Upload 2 files

app.py

@@ -0,0 +1,80 @@
+import feedparser
+import snscrape.modules.twitter as sntwitter
+from transformers import pipeline
+import gradio as gr
+from collections import Counter
+
+# Load sentiment pipeline once
+sentiment_model = pipeline("sentiment-analysis")
+
+def fetch_google_news(max_items=10):
+    url = "https://news.google.com/rss?hl=en-US&gl=US&ceid=US:en"
+    feed = feedparser.parse(url)
+    news = []
+    for entry in feed.entries[:max_items]:
+        title = entry.title
+        # Sometimes source is in 'source' key, else fallback to 'Unknown Source'
+        source = entry.get('source', {}).get('title', 'Unknown Source')
+        news.append({'text': title, 'source': source})
+    return news
+
+def fetch_twitter(max_items=10):
+    tweets = []
+    try:
+        scraper = sntwitter.TwitterHashtagScraper('mentalhealth')
+        for i, tweet in enumerate(scraper.get_items()):
+            if i >= max_items:
+                break
+            tweets.append({'text': tweet.content, 'source': 'Twitter'})
+    except Exception as e:
+        # Return empty list if scraping fails
+        print(f"Error fetching tweets: {e}")
+    return tweets
+
+def analyze_headlines():
+    news = fetch_google_news()
+    tweets = fetch_twitter()
+
+    combined = news + tweets
+    texts = [item['text'] for item in combined]
+
+    # Run sentiment analysis (batch)
+    results = sentiment_model(texts)
+
+    # Attach sentiment back
+    for item, res in zip(combined, results):
+        item['sentiment'] = res['label']
+        item['score'] = res['score']
+
+    pos_count = sum(1 for x in combined if x['sentiment'] == 'POSITIVE')
+    neg_count = sum(1 for x in combined if x['sentiment'] == 'NEGATIVE')
+    total = pos_count + neg_count
+    risk_ratio = neg_count / total if total > 0 else 0
+
+    # Count negative news by source
+    neg_sources = [x['source'] for x in combined if x['sentiment'] == 'NEGATIVE']
+    source_counts = Counter(neg_sources)
+
+    # Prepare output
+    lines = [
+        f"Positive News Count: {pos_count}",
+        f"Negative News Count: {neg_count}",
+        f"Risk Ratio (Negative / Total): {risk_ratio:.2f}",
+        "",
+        "Negative News by Source / Region:"
+    ]
+    for source, count in source_counts.most_common():
+        lines.append(f" - {source}: {count}")
+
+    return "\n".join(lines)
+
+app = gr.Interface(
+    fn=analyze_headlines,
+    inputs=[],
+    outputs="textbox",
+    title="🧠 Live Headline Sentiment Risk & Region Analyzer",
+    description="Fetches latest news and tweets, analyzes sentiment, counts risk, and shows areas with negative news."
+)
+
+if __name__ == "__main__":
+    app.launch()

requirements.txt

@@ -0,0 +1,5 @@
+gradio
+feedparser
+transformers
+snscrape
+torch