Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,19 +1,19 @@
|
|
1 |
import streamlit as st
|
2 |
import torch
|
3 |
-
from transformers import AutoTokenizer, AutoModelForSequenceClassification,pipeline
|
4 |
-
|
5 |
import requests
|
6 |
-
import
|
7 |
-
|
8 |
|
9 |
-
|
10 |
-
|
11 |
-
# Load tokenizer and model
|
12 |
tokenizer = AutoTokenizer.from_pretrained("hamzab/roberta-fake-news-classification")
|
13 |
model = AutoModelForSequenceClassification.from_pretrained("hamzab/roberta-fake-news-classification")
|
14 |
|
|
|
|
|
15 |
|
16 |
-
|
|
|
17 |
|
18 |
def translate_to_english(text):
|
19 |
try:
|
@@ -21,7 +21,12 @@ def translate_to_english(text):
|
|
21 |
except Exception as e:
|
22 |
return f"Error in translation: {e}"
|
23 |
|
24 |
-
|
|
|
|
|
|
|
|
|
|
|
25 |
|
26 |
def predict_fake(title, text):
|
27 |
input_str = "<title>" + title + "<content>" + text + "<end>"
|
@@ -34,111 +39,46 @@ def predict_fake(title, text):
|
|
34 |
|
35 |
def fact_check_with_google(api_key, query):
|
36 |
url = f"https://factchecktools.googleapis.com/v1alpha1/claims:search"
|
37 |
-
params = {
|
38 |
-
"query": query,
|
39 |
-
"key": api_key
|
40 |
-
}
|
41 |
response = requests.get(url, params=params)
|
42 |
if response.status_code == 200:
|
43 |
return response.json()
|
44 |
else:
|
45 |
-
return {"error": f"Unable to fetch
|
46 |
-
|
47 |
-
'''def main():
|
48 |
-
st.title("Fake News Prediction")
|
49 |
-
|
50 |
-
# Load Google API key from a secure location or environment variable
|
51 |
-
|
52 |
-
|
53 |
-
# Create the form for user input
|
54 |
-
with st.form("news_form"):
|
55 |
-
st.subheader("Enter News Details")
|
56 |
-
title = st.text_input("Title")
|
57 |
-
text = st.text_area("Text")
|
58 |
-
language = st.selectbox("Select Language", options=["English", "Other"])
|
59 |
-
submit_button = st.form_submit_button("Submit")
|
60 |
-
|
61 |
-
# Process form submission and make prediction
|
62 |
-
if submit_button:
|
63 |
-
if language == "Other":
|
64 |
-
title = translate_to_english(title)
|
65 |
-
text = translate_to_english(text)
|
66 |
-
|
67 |
-
prediction = predict_fake(title, text)
|
68 |
-
|
69 |
-
st.subheader("Prediction:")
|
70 |
-
st.write("Prediction: ", prediction)
|
71 |
-
|
72 |
-
if prediction.get("Real") > 0.5:
|
73 |
-
st.write("This news is predicted to be **real** :muscle:")
|
74 |
-
else:
|
75 |
-
st.write("This news is predicted to be **fake** :shit:")
|
76 |
-
|
77 |
-
'''
|
78 |
-
# Load summarizer
|
79 |
-
@st.cache_resource
|
80 |
-
def load_summarizer():
|
81 |
-
return pipeline("summarization", model="facebook/bart-large-cnn")
|
82 |
-
|
83 |
-
summarizer = load_summarizer()
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
def summarize_text(text):
|
88 |
-
try:
|
89 |
-
summary = summarizer(text, max_length=30, min_length=5, do_sample=False)
|
90 |
-
return summary[0]['summary_text']
|
91 |
-
except Exception as e:
|
92 |
-
return f"Error in summarization: {e}"
|
93 |
-
|
94 |
|
95 |
def main():
|
96 |
-
st.
|
97 |
-
|
98 |
-
# Store your API key here or load from environment variable
|
99 |
-
GOOGLE_API_KEY = "AIzaSyAf5v5380xkpo0Rk3kBiSxpxYVBQwcDi2A" # 🔐 Replace this!
|
100 |
|
101 |
with st.form("news_form"):
|
102 |
-
st.
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
submit_button = st.form_submit_button("Submit")
|
108 |
|
109 |
if submit_button:
|
110 |
if language == "Other":
|
111 |
title = translate_to_english(title)
|
112 |
text = translate_to_english(text)
|
113 |
-
|
114 |
prediction = predict_fake(title, text)
|
115 |
|
116 |
-
st.subheader("Prediction:")
|
117 |
-
st.write("Prediction:
|
118 |
-
|
119 |
-
|
120 |
-
st.write("This news is predicted to be **real** :muscle:")
|
121 |
-
else:
|
122 |
-
st.write("This news is predicted to be **fake** :shit:")
|
123 |
-
|
124 |
|
125 |
if check_fact and GOOGLE_API_KEY:
|
126 |
-
|
127 |
-
|
128 |
-
|
129 |
-
# custom_claim = st.text_input("Optional: Enter a specific claim to fact-check", "")
|
130 |
-
# query = custom_claim if custom_claim else title # Use custom claim if provided
|
131 |
-
summarized_claim = summarize_text(title)
|
132 |
-
st.info(f"🔍 Fact check query (summarized): **{summarized_claim}**")
|
133 |
-
|
134 |
-
fact_check_data = fact_check_with_google(GOOGLE_API_KEY, summarized_claim)
|
135 |
-
|
136 |
-
# Optional: show raw data for debugging
|
137 |
-
# st.json(fact_check_data)
|
138 |
-
|
139 |
|
|
|
|
|
|
|
140 |
|
141 |
-
if "claims" in fact_check_data
|
142 |
for claim in fact_check_data["claims"]:
|
143 |
st.markdown(f"**Claim:** {claim.get('text', 'N/A')}")
|
144 |
for review in claim.get("claimReview", []):
|
@@ -147,9 +87,7 @@ def main():
|
|
147 |
st.write(f"- **URL**: {review.get('url', 'N/A')}")
|
148 |
st.write("---")
|
149 |
else:
|
150 |
-
st.
