Spaces:
Running
Running
Update src/app.py
Browse files- src/app.py +180 -462
src/app.py
CHANGED
@@ -4,520 +4,238 @@ import pandas as pd
|
|
4 |
import numpy as np
|
5 |
from pathlib import Path
|
6 |
import sys
|
|
|
7 |
import plotly.graph_objects as go
|
8 |
from transformers import BertTokenizer
|
9 |
import nltk
|
10 |
|
11 |
# Download required NLTK data
|
12 |
-
|
13 |
-
'tokenizers/punkt'
|
14 |
-
|
15 |
-
'
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
|
|
|
|
|
|
|
|
|
|
23 |
|
24 |
# Add project root to Python path
|
25 |
project_root = Path(__file__).parent.parent
|
26 |
sys.path.append(str(project_root))
|
27 |
|
28 |
from src.models.hybrid_model import HybridFakeNewsDetector
|
29 |
-
from src.config.config import
|
30 |
from src.data.preprocessor import TextPreprocessor
|
31 |
|
32 |
-
#
|
33 |
-
st.markdown("""
|
34 |
-
<style>
|
35 |
-
@import url('https://fonts.googleapis.com/css2?family=Poppins:wght@200;300;400;500;600;700&display=swap');
|
36 |
-
|
37 |
-
* {
|
38 |
-
font-family: 'Poppins', sans-serif !important;
|
39 |
-
box-sizing: border-box;
|
40 |
-
}
|
41 |
-
|
42 |
-
.stApp {
|
43 |
-
background: #ffffff;
|
44 |
-
min-height: 100vh;
|
45 |
-
color: #1f2a44;
|
46 |
-
}
|
47 |
-
|
48 |
-
#MainMenu {visibility: hidden;}
|
49 |
-
footer {visibility: hidden;}
|
50 |
-
.stDeployButton {display: none;}
|
51 |
-
header {visibility: hidden;}
|
52 |
-
.stApp > header {visibility: hidden;}
|
53 |
-
|
54 |
-
/* Main Container */
|
55 |
-
.main-container {
|
56 |
-
max-width: 1200px;
|
57 |
-
margin: 0 auto;
|
58 |
-
padding: 1rem 2rem;
|
59 |
-
}
|
60 |
-
|
61 |
-
/* Header Section */
|
62 |
-
.header-section {
|
63 |
-
text-align: center;
|
64 |
-
margin-bottom: 2.5rem;
|
65 |
-
padding: 1.5rem 0;
|
66 |
-
}
|
67 |
-
|
68 |
-
.header-title {
|
69 |
-
font-size: 2.25rem;
|
70 |
-
font-weight: 700;
|
71 |
-
color: #1f2a44;
|
72 |
-
margin: 0;
|
73 |
-
}
|
74 |
-
|
75 |
-
/* Hero Section */
|
76 |
-
.hero {
|
77 |
-
display: flex;
|
78 |
-
align-items: center;
|
79 |
-
gap: 2rem;
|
80 |
-
margin-bottom: 2rem;
|
81 |
-
padding: 0 1rem;
|
82 |
-
}
|
83 |
-
|
84 |
-
.hero-left {
|
85 |
-
flex: 1;
|
86 |
-
padding: 1.5rem;
|
87 |
-
}
|
88 |
-
|
89 |
-
.hero-right {
|
90 |
-
flex: 1;
|
91 |
-
display: flex;
|
92 |
-
align-items: center;
|
93 |
-
justify-content: center;
|
94 |
-
}
|
95 |
-
|
96 |
-
.hero-right img {
|
97 |
-
max-width: 100%;
|
98 |
-
height: auto;
|
99 |
-
border-radius: 8px;
|
100 |
-
object-fit: cover;
|
101 |
-
}
|
102 |
-
|
103 |
-
.hero-title {
|
104 |
-
font-size: 2.5rem;
|
105 |
-
font-weight: 700;
|
106 |
-
color: #1f2a44;
|
107 |
-
margin-bottom: 0.5rem;
|
108 |
-
}
|
109 |
-
|
110 |
-
.hero-text {
|
111 |
-
