-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathapp_test.py
241 lines (215 loc) · 7.1 KB
/
app_test.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
231
232
233
234
235
236
237
238
239
240
241
# import streamlit as st
# from streamlit_chat import message
from chitchat.chitchat import chitchat, chitchat_batch
# import intent_classifier
# from nlp_pipeline import get_nel
from tqdm import tqdm
# from wikibot.wiki_ir import TopicBot
from wikibot.wikibot import get_wiki_batch_response
# import intent_classifier_albert
# from torch.utils.data import Dataset, DataLoader
import evaluate
import json
from PIL import Image
# image = Image.open('img-removebg-preview.png')
# st.image(image,width=300)
def batch(iterable, n=20):
l = len(iterable)
for ndx in range(0, l, n):
yield iterable[ndx:min(ndx + n, l)]
def get_response(prompt):
#Call the dialog manager api to get the response for the prompt
# message=prompt + ": Reply"
#dialogue manager rule or model get the prompt and call individual generator
# links = get_nel(prompt)
# if inference_intent_classifier_trained_albert.classify(prompt) == "chitchat":
# print("doing chitchat")
# message = chitchat_batch(prompt) #directly calling chitchat for testing
# else:
#perform entity recoq, linker, find relevant facts, perform paraphrasing and return
# print("doing wiki")
# message = topicBot.generator(prompt, links)
message = get_wiki_batch_response(prompt)
# if message == "":
# message = chitchat(prompt)
return message
# def get_intent():
def processpara(paras):
result ={}
for para in paras:
for qas in para['qas']:
question = qas['question']
print(question)
output = ""
try:
output=get_response(question)
except:
output="I didn't got that. I am sorry"
result[qas['id']] = output
return result
def squad_inference():
import pandas as pd
df = pd.read_csv('wiki_in.csv')
queries = df['bot_in'].to_list()
ans = []
que = []
for query in queries:
que.append(query)
if len(que) != 20:
continue
output=get_response(que)
que = []
ans.extend(output)
df = pd.DataFrame({'bot_out':ans})
df.to_csv("bot_out.csv")
# return ans
def squad ():
import pandas as pd
file = open("dev-v2.0.json")
data = json.load(file)
datas = data['data']
result = {}
ques = []
bot_out = []
for data in datas:
for para in data['paragraphs']:
for qas in para['qas']:
question = qas['question']
ques.append(question)
if len(ques) != 20:
continue
# print(question)
output = []
# output=get_response(ques)
# result[qas['id']] = output
# bot_out.extend(output)
# ques = []
# with open('result.json', 'w') as fp:
# json.dump(result, fp)
df = pd.DataFrame({'bot_in':ques})
df.to_csv("wiki_in.csv")
def bleu_inference():
import pandas as pd
count = 0
df = pd.read_csv("chitchat/emp_data.csv")
# data = CustomDataset(df)
# loader = DataLoader(data,batch_size=100)
# return
bot_out = []#["" for i in range(df.size)]
# print(len(bot_out))
for btch in tqdm(batch(df['send'].tolist())):
# for index, row in tqdm(df.iterrows()):
# print(btch)
count += 1
out = []
# try:
print(len(btch))
out = get_response(btch)
# except:
# out=["I didn't got that. I am sorry"]
# print(out)
# print(count)
# df['bot_out']= out
# print(out)
bot_out.extend(out)
# count += 1
df['bot_out']=bot_out
df.to_csv("emp_data_out.csv")
def bleu_inference_cc():
import pandas as pd
count = 0
df = pd.read_csv("chitchat/data_chitchat.csv")
# data = CustomDataset(df)
# loader = DataLoader(data,batch_size=100)
# return
bot_out = []#["" for i in range(df.size)]
# print(len(bot_out))
for btch in tqdm(batch(df['in'].tolist())):
# for index, row in tqdm(df.iterrows()):
# print(btch)
count += 1
out = []
# try:
print(len(btch))
out = get_response(btch)
# except:
# out=["I didn't got that. I am sorry"]
# print(out)
# print(count)
# df['bot_out']= out
# print(out)
bot_out.extend(out)
# count += 1
df['bot_out']=bot_out
df.to_csv("cc_data_out.csv")
def evaluator(metric, dataset):
import pandas as pd
bleu = evaluate.load(metric)
df = pd.read_csv(dataset)
bot_out = []
out = []
for index, row in df.iterrows():
if not pd.isnull(row['bot_out']) and not pd.isnull(row['recv']):
out.append(row['recv'].strip())
bot_out.append(row['bot_out'].strip())
results = bleu.compute(predictions=bot_out,references=out)
print(results)
def bert_score(metric, dataset):
import pandas as pd
bleu = evaluate.load(metric)
df = pd.read_csv(dataset)
bot_out = []
out = []
for index, row in df.iterrows():
if not pd.isnull(row['bot_out']) and not pd.isnull(row['recv']):
out.append(row['recv'].strip())
bot_out.append(row['bot_out'].strip())
results = bleu.compute(predictions=bot_out,references=out,lang="en")
print(results)
def bleurt_score(metric, dataset):
import pandas as pd
bleu = evaluate.load(metric, module_type="metric")
df = pd.read_csv(dataset)
bot_out = []
out = []
for index, row in df.iterrows():
if not pd.isnull(row['bot_out']) and not pd.isnull(row['recv']):
out.append(row['recv'].strip())
bot_out.append(row['bot_out'].strip())
results = bleu.compute(predictions=bot_out,references=out)
print(results)
def perplexity(metric, dataset):
import pandas as pd
bleu = evaluate.load(metric,module_type="metric")
df = pd.read_csv(dataset)
bot_out = []
out = []
for index, row in df.iterrows():
if not pd.isnull(row['bot_out']) and not pd.isnull(row['recv']):
out.append(row['recv'].strip())
bot_out.append(row['bot_out'].strip())
results = bleu.compute(model_id='gpt2',
add_start_token=False,
predictions=bot_out)
print(results)
def perplexity_wiki(metric, dataset):
import pandas as pd
bleu = evaluate.load(metric,module_type="metric")
df = pd.read_csv(dataset)
bot_out = []
out = []
for index, row in df.iterrows():
if not pd.isnull(row['bot_out']):
# out.append(row['recv'].strip())
bot_out.append(row['bot_out'].strip())
results = bleu.compute(model_id='gpt2',
add_start_token=False,
predictions=bot_out)
print(results)
# evaluator("bleu","chitchat/emp_data_out.csv")
# bert_score("bertscore","chitchat/emp_data_out.csv")
# evaluator("rouge","chitchat/emp_data_out.csv")
# bleurt_score("bleurt","chitchat/emp_data_out.csv")
# perplexity("perplexity","chitchat/emp_data_out.csv")
perplexity_wiki("perplexity","wiki_out_full.csv")
# squad()