-
Notifications
You must be signed in to change notification settings - Fork 1
/
app.py
272 lines (217 loc) · 12 KB
/
app.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
268
269
270
271
272
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sat Jul 4 20:27:17 2020
@author: kush
"""
## import libraries
import streamlit as st
import codecs
import pandas as pd
import unidecode
import re
from PIL import Image
import pickle
from wordcloud import WordCloud,STOPWORDS
from googletrans import Translator
import matplotlib.pyplot as plt
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from sklearn import metrics
from sklearn.model_selection import train_test_split
from sklearn.metrics import plot_confusion_matrix, plot_roc_curve, plot_precision_recall_curve
from sklearn.metrics import precision_score,recall_score
def main():
# Load Hindi stopwords
f=codecs.open("hindi_stopwords.txt",encoding='utf-8')
stopwords=[x.strip() for x in f.readlines()]
# Load dataset
@st.cache(allow_output_mutation=True)
def load_dataset():
df = pd.read_csv('DataCSV/emotion.csv',encoding='utf-8')
return df
pass
# Wordcloud function
def wordCloud(df,emotion):
if emotion == 'Angry':
num = 0
color = 'Reds'
elif emotion == 'Happy':
num=1
color = 'Greens'
elif emotion == 'Sad':
num=2
color = 'Purples'
else:
num=3
color = 'spring'
plt.figure(figsize = (8,8))
wc = WordCloud(font_path='Lohit-Devanagari.ttf',width=1400, height=1300,stopwords = stopwords,colormap='Reds',collocations=False).generate(" ".join(df[df.Emotion == num].Text))
plt.imshow(wc , interpolation = 'bilinear')
plt.axis('off')
plt.tight_layout(pad=0)
plt.savefig('worldCloud.jpg')
img = Image.open("worldCloud.jpg")
return img
def remove_stopwords(text):
text = [t for t in text.split() if t not in stopwords]
text = " ".join(text)
return text
def count_vect(df):
df['Text'] = df['Text'].apply(lambda x: remove_stopwords(x))
X = df['Text'].values
y = df['Emotion'].values
def tokenize(i):
return i.split(' ')
cv = CountVectorizer(min_df=2, ngram_range=(1, 3), encoding='utf-8',tokenizer=tokenize)
cv.fit(X)
X_vector = cv.transform(X) # Getting Bag of words representation for all the documents
X_train, X_test, y_train, y_test = train_test_split(X_vector, y, test_size=0.1, random_state=48,stratify=y)
return X_train, X_test, y_train, y_test,cv
def predict_emotion(model,cv,text):
text = [text]
cvect = cv.transform(text).toarray()
pred = model.predict(cvect)
return pred
def translate(text):
trans = Translator()
t = trans.translate(text,src='en',dest='hi')
return t.src, t.dest, t.text
st.markdown("<body style='background-color:white;'><h1 style='text-align: center; color: blue;'>Hindi Emotion Analysis</h1></body>", unsafe_allow_html=True)
st.markdown("<body style='background-color:white;'><h3 style='text-align: center; color: green;'>SELECT YOUR ACTIVITIES FROM THE SIDEBAR 👈</h3></body>", unsafe_allow_html=True)
if st.checkbox('Show data'):
df = load_dataset()
st.dataframe(df)
st.sidebar.subheader("Perform the following task")
select = ['Select','Word Cloud','Run Your Model','Run Pretrained Model']
option = st.sidebar.selectbox("",select)
if option==select[0]:
pass
# WordCloud
elif option == select[1]:
st.markdown("<body style='background-color:white;'><h2 style='text-align: center; color: orange;'>Word Cloud For</h2></body>", unsafe_allow_html=True)
emotion = st.radio(' ',('Angry','Happy','Sad','Neutral'))
df = load_dataset()
img = wordCloud(df,emotion)
st.image(img)
# Train your model
elif option == select[2]:
classifier = st.sidebar.selectbox("Classifier",("Support Vector Machine","Logistic Regression","Random Forest"))
if classifier =="Support Vector Machine":
df = load_dataset()
X_train, X_test, y_train, y_test,cv = count_vect(df)
st.sidebar.subheader("Model Hyperparameters")
C = st.sidebar.number_input("C (Regularization parameter)",0.01,10.0,step=0.01,key='C')
kernel = st.sidebar.radio("kernel",("rbf","linear"),key='kernel')
gamma = st.sidebar.radio("Gamma (Kernel Coefficientt)", ("scale","auto"),key='gamma')
if st.sidebar.button("Classify",key='classify1'):
st.markdown("<body style='background-color:white;'><h2 style='text-align: center; color: orange;'>Support Vector Machine Results</h2></body>", unsafe_allow_html=True)
model = SVC(C=C,kernel=kernel,gamma=gamma)
model.fit(X_train,y_train)
accuracy = model.score(X_test,y_test)
y_pred = model.predict(X_test)
precision = metrics.precision_score(y_pred, y_test,average='weighted')
f1_score = metrics.f1_score(y_pred, y_test,average='weighted')
recall = metrics.recall_score(y_pred, y_test,average='weighted')
st.success("Accuracy ---> {} % ".format(accuracy.round(5)*100))
st.success("Precision---> {}".format(precision.round(2)))
st.success("Recall ---> {}".format(recall.round(2)))
st.success("F1-score ---> {} ".format(f1_score.round(2)))
if classifier == "Logistic Regression":
