def rankNew():
fils = glob.glob('sentenceSplit/*') # why sentenceSplit?
lengs = ["File\tnltk sentence\tlength by '.'\t% capitalized after '.'\tlength by ','\t% capitalized after ','\t"
+". or , and capital\tlength by capitals\tlength by tab\ttotal length"""]
model = getModel()
for fil in fils:
print 'file',fil
#txt = removePageN(extractText(fil))
txt = (extractText(fil))
# first count the sentences with nltk
nltk = model.tokenize(txt)
# then rank '.'
len2,caps = countSentence(txt,'.')
# then count the sentences separated by ','
(klen,caps2) = countSentence(txt,',')
# then count the sentences separated by cpC
ss = chunker.grouptext(txt,'cpC')
(cpC,_) = countWords(ss)
# then count the sentences separated by capitalized words
ss = chunker.grouptext(txt,'cap')
(cap,_) = countWords(ss)
# then count the sentences separated by tabs
(tlen,_) = countSentence(txt,'\t')
(tot,_) = countWords([txt])
print tot
totlen = tot[0]
# calculate medium for the file
normalleng = div(totlen,nltk)
dotleng = div(totlen,len2)
caps = 0 if float(len(caps))==0 else len(filter(lambda x: x, caps))/float(len(caps))
cleng = div(totlen,klen)
caps2 = 0 if float(len(caps2))==0 else len(filter(lambda x: x, caps2))/float(len(caps2))
cpCleng = div(totlen,cpC)
capleng = div(totlen,cap)
tleng = div(totlen,tlen)
# then count the sentences separated by paras
inp = open(fil,'r').read()
xml = etree.fromstring(inp)
body = xml.find(usefuls.prefix+'body')
paras = body.findall('paragraph')
row = [fil,normalleng,dotleng,caps,cleng,caps2,cpCleng,capleng,tleng,totlen]
lengs += ['\t'.join(map(lambda x: str(x),row))]
table = '\n'.join(lengs)
open('data3.txt','w').write(table)
def rankSuper():
fils = glob.glob('sentenceSplit/*')
lengs = ["File\tnltk sentence\tlength by '.'\t% capitalized after '.'\tlength by ','\t% capitalized after ','\t"
+". or , and capital\tlength by capitals\tlength by tab\tparagraph length\ttotal length"""]
model = getModel()
for fil in fils:
print 'file',fil
inp = open(fil,'r').read()
xml = etree.fromstring(inp)
body = xml.find(usefuls.prefix+'body')
paras = body.findall('paragraph')
ss = []
len1,len2,cleng,caps,caps2,cpC,cap,tlen = [[] for x in range(8)]
for para in paras:
txt = list(para.itertext())
if txt is not None and len(txt)>0:
#txt = removePageN(' '.join(txt))
txt = ' '.join(txt)
# first count the sentences with nltk
s = model.tokenize(txt)
(l,_) = countWords(s)
len1 += (l)
# then rank '.'
l,c = countSentence(txt,'.')
len2 += (l)
caps += (c)
# then count the sentences separated by ','
(klen,capsC) = countSentence(txt,',')
cleng += (klen)
caps2 += (capsC)
# then count the sentences separated by cpC
s = chunker.grouptext(txt,'cpC')
(l,_) = countWords(s)
cpC += l
# then count the sentences separated by capitalized words
s = chunker.grouptext(txt,'cap')
(l,_) = countWords(s)
cap += (l)
# then count the sentences separated by tabs
(l,_) = countSentence(txt,'\t')
tlen += (l)
ss += [' '+txt]
# calculate medium for the file
normalleng = mediumLen(len1)
dotleng = mediumLen(len2)
caps = 0 if float(len(caps))==0 else len(filter(lambda x: x, caps))/float(len(caps))
cleng = mediumLen(cleng)
caps2 = 0 if float(len(caps2))==0 else len(filter(lambda x: x, caps2))/float(len(caps2))
cpCleng = mediumLen(cpC)
capleng = mediumLen(cap)
tleng = mediumLen(tlen)
txt = extractText(fil)
(tot,_) = countWords([txt])
totleng = mediumLen(tot)
# then count the sentences separated by paras
# inp = open(fil,'r').read()
# xml = etree.fromstring(inp)
# body = xml.find(usefuls.prefix+'body')
# paras = body.findall('paragraph')
# for para in paras:
# txt = list(para.itertext())
# if txt is not None and len(txt)>0:
# ss.append(' '.join(txt))
ss = filter(lambda x : len(x.strip())>0,ss)
(lens,_) = countWords(ss)
paraleng = mediumLen(lens)
row = [fil,normalleng,dotleng,caps,cleng,caps2,cpCleng,capleng,tleng,paraleng,totleng]
lengs += ['\t'.join(map(lambda x: str(x),row))]
table = '\n'.join(lengs)
open('data3.txt','w').write(table)
