-
Notifications
You must be signed in to change notification settings - Fork 7
/
pipeline.py
230 lines (176 loc) · 7.17 KB
/
pipeline.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
import re
import os
import subprocess
import unicodedata
import argparse
from shutil import rmtree
stopping_chars=["@", "#", "&", "$"]
def is_number(s):
try:
float(s)
return True
except ValueError:
pass
try:
unicodedata.numeric(s)
return True
except (TypeError, ValueError):
pass
return False
def OPEN(filename):
lines=[]
with open(filename, "r") as f:
for line in f:
line=line.strip()
lines.append(line)
return lines
def savefile(filename,LIST):
with open(filename, 'w') as f:
for line in LIST:
f.write("%s\n" % line)
def processing_1(form):
form=form.replace('#', '').replace('[', '').replace(']', '').replace('<', '').replace('>', '').replace( '!', '').replace('?', '').replace('@c','').replace('@t','').replace('_','').replace(',','')
#x = re.sub(r"\[\.+\]","unk",text_line)
#x = re.sub(r"...","unk",x)
'''x = re.sub(r'\#', '', text_line)
x = re.sub(r"\_", "", x)
x = re.sub(r"\[", "", x)
x = re.sub(r"\]", "", x)
x = re.sub(r"\<", "", x)
x = re.sub(r"\>", "", x)
x = re.sub(r"\!", "", x)
x = re.sub(r"@c", "", x)
x = re.sub(r"@t", "", x)
x = re.sub(r",", "", x)
#x=re.sub(r"(x)+","x",x)
x = re.sub(r"\?", "", x)
x = x.split()
x = " ".join(x)
k = re.search(r"[a-wyzA-Z]+",x)
if k:
return x
else:
return "" '''
if(len(form)==0):
form="..."
return form
def Pipeline_start(lines):
Pipeline=[]
for i,line in enumerate(lines):
if len(line)>0 and is_number(line[0]):
index=line.find(".")
line=line[index+1:].strip()
text=processing_1(line)
Pipeline.append(text)
return Pipeline
def Pipeline_end(lines):
pipeline_result=[]
#POS=OPEN(output_dir+'pos_pipeline.txt')
#NER=OPEN(output_dir+'ner_pipeline.txt')
tr_en=OPEN(output_dir+'trans_pipeline.txt')
#pos_index=2;
index=0
for i,line in enumerate(lines):
pipeline_result.append(line)
if len(line)>0 and is_number(line[0]):
#pipeline_result.append(POS[pos_index])
#pipeline_result.append(NER[pos_index])
pipeline_result.append('#tr.en:'+tr_en[index])
index+=1
#pos_index+=3
return pipeline_result
def main():
if os.path.exists(output_dir):
rmtree(output_dir)
os.makedirs(output_dir)
lines=OPEN(input_path)
if(atf_file=="False"):
Pipeline=[]
print(len(lines))
for line in lines:
Pipeline.append(processing_1(line))
savefile(output_dir+'pipeline.txt',Pipeline)
else:
Pipeline=Pipeline_start(lines)
savefile(output_dir+'pipeline.txt',Pipeline)
#POS MODEL
print("\n Running Part of speech Model for Sumerian Language")
if Flair==False:
os.system(f'python3 {pos_path} -i {output_dir}pipeline.txt -o {output_dir}pos_pipeline.txt')
else:
from FLAIR.predict import Predictions
print("Using Flair Model")
inp=output_dir+'/pipeline.txt'
out=output_dir+'/pos_pipeline.txt'
obj=Predictions(inp,out,True,False)
obj.predict()
#NER MODEL
print("Running Named entity recognation Model for Sumerian Language")
if Flair==False:
os.system(f'python3 {ner_path} -i {output_dir}pipeline.txt -o {output_dir}ner_pipeline.txt')
else:
print("Using Flair Model")
inp=output_dir+'/pipeline.txt'
out=output_dir+'/ner_pipeline.txt'
obj=Predictions(inp,out,False,True)
obj.predict()
