wresult2 = []
for result,sentence in zip(wresult1,testlocs):
wresult2.append([])
print sentence
for entity in result:
print entity
wresult2[-1].append((sentence[entity[0]], sentence[entity[1]], entity[2]))
wresult = wresult2
# char model
data = preprocess.getcData(datafile)
cdic, cvectors = preprocess.chars2dic2("char_vector_50",chardim)
padding_id = cdic["<padding>"] = len(cvectors)
pids = [padding_id]
cvectors.append(np.random.randn(chardim))
embedding = cvectors
indexdata = preprocess.raw2num1(data,cdic,tags,0,padding_id)
traindata = indexdata[0:len(indexdata)/20*16]
devdata = indexdata[len(indexdata)/20*16:len(indexdata)/20*18]
testdata = indexdata[len(indexdata)/20*18:len(indexdata)]
testwdata = data[len(indexdata)/20*18:len(indexdata)]
cgold = []
for item in testdata:
af = open(nerstr+"/data", 'wb')
pickle.dump(indexdata,af)
af.close()
df = open(nerstr+"/worddic", 'wb')
pickle.dump(wdic,df)
df.close()
tf = open(nerstr+"/tags",'wb')
pickle.dump(tags,tf)
tf.close()
elif mtype == "char":
data = preprocess.getcData(datafile)
tags = preprocess.tags2dic(map(lambda x:x[1], data))
cdic, cvectors = preprocess.chars2dic2("char_vector_50",chardim)
padding_id = cdic["<padding>"] = len(cvectors)
pids = [padding_id]
cvectors.append(np.random.randn(chardim))
embedding = cvectors
indexdata = preprocess.raw2num1(data,cdic,tags,0,padding_id)
indim = chardim
nerstr = sys.argv[1]
if os.path.exists(nerstr):
shutil.rmtree(nerstr)
os.mkdir(nerstr)
af = open(nerstr+"/data", 'wb')
pickle.dump(indexdata,af)
af.close()
df = open(nerstr+"/chardic", 'wb')