font-size: 1rem;
|
112 |
-
color: #6b7280;
|
113 |
-
line-height: 1.6;
|
114 |
-
max-width: 450px;
|
115 |
-
}
|
116 |
-
|
117 |
-
/* About Section */
|
118 |
-
.about-section {
|
119 |
-
margin-bottom: 2rem;
|
120 |
-
text-align: center;
|
121 |
-
padding: 0 1rem;
|
122 |
-
}
|
123 |
-
|
124 |
-
.about-title {
|
125 |
-
font-size: 1.75rem;
|
126 |
-
font-weight: 600;
|
127 |
-
color: #1f2a44;
|
128 |
-
margin-bottom: 0.5rem;
|
129 |
-
}
|
130 |
-
|
131 |
-
.about-text {
|
132 |
-
font-size: 0.95rem;
|
133 |
-
color: #6b7280;
|
134 |
-
line-height: 1.6;
|
135 |
-
max-width: 600px;
|
136 |
-
margin: 0 auto;
|
137 |
-
}
|
138 |
-
|
139 |
-
/* Input Section */
|
140 |
-
.input-container {
|
141 |
-
max-width: 800px;
|
142 |
-
margin: 0 auto;
|
143 |
-
}
|
144 |
-
|
145 |
-
.stTextArea > div > div > textarea {
|
146 |
-
border-radius: 8px !important;
|
147 |
-
border: 1px solid #d1d5db !important;
|
148 |
-
padding: 1rem !important;
|
149 |
-
font-size: 1rem !important;
|
150 |
-
background: #ffffff !important;
|
151 |
-
min-height: 150px !important;
|
152 |
-
transition: all 0.2s ease !important;
|
153 |
-
}
|
154 |
-
|
155 |
-
.stTextArea > div > div > textarea:focus {
|
156 |
-
border-color: #6366f1 !important;
|
157 |
-
box-shadow: 0 0 0 2px rgba(99, 102, 241, 0.1) !important;
|
158 |
-
outline: none !important;
|
159 |
-
}
|
160 |
-
|
161 |
-
.stTextArea > div > div > textarea::placeholder {
|
162 |
-
color: #9ca3af !important;
|
163 |
-
}
|
164 |
-
|
165 |
-
/* Button Styling */
|
166 |
-
.stButton > button {
|
167 |
-
background: #6366f1 !important;
|
168 |
-
color: white !important;
|
169 |
-
border-radius: 8px !important;
|
170 |
-
padding: 0.75rem 2rem !important;
|
171 |
-
font-size: 1rem !important;
|
172 |
-
font-weight: 600 !important;
|
173 |
-
transition: all 0.2s ease !important;
|
174 |
-
border: none !important;
|
175 |
-
width: 100% !important;
|
176 |
-
max-width: 300px;
|
177 |
-
}
|
178 |
-
|
179 |
-
.stButton > button:hover {
|
180 |
-
background: #4f46e5 !important;
|
181 |
-
transform: translateY(-1px) !important;
|
182 |
-
}
|
183 |
-
|
184 |
-
/* Results Section */
|
185 |
-
.results-container {
|
186 |
-
margin-top: 1rem;
|
187 |
-
padding: 1rem;
|
188 |
-
border-radius: 8px;
|
189 |
-
max-width: 1200px;
|
190 |
-
margin-left: auto;
|
191 |
-
margin-right: auto;
|
192 |
-
}
|
193 |
-
|
194 |
-
.result-card {
|
195 |
-
padding: 1rem;
|
196 |
-
border-radius: 8px;
|
197 |
-
border-left: 4px solid transparent;
|
198 |
-
margin-bottom: 1rem;
|
199 |
-
}
|
200 |
-
|
201 |
-
.fake-news {
|
202 |
-
background: #fef2f2;
|
203 |
-
border-left-color: #ef4444;
|
204 |
-
}
|
205 |
-
|
206 |
-
.real-news {
|
207 |
-
background: #ecfdf5;
|
208 |
