231     ngram.sort()
 232     ngram.reverse()
 233     if len(ngram)>0:
 234         #recdblist.printutf8(title + ngram[0][1] + " : "+str(ngram[0][0]))
 235         if ngram[0][0]>threshold:
 236             return ngram[0][1]
 237         else:
 238             return ""
 239     else:
 240         return ""