Update app.py

app.py

@@ -6,12 +6,13 @@ from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassifica
 from keybert import KeyBERT
 
 # ─── DeepSeek Model Client ────────────────────────────────────────────────────
-# High-level helper pipeline for text-generation (flattened chat messages)
+# High-level helper pipeline for text-generation
 pipe = pipeline(
     "text-generation",
     model="deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",
     trust_remote_code=True
 )
+
 @st.cache_resource
 def load_sentiment_pipeline():
     model_name = "mayf/amazon_reviews_bert_ft"
@@ -39,6 +40,7 @@ LABEL_MAP = {
     "LABEL_4": "Very Positive"
 }
 
+
 def main():
     st.title("📊 Amazon Review Analyzer")
 
@@ -49,7 +51,6 @@ def main():
         st.warning("Please enter a review to analyze.")
         return
 
-    # Initialize progress bar
     progress = st.progress(0)
 
     # Load models
@@ -58,13 +59,15 @@ def main():
     kw_model = load_keybert_model()
     progress.progress(20)
 
-    # Run sentiment analysis
+    # Sentiment analysis
     progress.text("Analyzing sentiment...")
     raw_scores = sentiment_pipeline(review)[0]
-    sentiment_results = {LABEL_MAP[item['label']]: float(item['score']) for item in raw_scores}
+    sentiment_results = {
+        LABEL_MAP[item['label']]: float(item['score']) for item in raw_scores
+    }
     progress.progress(40)
 
-    # Extract keywords
+    # Keyword extraction
     progress.text("Extracting keywords...")
     keywords = kw_model.extract_keywords(
         review,
@@ -74,54 +77,56 @@ def main():
     )
     progress.progress(60)
 
-    # Display scores and keywords side by side
+    # Display results
     col1, col2 = st.columns(2)
     with col1:
         st.subheader("Sentiment Scores")
         st.json({k: round(v, 4) for k, v in sentiment_results.items()})
     with col2:
-        st.subheader("Top 3 Keywords")
+        st.subheader("Top Keywords")
         for kw, score in keywords:
             st.write(f"• {kw} ({score:.4f})")
 
     # Bar chart
     progress.text("Rendering chart...")
-    df_scores = pd.DataFrame.from_dict(sentiment_results, orient='index', columns=['score'])
+    df_scores = pd.DataFrame.from_dict(
+        sentiment_results, orient='index', columns=['score']
+    )
     df_scores.index.name = 'Sentiment'
     st.bar_chart(df_scores)
     progress.progress(80)
 
     # Highlight highest sentiment
-    max_label, max_score = max(sentiment_results.items(), key=lambda x: x[1])
+    max_label, max_score = max(
+        sentiment_results.items(), key=lambda x: x[1]
+    )
     st.markdown(f"**Highest Sentiment:** **{max_label}** ({max_score:.4f})")
 
     # GPT-Driven Analysis & Suggestions
     progress.text("Generating insights...")
+    # Build the prompt
     prompt = f"""
 You are an analytical amazon feedback expert.
-Review: "{review}"
+Review: \"{review}\"
 Sentiment Scores: {sentiment_results}
 Top Keywords: {[kw for kw, _ in keywords]}
 Tasks:
 1. Analysis: Write a concise paragraph (3 sentences) interpreting customer sentiment by combining the scores and keywords.
 2. Recommendations: Three separate paragraphs with actionable suggestions (max 30 words each).
 """
-
-    # Use the high-level text-generation pipeline with flattened prompt
-    # Flatten roles into a single text prompt
-    flat_prompt = "
-".join(f"{msg['role'].upper()}: {msg['content']}" for msg in chat_input)
+    # Prepare chat messages
+    chat_input = [
+        {"role": "system", "content": "You are a product-feedback analyst."},
+        {"role": "user", "content": prompt}
+    ]
+    # Flatten into a single text prompt
+    flat_prompt = "\n".join(
+        f"{msg['role'].upper()}: {msg['content']}" for msg in chat_input
+    )
+    # Generate
     gen_output = pipe(flat_prompt, max_new_tokens=200)
     gpt_reply = gen_output[0]['generated_text']
 
-    # Alternative: direct model invocation
-    # inputs = tokenizer_ds(flat_prompt, return_tensors="pt")
-    # outputs = model_ds.generate(**inputs, max_new_tokens=200)
-    # gpt_reply = tokenizer_ds.decode(outputs[0], skip_special_tokens=True)
-    # inputs = tokenizer_ds(prompt, return_tensors="pt")
-    # outputs = model_ds.generate(**inputs, max_new_tokens=200)
-    # gpt_reply = tokenizer_ds.decode(outputs[0], skip_special_tokens=True)
-
     st.markdown(gpt_reply)
     progress.progress(100)
     progress.text("Done!")
@@ -129,3 +134,4 @@ Tasks:
 if __name__ == "__main__":
     main()
 
+