-
border-left-color: #10b981;
|
209 |
-
}
|
210 |
-
|
211 |
-
.prediction-badge {
|
212 |
-
font-weight: 600;
|
213 |
-
font-size: 1rem;
|
214 |
-
margin-bottom: 0.5rem;
|
215 |
-
display: flex;
|
216 |
-
align-items: center;
|
217 |
-
gap: 0.5rem;
|
218 |
-
}
|
219 |
-
|
220 |
-
.confidence-score {
|
221 |
-
font-weight: 600;
|
222 |
-
margin-left: auto;
|
223 |
-
font-size: 1rem;
|
224 |
-
}
|
225 |
-
|
226 |
-
/* Chart Containers */
|
227 |
-
.chart-container {
|
228 |
-
padding: 1rem;
|
229 |
-
border-radius: 8px;
|
230 |
-
margin: 1rem 0;
|
231 |
-
max-width: 1200px;
|
232 |
-
margin-left: auto;
|
233 |
-
margin-right: auto;
|
234 |
-
}
|
235 |
-
|
236 |
-
/* Footer */
|
237 |
-
.footer {
|
238 |
-
border-top: 1px solid #e5e7eb;
|
239 |
-
padding: 1.5rem 0;
|
240 |
-
text-align: center;
|
241 |
-
max-width: 1200px;
|
242 |
-
margin: 2rem auto 0;
|
243 |
-
}
|
244 |
-
|
245 |
-
/* Responsive Design */
|
246 |
-
@media (max-width: 1024px) {
|
247 |
-
.hero {
|
248 |
-
flex-direction: column;
|
249 |
-
text-align: center;
|
250 |
-
}
|
251 |
-
.hero-right img {
|
252 |
-
max-width: 80%;
|
253 |
-
}
|
254 |
-
}
|
255 |
-
|
256 |
-
@media (max-width: 768px) {
|
257 |
-
.header-title {
|
258 |
-
font-size: 1.75rem;
|
259 |
-
}
|
260 |
-
.hero-title {
|
261 |
-
font-size: 2rem;
|
262 |
-
}
|
263 |
-
.hero-text {
|
264 |
-
font-size: 0.9rem;
|
265 |
-
}
|
266 |
-
.about-title {
|
267 |
-
font-size: 1.5rem;
|
268 |
-
}
|
269 |
-
.about-text {
|
270 |
-
font-size: 0.9rem;
|
271 |
-
}
|
272 |
-
}
|
273 |
-
|
274 |
-
@media (max-width: 480px) {
|
275 |
-
.header-title {
|
276 |
-
font-size: 1.5rem;
|
277 |
-
}
|
278 |
-
.hero-title {
|
279 |
-
font-size: 1.75rem;
|
280 |
-
}
|
281 |
-
.hero-text {
|
282 |
-
font-size: 0.85rem;
|
283 |
-
}
|
284 |
-
.about-title {
|
285 |
-
font-size: 1.25rem;
|
286 |
-
}
|
287 |
-
.about-text {
|
288 |
-
font-size: 0.85rem;
|
289 |
-
}
|
290 |
-
}
|
291 |
-
</style>
|
292 |
-
""", unsafe_allow_html=True)
|
293 |
|
294 |
@st.cache_resource
|
295 |
-
def load_model_and_tokenizer()
|
296 |
"""Load the model and tokenizer (cached)."""
|
297 |
-
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
|
303 |
-
|
304 |
-
|
305 |
-
|
306 |
-
|
307 |
-
|
308 |
-
|
309 |
-
|
310 |
-
|
311 |
-
|
312 |
-
|
313 |
-
|
314 |
-
|
315 |
-
|
316 |
-
|
317 |
-
|
|
|
|
|
318 |
|
319 |
@st.cache_resource
|
320 |
-
def get_preprocessor()
|
321 |
"""Get the text preprocessor (cached)."""
|
322 |
-
|
323 |
-
return TextPreprocessor()
|
324 |
-
except Exception as e:
|
325 |
-
st.error(f"Error initializing preprocessor: {str(e)}")
|
326 |
-
return None
|
327 |
|
328 |
-
def predict_news(text
|
329 |
"""Predict if the given news is fake or real."""
