Corpus1=[]
Corpus2=[]
Corpus3=[]
print('\nGenerating three vesions of word-structure annotation...')
count=0
for word in Word2treeID:
index=Word2treeID[word]
tree=NewForest[index]
tree_str1=tree.pprint(margin=10000)
tree_str2=remove_all_subscript(tree).pprint(margin=10000)
tree_str3=remove_crl_subscript(tree).pprint(margin=10000)
# removing '_' between tag and subtag, as Stanford parser will remove the subtag when seeing '_'
new_tree_str1=' '.join([i[:-2]+i[-1] if len(i)>1 and i[-2]=='_' else i for i in tree_str1.split()]) # remove '_' to merge subscript to merge it to the non-terminal
new_tree_str2=tree_str2 # tree_str2 have already remove all the subscripts
new_tree_str3=' '.join([i[:-2]+i[-1] if len(i)>1 and i[-2]=='_' else i for i in tree_str3.split()]) # remove '_' to merge subscript to merge Annotation1.append((new_tree_str1, counter))
Corpus1.append(new_tree_str1)
Corpus2.append(new_tree_str2)
Corpus3.append(new_tree_str3)
print('\ndone!')
if count%int(len(lines)/10)==0:
print(int(count/len(lines)*100),'% finished...')
count +=1
mini_tree_seq1=[]
mini_tree_seq2=[]
mini_tree_seq3=[]
for word in sent:
index=Word2treeID[word]
tree=NewForest[index]
mini_tree_seq1.append(tree.pprint(margin=10000))
mini_tree_seq2.append(remove_all_subscript(tree).pprint(margin=10000))
mini_tree_seq3.append(remove_crl_subscript(tree).pprint(margin=10000))
#mini_tree_seq1.append(keep_all_subscript(tree, Encoder1).pprint(margin=10000))
#mini_tree_seq2.append(remove_all_subscript(tree,Encoder2).pprint(margin=10000))
#mini_tree_seq3.append(remove_crl_subscript(tree, Encoder3).pprint(margin=10000))
Corpus1.append(mini_tree_seq1)
Corpus2.append(mini_tree_seq2)
Corpus3.append(mini_tree_seq3)
print('\ndone!')