-
Notifications
You must be signed in to change notification settings - Fork 0
/
fasttext_flask.py
267 lines (223 loc) · 9.06 KB
/
fasttext_flask.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
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
import flask
from flask import request
from flask import jsonify
import numpy as np
import pandas as pd
app = flask.Flask(__name__)
import subprocess
from pathlib import Path
import json
import string
import random
import fasttext
from sklearn.metrics import accuracy_score
from sklearn.metrics import classification_report
from mlflow import log_metric, set_tag, log_param, log_params, log_artifact, set_experiment, end_run, start_run, set_tracking_uri
do_ml_flow = False
def training(df_train, df_valid, model_name, clf_params):
params = {'lr':0.1, 'epoch':200, 'word_ngrams':2, 'bucket':200000, 'dim':300, 'loss':'softmax',
'autotuneDuration':1200, 'autotuneMetric':'f1'}
params.update(clf_params)
global do_ml_flow
if do_ml_flow:
log_params(params)
lr = params['lr']
epoch = params['epoch']
word_ngrams = params['word_ngrams']
bucket = params['bucket']
dim = params['dim']
loss = params['loss']
autotuneDuration = params['autotuneDuration'] # 300 for 5 minutes
autotuneMetric = params['autotuneMetric']
f_name_trn = 'tmp/'+''.join(random.choices(string.ascii_uppercase + string.digits, k=7))+'.txt'
f = open(f_name_trn,'w',encoding='utf8')
train_size=df_train.shape[0]
for i in range(df_train.shape[0]):
# print(i, end='\r')
label=str(df_train.loc[i,'label'])
label=label.replace(" ", "-")
label=label.lower()
body=str(df_train.loc[i,'text'])
body=body.translate(str.maketrans('','',string.punctuation))
body=body.replace('\n',' ')
body=body.lower()
f.write('__label__'+label+' '+body+'\n')
f.close()
# print('\nWritten')
if df_valid:
f_name_val = 'tmp/'+''.join(random.choices(string.ascii_uppercase + string.digits, k=7))+'.txt'
f = open(f_name_val,'w',encoding='utf8')
valid_size=df_valid.shape[0]
# print(valid_size)
for i in range(df_valid.shape[0]):
# print(i)
label = str(df_valid.loc[i,'label'])
label = label.replace(" ", "-")
label = label.lower()
body = str(df_valid.loc[i,'text'])
body = body.translate(str.maketrans('','',string.punctuation))
body = body.replace('\n',' ')
body = body.lower()
f.write('__label__'+label+' '+body+'\n')
f.close()
# print('\nWritten')
classifier = fasttext.train_supervised(f_name_trn, lr=lr, epoch=epoch,
word_ngrams=word_ngrams, bucket=bucket, dim=dim,
autotuneValidationFile=f_name_val,
autotuneDuration=autotuneDuration, #300 for 5 minutes
autotuneMetric=autotuneMetric,
loss=loss)
subprocess.call(f'rm {f_name_trn}', shell=True)
subprocess.call(f'rm {f_name_val}', shell=True)
return classifier
classifier = fasttext.train_supervised(f_name_trn, lr=lr, epoch=epoch,
word_ngrams=word_ngrams, bucket=bucket, dim=dim,
loss=loss)
subprocess.call(f'rm {f_name_trn}', shell=True)
return classifier
def testing(df_test, classifier=None, model_name=''):
if classifier==None:
classifier = fasttext.load_model(f'fasttext_models/{model_name}.bin')
sent = []
labels = []
conf = []
numbers = []
t_size = df_test.shape[0]
for i in range(df_test.shape[0]):
numbers.append(df_test.loc[i,'id'])
label = str(df_test.loc[i,'label'])
label = label.replace(" ", "-")
label = label.lower()
labels.append(label)
body = str(df_test.loc[i,'text'])
body = body.translate(str.maketrans('','',string.punctuation))
body = body.replace('\n',' ')
body = body.lower()
sent.append(body)
pred = []
for i, s in enumerate(sent):
prediction = classifier.predict(s)
pred.append(prediction[0][0][9:])
conf.append(prediction[1][0])
df_result = pd.DataFrame(columns=['id','real','pred','conf'])
df_result['id'] = numbers
df_result['pred'] = pred
df_result['conf'] = conf
df_result['real'] = labels
acc = accuracy_score(labels, pred)
report = classification_report(labels, pred, output_dict=True)
report = pd.DataFrame(report).T