|
|
|
330 |
model, tokenizer = load_model_and_tokenizer()
|
331 |
-
if model is None or tokenizer is None:
|
332 |
-
return None
|
333 |
preprocessor = get_preprocessor()
|
334 |
-
|
335 |
-
|
336 |
-
|
337 |
-
|
338 |
-
|
339 |
-
|
340 |
-
|
341 |
-
|
342 |
-
|
343 |
-
|
344 |
-
|
345 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
346 |
)
|
347 |
-
|
348 |
-
|
349 |
-
|
350 |
-
|
351 |
-
|
352 |
-
|
353 |
-
|
354 |
-
|
355 |
-
|
356 |
-
|
357 |
-
|
358 |
-
|
359 |
-
'
|
360 |
-
'
|
361 |
-
|
362 |
-
|
363 |
-
|
364 |
-
|
365 |
-
|
366 |
-
|
367 |
-
st.error(f"Prediction error: {str(e)}")
|
368 |
-
return None
|
369 |
-
|
370 |
-
def plot_confidence(probabilities: dict) -> go.Figure:
|
371 |
-
"""Plot prediction confidence with simplified styling."""
|
372 |
-
if not probabilities or not isinstance(probabilities, dict):
|
373 |
-
return go.Figure()
|
374 |
fig = go.Figure(data=[
|
375 |
go.Bar(
|
376 |
x=list(probabilities.keys()),
|
377 |
y=list(probabilities.values()),
|
378 |
-
text=[f'{p:.
|
379 |
textposition='auto',
|
380 |
-
marker=dict(
|
381 |
-
color=['#10b981', '#ef4444'],
|
382 |
-
line=dict(color='#ffffff', width=1),
|
383 |
-
),
|
384 |
)
|
385 |
])
|
|
|
386 |
fig.update_layout(
|
387 |
-
title=
|
388 |
-
|
389 |
-
|
390 |
-
|
391 |
-
height=300,
|
392 |
-
margin=dict(t=60, b=60)
|
393 |
)
|
|
|
394 |
return fig
|
395 |
|
396 |
-
def plot_attention(text
|
397 |
-
"""Plot attention weights
|
398 |
-
|
399 |
-
|
400 |
-
|
401 |
-
|
402 |
if isinstance(attention_weights, (list, np.ndarray)):
|
403 |
attention_weights = np.array(attention_weights).flatten()
|
404 |
-
|
405 |
-
|
|
|
|
|
406 |
fig = go.Figure(data=[
|
407 |
go.Bar(
|
408 |
x=tokens,
|
409 |
y=attention_weights,
|
410 |
-
text=
|
411 |
textposition='auto',
|
412 |
-
marker=dict(color=colors),
|
413 |
)
|
414 |
])
|
|
|
415 |
fig.update_layout(
|
416 |
-
title=
|
417 |
-
|
418 |
-
|
419 |
-
|
420 |
-
height=350,
|
421 |
-
margin=dict(t=60, b=80)
|
422 |
)
|
|
|
423 |
return fig
|
424 |
|
425 |
def main():
|
426 |
-
|
427 |
-
st.
|
428 |
-
|
429 |
-
|
430 |
-
|
431 |
-
|
432 |
-
|
433 |
-
|
434 |
-
"""
|
435 |
-
|
436 |
-
|
437 |
-
|
438 |
-
|
439 |
-
|
440 |
-
|
441 |
-
|
442 |
-
|
443 |
-
|
444 |
-
|
445 |
-
|
446 |
-
<img src="https://images.pexels.com/photos/267350/pexels-photo-267350.jpeg?auto=compress&cs=tinysrgb&w=500" alt="Fake News Illustration" onerror="this.src='https://via.placeholder.com/500x300.png?text=Fake+News+Illustration'">
|
447 |
-
</div>
|
448 |
-
</div>
|
449 |
-
""", unsafe_allow_html=True)
|
450 |
-
|
451 |
-
# About Section
|
452 |
-
st.markdown("""
|
453 |
-
<div class="about-section">
|
454 |
-
<h2 class="about-title">About TruthCheck</h2>
|
455 |
-
<p class="about-text">
|
456 |
-
TruthCheck harnesses a hybrid BERT-BiLSTM model to detect fake news with high precision. Simply paste an article below to analyze its authenticity instantly.
