Create app.py

app.py

@@ -0,0 +1,39 @@
+# lab1_sentiment_local.py
+
+from transformers import pipeline
+
+# Load sentiment pipeline locally (downloads model once)
+MODEL = "cardiffnlp/twitter-roberta-base-sentiment-latest"
+classifier = pipeline("sentiment-analysis", model=MODEL)
+
+# Our test reviews (in production, load from CSV or database)
+reviews = [
+    {"id": 1, "product": "Wireless Headphones",
+     "text": "Incredible sound quality and the noise cancellation is superb."},
+    {"id": 2, "product": "Wireless Headphones",
+     "text": "Broke after 2 weeks. Cheap plastic. Do not buy."},
+    {"id": 3, "product": "Standing Desk",
+     "text": "Easy to assemble. Sturdy build. Motor is a bit noisy."},
+    {"id": 4, "product": "Standing Desk",
+     "text": "Arrived with a huge scratch on the surface. Disappointed."},
+    {"id": 5, "product": "Coffee Maker",
+     "text": "Makes decent coffee. Nothing special for the price."},
+    {"id": 6, "product": "Coffee Maker",
+     "text": "Best purchase this year! Perfect espresso every morning."},
+    {"id": 7, "product": "Laptop Stand",
+     "text": "It does the job. Average quality."},
+    {"id": 8, "product": "Laptop Stand",
+     "text": "Wobbly and unstable. Returned it immediately."},
+]
+
+print(f"{'ID':<4} {'Product':<20} {'Sentiment':<10} {'Conf':>6}  Text")
+print("-" * 90)
+
+for review in reviews:
+    result = classifier(review["text"])[0]
+
+    print(f"{review['id']:<4} "
+          f"{review['product']:<20} "
+          f"{result['label']:<10} "
+          f"{result['score']:>5.1%}  "
+          f"{review['text'][:40]}...")