#Translation MODEL
print("Running Translation Model for Sumerian Language")
model_name = trans_path.split('/')[-1].split('.')[0]
if model_name == 'Transformer' or model_name == 'Back_Translation':
if GPU==False:
os.system(f'onmt_translate -model {trans_path} -src {output_dir}pipeline.txt -output {output_dir}trans_pipeline.txt -replace_unk -verbose')
else:
os.system(f'CUDA_VISIBLE_DEVICES=0 onmt_translate -model {trans_path} -src {output_dir}pipeline.txt -output {output_dir}trans_pipeline.txt -replace_unk -verbose -gpu 0')
if model_name == 'XLM' or model_name == 'MASS':
if GPU==False:
os.system(f'sh Translation_Models/evalXLM.sh ../{output_dir}pipeline.txt {model_name} ../{output_dir}trans_pipeline.txt')
else:
os.system(f'CUDA_VISIBLE_DEVICES=0 sh Translation_Models/evalXLM.sh ../{output_dir}pipeline.txt {model_name} ../{output_dir}trans_pipeline.txt')
#Converting POS_NER to conll
if(atf_file!="False"):
print("\n converting POS_NER to conll form")
os.system(f'python3 ATF_2_Conll/atf2conll_tags.py -i {input_path}')
if(atf_file=="False"):
print("\n Generating atf file output")
pipeline_result=Pipeline_end(lines)
savefile(output_dir+'pipeline_output.atf',pipeline_result)
print("\n Done!.. \n")
if __name__=='__main__':
parser = argparse.ArgumentParser()
parser.add_argument("-i","--input",help="Location of the Input ATF File ", default="ATF_INPUT/demo.atf")
parser.add_argument("-p","--pos",help="POS Model to be used from ['POS_CRF','POS_HMM','POS_Bi_LSTM','POS_Bi_LSTM_CRF'] (Case sensitive)", choices=['POS_CRF','POS_HMM','POS_Bi_LSTM','POS_Bi_LSTM_CRF'],default="POS_CRF" )
parser.add_argument("-n","--ner",help="NER Model to be used from ['NER_CRF','NER_Bi_LSTM','NER_Bi_LSTM_CRF'] (Case_sensitive)", choices=['NER_CRF','NER_Bi_LSTM','NER_Bi_LSTM_CRF'],default="NER_CRF" )
parser.add_argument("-t","--trans",help="Machine Translation Model to be used",choices=['Transformer', 'Back_Translation', 'XLM', 'MASS'], default="Back_Translation" )
parser.add_argument("-o","--output",help="Location of output Directory", default='ATF_OUTPUT/')
parser.add_argument("-a","--atf",help="Use of Text file rather than ATF", default=True)
parser.add_argument("-g","--gpu",help="Use of GPU if avaliable", default=False)
parser.add_argument("-f","--flair",help="Use of flair language model", default=False)
args=parser.parse_args()
input_path=args.input
pos_path='POS_Models/'+args.pos+'/prediction.py'
ner_path='NER_Models/'+args.ner+'/prediction.py'
trans_path='Translation_Models/'+args.trans+'.pt'
output_dir=args.output
atf_file=args.atf
GPU=args.gpu
Flair=args.flair
print("\n")
print("Input file is ", input_path)
print("POS Model path is ", pos_path)
print("NER Model path is ", ner_path)
print("Translation Model path is ", trans_path)
print("Output directory is", output_dir)
print("Using ATF file", atf_file)
print("GPU", GPU)
print("Flair_Model", Flair)
print("\n")
main()
#IF WE WANT TO USE CLASS
#pipeline=OPEN('ATF_OUTPUT/pipeline.txt')
#POS=POS_tag(Pipeline)
#savefile('ATF_OUTPUT/pipeline1.txt',POS)
#NER=NER_tag(Pipeline)
#savefile('ATF_OUTPUT/pipeline2.txt',NER)
#translations=Translation_tag(Pipeline)
#savefile('ATF_OUTPUT/pipeline3.txt',translations)