|
457 |
-
</p>
|
458 |
-
</div>
|
459 |
-
""", unsafe_allow_html=True)
|
460 |
-
|
461 |
-
# Input Section
|
462 |
-
st.markdown('<div class="input-container">', unsafe_allow_html=True)
|
463 |
news_text = st.text_area(
|
464 |
-
"
|
465 |
-
height=
|
466 |
-
placeholder="Paste your news article here
|
467 |
-
key="news_input"
|
468 |
)
|
469 |
-
|
470 |
-
|
471 |
-
|
472 |
-
|
473 |
-
|
474 |
-
analyze_button = st.button("🔍 Analyze Now", key="analyze_button")
|
475 |
-
|
476 |
-
if analyze_button:
|
477 |
-
if news_text and len(news_text.strip()) > 10:
|
478 |
-
with st.spinner("Analyzing article..."):
|
479 |
result = predict_news(news_text)
|
480 |
-
|
481 |
-
|
482 |
-
|
483 |
-
|
484 |
-
|
485 |
-
|
486 |
-
|
487 |
-
|
488 |
-
|
489 |
-
|
490 |
-
|
491 |
-
|
492 |
-
|
493 |
-
|
494 |
-
|
495 |
-
|
496 |
-
|
497 |
-
|
498 |
-
|
499 |
-
|
500 |
-
|
501 |
-
|
502 |
-
|
503 |
-
|
504 |
-
|
505 |
-
|
506 |
-
|
507 |
-
|
508 |
-
|
509 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
510 |
else:
|
511 |
-
st.
|
512 |
-
|
513 |
-
# Footer
|
514 |
-
st.markdown("---")
|
515 |
-
st.markdown(
|
516 |
-
'<p style="text-align: center; font-weight: 600; font-size: 16px;">💻 Developed with ❤️ using Streamlit | © 2025</p>',
|
517 |
-
unsafe_allow_html=True
|
518 |
-
)
|
519 |
-
|
520 |
-
st.markdown('</div>', unsafe_allow_html=True) # Close main-container
|
521 |
|
522 |
if __name__ == "__main__":
|
523 |
-
main()
|
|
|
4 |
import numpy as np
|
5 |
from pathlib import Path
|
6 |
import sys
|
7 |
+
import plotly.express as px
|
8 |
import plotly.graph_objects as go
|
9 |
from transformers import BertTokenizer
|
10 |
import nltk
|
11 |
|
12 |
# Download required NLTK data
|
13 |
+
try:
|
14 |
+
nltk.data.find('tokenizers/punkt')
|
15 |
+
except LookupError:
|
16 |
+
nltk.download('punkt')
|
17 |
+
try:
|
18 |
+
nltk.data.find('corpora/stopwords')
|
19 |
+
except LookupError:
|
20 |
+
nltk.download('stopwords')
|
21 |
+
try:
|
22 |
+
nltk.data.find('tokenizers/punkt_tab')
|
23 |
+
except LookupError:
|
24 |
+
nltk.download('punkt_tab')
|
25 |
+
try:
|
26 |
+
nltk.data.find('corpora/wordnet')
|
27 |
+
except LookupError:
|
28 |
+
nltk.download('wordnet')
|
29 |
|
30 |
# Add project root to Python path
|
31 |
project_root = Path(__file__).parent.parent
|
32 |
sys.path.append(str(project_root))
|
33 |
|
34 |
from src.models.hybrid_model import HybridFakeNewsDetector
|
35 |
+
from src.config.config import *
|
36 |
from src.data.preprocessor import TextPreprocessor
|
37 |
|
38 |
+
# Page config is set in main app.py
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
|
40 |
@st.cache_resource
|
41 |
+
def load_model_and_tokenizer():
|
42 |
"""Load the model and tokenizer (cached)."""
|
43 |
+
# Initialize model
|
44 |
+
model = HybridFakeNewsDetector(
|
45 |
+
bert_model_name=BERT_MODEL_NAME,
|
46 |
+
lstm_hidden_size=LSTM_HIDDEN_SIZE,
|
47 |
+
lstm_num_layers=LSTM_NUM_LAYERS,
|
48 |
+
dropout_rate=DROPOUT_RATE
|
49 |
+
)
|
50 |
+
|
51 |
+
# Load trained weights
|
52 |
+
state_dict = torch.load(SAVED_MODELS_DIR / "final_model.pt", map_location=torch.device('cpu'))
|
53 |
+
|
54 |
+
# Filter out unexpected keys
|
55 |
+
model_state_dict = model.state_dict()
|
56 |
+
filtered_state_dict = {k: v for k, v in state_dict.items() if k in model_state_dict}
|
57 |
+
|
58 |
+
# Load the filtered state dict
|
59 |
+
model.load_state_dict(filtered_state_dict, strict=False)
|
60 |
+
model.eval()
|
61 |
+
|
62 |
+
# Initialize tokenizer
|
63 |
+
tokenizer = BertTokenizer.from_pretrained(BERT_MODEL_NAME)
|
64 |
+
|
65 |
+
return model, tokenizer
|
66 |
|
67 |
@st.cache_resource
|
68 |
+
def get_preprocessor():
|
69 |
"""Get the text preprocessor (cached)."""