df = load_dataset()
X_train, X_test, y_train, y_test,cv = count_vect(df)
st.sidebar.subheader("Model Hyperparameters")
C = st.sidebar.number_input("C (Regularization parameter)",0.01,10.0,step=0.01,key='C_LR')
max_iter = st.sidebar.slider("Maximum number of iterations",100,500,key='max_iter')
if st.sidebar.button("Classify",key='classify2'):
st.markdown("<body style='background-color:white;'><h2 style='text-align: center; color: orange;'>Logistic Regression Results</h2></body>", unsafe_allow_html=True)
model = LogisticRegression(C=C,max_iter=max_iter)
model.fit(X_train,y_train)
accuracy = model.score(X_test,y_test)
y_pred = model.predict(X_test)
precision = metrics.precision_score(y_pred, y_test,average='weighted')
f1_score = metrics.f1_score(y_pred, y_test,average='weighted')
recall = metrics.recall_score(y_pred, y_test,average='weighted')
st.success("Accuracy ---> {} % ".format(accuracy.round(5)*100))
st.success("Precision---> {}".format(precision.round(2)))
st.success("Recall ---> {}".format(recall.round(2)))
st.success("F1-score ---> {} ".format(f1_score.round(2)))
if classifier == "Random Forest":
df = load_dataset()
X_train, X_test, y_train, y_test,cv = count_vect(df)
st.sidebar.subheader("Model Hyperparameters")
n_estim = st.sidebar.number_input("The number of trees in forest",100,5000,step=10,key='nest')
max_depth = st.sidebar.number_input("The max depth of the trees",1,20,step=1,key='mdepth')
if st.sidebar.button("Classify",key='classify3'):
st.markdown("<body style='background-color:white;'><h2 style='text-align: center; color: orange;'>Random Forest Results</h2></body>", unsafe_allow_html=True)
model = RandomForestClassifier(n_estimators=n_estim,max_depth=max_depth)
model.fit(X_train,y_train)
accuracy = model.score(X_test,y_test)
y_pred = model.predict(X_test)
precision = metrics.precision_score(y_pred, y_test,average='weighted')
f1_score = metrics.f1_score(y_pred, y_test,average='weighted')
recall = metrics.recall_score(y_pred, y_test,average='weighted')
st.success("Accuracy ---> {} % ".format(accuracy.round(5)*100))
st.success("Precision---> {}".format(precision.round(2)))
st.success("Recall ---> {}".format(recall.round(2)))
st.success("F1-score ---> {} ".format(f1_score.round(2)))
# Train pre-trained model
else:
model = pickle.load(open('hindi_model.pkl','rb'))
df = load_dataset()
X_train, X_test, y_train, y_test,cv = count_vect(df)
accuracy = model.score(X_test,y_test)
y_pred = model.predict(X_test)
precision = metrics.precision_score(y_pred, y_test,average='weighted')
f1_score = metrics.f1_score(y_pred, y_test,average='weighted')
recall = metrics.recall_score(y_pred, y_test,average='weighted')
st.markdown("<body style='background-color:white;'><h2 style='text-align: center; color: orange;'>Pre-trained Model Results</h2></body>", unsafe_allow_html=True)
st.success("Accuracy ---> {} % ".format(accuracy.round(5)*100))
st.success("Precision---> {}".format(precision.round(2)))
st.success("Recall ---> {}".format(recall.round(2)))
st.success("F1-score ---> {} ".format(f1_score.round(2)))
# Predict on New Text
st.markdown("<body style='background-color:blue;'><h2 style='text-align: center; color: white;'> Predict on New Text</h2></body>", unsafe_allow_html=True)
st.markdown("<body style='background-color:white;'><h3 style='text-align: center; color: red;'> Do You Know Hindi ????</h3></body>", unsafe_allow_html=True)
sel = ['Select','Yes, I Know',"No, I don't Know"]
yn = st.selectbox("",sel)
if yn == 'Select':
pass
elif yn == 'Yes, I Know':
st.markdown("<body style='background-color:white;'><h3 style='text-align: center; color: brown;'> Great, Now enter the text in Hindi to predict the emotion</h3></body>", unsafe_allow_html=True)
text = st.text_input("")
if st.button('Predict Emotion'):
pred = predict_emotion(model,cv,text)
prediction = int(pred)
if prediction ==0:
prediction = 'Angry'
elif prediction == 1:
prediction ='Happy'
elif prediction == 2:
prediction = 'Sad'
else:
prediction = 'Neutral'
st.success('The emotion behind the sentence "{}" is {}: '.format(text,prediction))
else:
st.markdown("<body style='background-color:white;'><h3 style='text-align: center; color: brown;'> Don't Worry, enter the text in English to translate it in Hindi</h3></body>", unsafe_allow_html=True)
st.markdown("#### ")
st.write("Please enter a long sentence for better prediction")
text = st.text_input("")
if st.button('Predict Emotion'):
src,dest, tex = translate(text)
st.info("Your entered text is translated to {}".format(tex))
pred = predict_emotion(model,cv,tex)
prediction = int(pred)
if prediction ==0:
prediction = 'Angry 😡'
elif prediction == 1:
prediction ='Happy 😀'
elif prediction == 2:
prediction = 'Sad 😔'
else:
prediction = 'Neutral 😐'
st.success('The emotion behind the sentence "{}" is {}: '.format(tex,prediction))
if __name__ == '__main__':
main()