# report = report[~report.index.isin(['accuracy', 'macro avg', 'weighted avg'])]
report_txt = classification_report(labels, pred)
conf_mat = pd.crosstab(df_result['real'], df_result['pred'], rownames=['Actual'], colnames=['Pred'])
global do_ml_flow
if do_ml_flow:
log_metric("Test Accuracy", acc)
df_result.to_excel(f'tmp/df_result -- {model_name}.xlsx', index=False)
report.to_excel(f'tmp/report -- {model_name}.xlsx')
with open(f'tmp/report -- {model_name}.txt', 'w') as f:
print(report_txt, file=f)
conf_mat.to_excel(f'tmp/conf_mat -- {model_name}.xlsx')
log_artifact(f'tmp/df_result -- {model_name}.xlsx')
log_artifact(f'tmp/report -- {model_name}.xlsx')
log_artifact(f'tmp/report -- {model_name}.txt')
log_artifact(f'tmp/conf_mat -- {model_name}.xlsx')
subprocess.call(f'rm "tmp/df_result -- {model_name}.xlsx"', shell=True)
subprocess.call(f'rm "tmp/report -- {model_name}.xlsx"', shell=True)
subprocess.call(f'rm "tmp/report -- {model_name}.txt"', shell=True)
subprocess.call(f'rm "tmp/conf_mat -- {model_name}.xlsx"', shell=True)
res = {}
res['df_res'] = df_result.to_dict()
res['acc'] = acc
res['report'] = report.to_dict()
res['report_txt'] = report_txt
res['conf_mat'] = conf_mat.to_dict()
return res
@app.route('/fasttext/', methods=['POST'])
def home():
df_train = request.json['train_data']
df_train = pd.DataFrame(df_train)
df_train.index = df_train.index.astype(int)
# print(type(df_train), df_train.shape, df_train.columns)
# print(df_train.index)
df_test = request.json['test_data']
df_test = pd.DataFrame(df_test)
# print(type(df_test), df_test.shape, df_test.columns)
df_test.index = df_test.index.astype(int)
try:
df_valid = request.json['valid_data']
df_valid = pd.DataFrame(df_valid)
# print(type(df_valid), df_valid.shape, df_valid.columns)
df_valid.index = df_valid.index.astype(int)
except:
df_valid = None
try:
clf_params = request.json['clf_params']
except:
clf_params = {}
try:
save_model = request.json['save_model']
except:
save_model = False
try:
model_name = request.json['model_name']
except:
model_name = ''.join(random.choices(string.ascii_uppercase + string.digits, k=7))
try:
ml_flow_params = request.json['ml_flow_params']
experiment_name = ml_flow_params['experiment_name']
run_name = ml_flow_params['run_name']
description = ml_flow_params['description']
global do_ml_flow
do_ml_flow = True
except:
pass
if do_ml_flow:
set_experiment(experiment_name)
start_run(run_name=run_name)
set_tag("mlflow.note.content", description)
clf = training(df_train, df_valid, model_name, clf_params)
if save_model:
clf.save_model(f'fasttext_models/{model_name}.bin')
res = testing(df_test, clf, model_name)
res['model_name'] = model_name
if do_ml_flow:
log_param("model_name", model_name)
end_run()
return jsonify(res)
@app.route('/fasttext/test', methods=['POST'])
def test():
df_test = request.json['test_data']
df_test = pd.DataFrame(df_test)
# print(type(df_test), df_test.shape, df_test.columns)
df_test.index = df_test.index.astype(int)
model_name = request.json['model_name']
my_file = Path(f'fasttext_models/{model_name}.bin')
if not my_file.is_file():
res = {'status':f'Model \'{model_name}\' Doesn\'t Exists'}
else:
res = testing(df_test, model_name=model_name)
return jsonify(res)
@app.route('/fasttext/delete', methods=['POST'])
def delete():
model_name = request.json['model_name']
my_file = Path(f'fasttext_models/{model_name}.bin')
if not my_file.is_file():
res = {'status':f'Model \'{model_name}\' Doesn\'t Exists'}
else:
subprocess.call(f'rm fasttext_models/{model_name}.bin', shell=True)
res = {'status':f'Model \'{model_name}\' Deleted'}
return jsonify(res)
@app.route('/fasttext/mlflow', methods=['POST'])
def mlflow():
res = {'url':'http://{system_ip}:3456/'}
return jsonify(res)
if __name__ == '__main__':
DATA_PATH = Path('fasttext_models/')
DATA_PATH.mkdir(exist_ok=True)
DATA_PATH = Path('fasttext_mlruns/')
DATA_PATH.mkdir(exist_ok=True)
DATA_PATH = Path('tmp/')
DATA_PATH.mkdir(exist_ok=True)
set_tracking_uri('./fasttext_mlruns/')
app.run(host='0.0.0.0', port=7655)