|
70 |
+
return TextPreprocessor()
|
|
|
|
|
|
|
|
|
71 |
|
72 |
+
def predict_news(text):
|
73 |
"""Predict if the given news is fake or real."""
|
74 |
+
# Get model, tokenizer, and preprocessor from cache
|
75 |
model, tokenizer = load_model_and_tokenizer()
|
|
|
|
|
76 |
preprocessor = get_preprocessor()
|
77 |
+
|
78 |
+
# Preprocess text
|
79 |
+
processed_text = preprocessor.preprocess_text(text)
|
80 |
+
|
81 |
+
# Tokenize
|
82 |
+
encoding = tokenizer.encode_plus(
|
83 |
+
processed_text,
|
84 |
+
add_special_tokens=True,
|
85 |
+
max_length=MAX_SEQUENCE_LENGTH,
|
86 |
+
padding='max_length',
|
87 |
+
truncation=True,
|
88 |
+
return_attention_mask=True,
|
89 |
+
return_tensors='pt'
|
90 |
+
)
|
91 |
+
|
92 |
+
# Get prediction
|
93 |
+
with torch.no_grad():
|
94 |
+
outputs = model(
|
95 |
+
encoding['input_ids'],
|
96 |
+
encoding['attention_mask']
|
97 |
)
|
98 |
+
probabilities = torch.softmax(outputs['logits'], dim=1)
|
99 |
+
prediction = torch.argmax(outputs['logits'], dim=1)
|
100 |
+
attention_weights = outputs['attention_weights']
|
101 |
+
|
102 |
+
# Convert attention weights to numpy and get the first sequence
|
103 |
+
attention_weights_np = attention_weights[0].cpu().numpy()
|
104 |
+
|
105 |
+
return {
|
106 |
+
'prediction': prediction.item(),
|
107 |
+
'label': 'FAKE' if prediction.item() == 1 else 'REAL',
|
108 |
+
'confidence': torch.max(probabilities, dim=1)[0].item(),
|
109 |
+
'probabilities': {
|
110 |
+
'REAL': probabilities[0][0].item(),
|
111 |
+
'FAKE': probabilities[0][1].item()
|
112 |
+
},
|
113 |
+
'attention_weights': attention_weights_np
|
114 |
+
}
|
115 |
+
|
116 |
+
def plot_confidence(probabilities):
|
117 |
+
"""Plot prediction confidence."""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
118 |
fig = go.Figure(data=[
|
119 |
go.Bar(
|
120 |
x=list(probabilities.keys()),
|
121 |
y=list(probabilities.values()),
|
122 |
+
text=[f'{p:.2%}' for p in probabilities.values()],
|
123 |
textposition='auto',
|
|
|
|
|
|
|
|
|
124 |
)
|
125 |
])
|
126 |
+
|
127 |
fig.update_layout(
|
128 |
+
title='Prediction Confidence',
|
129 |
+
xaxis_title='Class',
|
130 |
+
yaxis_title='Probability',
|
131 |
+
yaxis_range=[0, 1]
|
|
|
|
|
132 |
)
|
133 |
+
|
134 |
return fig
|
135 |
|
136 |
+
def plot_attention(text, attention_weights):
|
137 |
+
"""Plot attention weights."""
