def CalcTFIDF(self):
sh = show()
count = sh.showcount()
docArray = self.loadDataFromCutFile(count)
#docArray = self.loadDataFromCutFile(10)
vectorizer = CountVectorizer()
transformer = TfidfTransformer()
tfidf = transformer.fit_transform(vectorizer.fit_transform(docArray))
print 'done'
#write index-doc to file
i = 0
indexdoc = dict()
f = open(Global.inverse_dir+'id.txt','wb')
word = vectorizer.get_feature_names()
for name in vectorizer.get_feature_names():
i+=1
indexdoc[name] = i
f.write(json.dumps(indexdoc))
f.close()
colnum = tfidf.shape[1]
#for i in range(0,colnum):
# filename = Global.inverse_dir+str(i/Global.filesize)+'.txt'
# f = open(filename,'a')
# idx_list = dict()
# for j in range(0,row):
# val = tfidf[j,i]
# if val > 0:
# idx_list[j+1] = val
# f.write(json.dumps(idx_list)+'\n')
# f.close()
#i表示词项的编号，row表示非零文档所在的行
for i in range(0,colnum):
filename = Global.inverse_dir+str(i/Global.filesize)+'.txt'
coldata = tfidf.getcol(i)
col_nonzero_index = np.nonzero(coldata)
item_weight_dict = dict()
for row in col_nonzero_index[0]:
item_weight_dict[row+1] = coldata[row][0].data[0]
f = open(filename,'a')
f.write(json.dumps(item_weight_dict)+'\n')
f.close()
print 'item ',i,'calculate done'
def CalcTFIDF(self):
sh = show()
count = sh.showcount()
docArray = self.loadDataFromCutFile(count)
# docArray = self.loadDataFromCutFile(10)
vectorizer = CountVectorizer()
transformer = TfidfTransformer()
tfidf = transformer.fit_transform(vectorizer.fit_transform(docArray))
print 'done'
# write index-doc to file
i = 0
indexdoc = dict()
f = open(Global.inverse_dir + 'id.txt', 'wb')
word = vectorizer.get_feature_names()
for name in vectorizer.get_feature_names():
i += 1
indexdoc[name] = i
f.write(json.dumps(indexdoc))
f.close()
colnum = tfidf.shape[1]
# for i in range(0,colnum):
# filename = Global.inverse_dir+str(i/Global.filesize)+'.txt'
# f = open(filename,'a')
# idx_list = dict()
# for j in range(0,row):
# val = tfidf[j,i]
# if val > 0:
# idx_list[j+1] = val
# f.write(json.dumps(idx_list)+'\n')
# f.close()
# i表示词项的编号，row表示非零文档所在的行
for i in range(0, colnum):
filename = Global.inverse_dir + str(i / Global.filesize) + '.txt'
coldata = tfidf.getcol(i)
col_nonzero_index = np.nonzero(coldata)
item_weight_dict = dict()
for row in col_nonzero_index[0]:
item_weight_dict[row + 1] = coldata[row][0].data[0]
f = open(filename, 'a')
f.write(json.dumps(item_weight_dict) + '\n')
f.close()
print 'item ', i, 'calculate done'
import sys
sys.path.append("..")
from tools.show import show
import tools.Global as Global
from ml.Cut import Cut
s = show()
#s.showcount()
#s.shownews(1)
s.showKeyWord()
#s.showitem(2608)
#c = Cut()
#line = c.getRow(50,Global.cutnews_origin_dir,Global.filesize)
#s.showitem(line)