|
151 |
|
152 |
-
|
153 |
if __name__ == "__main__":
|
154 |
main()
|
155 |
-
|
|
|
1 |
import streamlit as st
|
2 |
import torch
|
3 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
|
4 |
+
from deep_translator import GoogleTranslator
|
5 |
import requests
|
6 |
+
import os
|
|
|
7 |
|
8 |
+
# Load tokenizer and fake news model
|
|
|
|
|
9 |
tokenizer = AutoTokenizer.from_pretrained("hamzab/roberta-fake-news-classification")
|
10 |
model = AutoModelForSequenceClassification.from_pretrained("hamzab/roberta-fake-news-classification")
|
11 |
|
12 |
+
# Load summarizer pipeline
|
13 |
+
summarizer = pipeline("summarization", model="sshleifer/distilbart-cnn-12-6")
|
14 |
|
15 |
+
# Google Fact Check API key (replace with your actual key or use st.secrets)
|
16 |
+
GOOGLE_API_KEY = "AIzaSyAf5v5380xkpo0Rk3kBiSxpxYVBQwcDi2A"
|
17 |
|
18 |
def translate_to_english(text):
|
19 |
try:
|
|
|
21 |
except Exception as e:
|
22 |
return f"Error in translation: {e}"
|
23 |
|
24 |
+
def summarize_text(text):
|
25 |
+
try:
|
26 |
+
summary = summarizer(text, max_length=60, min_length=15, do_sample=False)
|
27 |
+
return summary[0]['summary_text']
|
28 |
+
except Exception as e:
|
29 |
+
return f"Error in summarization: {e}"
|
30 |
|
31 |
def predict_fake(title, text):
|
32 |
input_str = "<title>" + title + "<content>" + text + "<end>"
|
|
|
39 |
|
40 |
def fact_check_with_google(api_key, query):
|
41 |
url = f"https://factchecktools.googleapis.com/v1alpha1/claims:search"
|
42 |
+
params = {"query": query, "key": api_key}
|
|
|
|
|
|
|
43 |
response = requests.get(url, params=params)
|
44 |
if response.status_code == 200:
|
45 |
return response.json()
|
46 |
else:
|
47 |
+
return {"error": f"Unable to fetch fact-checks. HTTP {response.status_code}: {response.text}"}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
48 |
|
49 |
def main():
|
50 |
+
st.set_page_config(page_title="News Credibility Checker", page_icon="🧠")
|
51 |
+
st.title("🧠 Fake News Detection & Fact Check")
|
|
|
|
|
52 |
|
53 |
with st.form("news_form"):
|
54 |
+
title = st.text_input("News Title")
|
55 |
+
text = st.text_area("News Content")
|
56 |
+
language = st.selectbox("Select Language", ["English", "Other"])
|
57 |
+
check_fact = st.checkbox("Check with Google Fact Check API")
|
58 |
+
submit_button = st.form_submit_button("Analyze")
|
|
|
59 |
|
60 |
if submit_button:
|
61 |
if language == "Other":
|
62 |
title = translate_to_english(title)
|
63 |
text = translate_to_english(text)
|
64 |
+
|
65 |
prediction = predict_fake(title, text)
|
66 |
|
67 |
+
st.subheader("Prediction Result:")
|
68 |
+
st.write("Prediction Score:", prediction)
|
69 |
+
verdict = "real" if prediction.get("Real") > 0.5 else "fake"
|
70 |
+
st.success(f"This news is predicted to be **{verdict}**.")
|
|
|
|
|
|
|
|
|
71 |
|
72 |
if check_fact and GOOGLE_API_KEY:
|
73 |
+
# Generate summary for fact-checking
|
74 |
+
summary_text = summarize_text(title + ". " + text)
|
75 |
+
st.markdown("**Fact-check Query (Summary):** " + summary_text)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
76 |
|
77 |
+
# Call Google Fact Check API
|
78 |
+
fact_check_data = fact_check_with_google(GOOGLE_API_KEY, summary_text)
|
79 |
+
st.subheader("Google Fact Check Results")
|
80 |
|
81 |
+
if "claims" in fact_check_data:
|
82 |
for claim in fact_check_data["claims"]:
|
83 |
st.markdown(f"**Claim:** {claim.get('text', 'N/A')}")
|
84 |
for review in claim.get("claimReview", []):
|
|
|
87 |
st.write(f"- **URL**: {review.get('url', 'N/A')}")
|
88 |
st.write("---")
|
89 |
else:
|
90 |
+
st.info("No fact-check results found.")
|
91 |
|
|
|
92 |
if __name__ == "__main__":
|
93 |
main()
|
|