|
138 |
+
tokens = text.split()
|
139 |
+
attention_weights = attention_weights[:len(tokens)] # Truncate to match tokens
|
140 |
+
|
141 |
+
# Ensure attention weights are in the correct format
|
142 |
if isinstance(attention_weights, (list, np.ndarray)):
|
143 |
attention_weights = np.array(attention_weights).flatten()
|
144 |
+
|
145 |
+
# Format weights for display
|
146 |
+
formatted_weights = [f'{float(w):.2f}' for w in attention_weights]
|
147 |
+
|
148 |
fig = go.Figure(data=[
|
149 |
go.Bar(
|
150 |
x=tokens,
|
151 |
y=attention_weights,
|
152 |
+
text=formatted_weights,
|
153 |
textposition='auto',
|
|
|
154 |
)
|
155 |
])
|
156 |
+
|
157 |
fig.update_layout(
|
158 |
+
title='Attention Weights',
|
159 |
+
xaxis_title='Tokens',
|
160 |
+
yaxis_title='Attention Weight',
|
161 |
+
xaxis_tickangle=45
|
|
|
|
|
162 |
)
|
163 |
+
|
164 |
return fig
|
165 |
|
166 |
def main():
|
167 |
+
st.title("📰 Fake News Detection System")
|
168 |
+
st.write("""
|
169 |
+
This application uses a hybrid deep learning model (BERT + BiLSTM + Attention)
|
170 |
+
to detect fake news articles. Enter a news article below to analyze it.
|
171 |
+
""")
|
172 |
+
|
173 |
+
# Sidebar
|
174 |
+
st.sidebar.title("About")
|
175 |
+
st.sidebar.info("""
|
176 |
+
|
177 |
+
The model combines:
|
178 |
+
- BERT for contextual embeddings
|
179 |
+
- BiLSTM for sequence modeling
|
180 |
+
- Attention mechanism for interpretability
|
181 |
+
""")
|
182 |
+
|
183 |
+
# Main content
|
184 |
+
st.header("News Analysis")
|
185 |
+
|
186 |
+
# Text input
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
187 |
news_text = st.text_area(
|
188 |
+
"Enter the news article to analyze:",
|
189 |
+
height=200,
|
190 |
+
placeholder="Paste your news article here..."
|
|
|
191 |
)
|
192 |
+
|
193 |
+
if st.button("Analyze"):
|
194 |
+
if news_text:
|
195 |
+
with st.spinner("Analyzing the news article..."):
|
196 |
+
# Get prediction
|
|
|
|
|
|
|
|
|
|
|
197 |
result = predict_news(news_text)
|
198 |
+
|
199 |
+
# Display result
|
200 |
+
col1, col2 = st.columns(2)
|
201 |
+
|
202 |
+
with col1:
|
203 |
+
st.subheader("Prediction")
|
204 |
+
if result['label'] == 'FAKE':
|
205 |
+
st.error(f"🔴 This news is likely FAKE (Confidence: {result['confidence']:.2%})")
|
206 |
+
else:
|
207 |
+
st.success(f"🟢 This news is likely REAL (Confidence: {result['confidence']:.2%})")
|
208 |
+
|
209 |
+
with col2:
|
210 |
+
st.subheader("Confidence Scores")
|
211 |
+
st.plotly_chart(plot_confidence(result['probabilities']), use_container_width=True)
|
212 |
+
|
213 |
+
# Show attention visualization
|
214 |
+
st.subheader("Attention Analysis")
|
215 |
+
st.write("""
|
216 |
+
The attention weights show which parts of the text the model focused on
|
217 |
+
while making its prediction. Higher weights indicate more important tokens.
|
218 |
+
""")
|
219 |
+
st.plotly_chart(plot_attention(news_text, result['attention_weights']), use_container_width=True)
|
220 |
+
|
221 |
+
# Show model explanation
|
222 |
+
st.subheader("Model Explanation")
|
223 |
+
if result['label'] == 'FAKE':
|
224 |
+
st.write("""
|
225 |
+
The model identified this as fake news based on:
|
226 |
+
- Linguistic patterns typical of fake news
|
227 |
+
- Inconsistencies in the content
|
228 |
+
- Attention weights on suspicious phrases
|
229 |
+
""")
|
230 |
+
else:
|
231 |
+
st.write("""
|
232 |
+
The model identified this as real news based on:
|
233 |
+
- Credible language patterns
|
234 |
+
- Consistent information
|
235 |
+
- Attention weights on factual statements
|
236 |
+
""")
|
237 |
else:
|
238 |
+
st.warning("Please enter a news article to analyze.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
239 |
|
240 |
if __name__ == "__main__":
|
241 |
+
main()
|