Update app.py

app.py

@@ -2,25 +2,23 @@ import os
 import numpy as np
 import pandas as pd
 import streamlit as st
-from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification, AutoModelForCausalLM
+from huggingface_hub import login
+from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification
 from keybert import KeyBERT
+from openai import AzureOpenAI  # new
 
-# ─── DeepSeek Model Client ────────────────────────────────────────────────────
-# High-level helper pipeline for text-generation
-pipe = pipeline(
-    "text-generation",
-    model="deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",
-    trust_remote_code=True
+# ─── Azure OpenAI Client ─────────────────────────────────────────────────────
+openai_client = AzureOpenAI(
+  api_key = "fbca46bfd8814334be46a2e5c323904c", # use your key here
+  api_version = "2023-05-15", # apparently HKUST uses a deprecated version
+  azure_endpoint = "https://hkust.azure-api.net" # per HKUST instructions
 )
 
 @st.cache_resource
 def load_sentiment_pipeline():
     model_name = "mayf/amazon_reviews_bert_ft"
     tok = AutoTokenizer.from_pretrained(model_name, use_auth_token=True)
-    mdl = AutoModelForSequenceClassification.from_pretrained(
-        model_name,
-        use_auth_token=True
-    )
+    mdl = AutoModelForSequenceClassification.from_pretrained(model_name, use_auth_token=True)
     return pipeline(
         "sentiment-analysis",
         model=mdl,
@@ -51,6 +49,7 @@ def main():
         st.warning("Please enter a review to analyze.")
         return
 
+    # Initialize progress bar
     progress = st.progress(0)
 
     # Load models
@@ -59,15 +58,14 @@ def main():
     kw_model = load_keybert_model()
     progress.progress(20)
 
-    # Sentiment analysis
+    # Run 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
-    }
+    # Map labels
+    sentiment_results = {LABEL_MAP[item['label']]: float(item['score']) for item in raw_scores}
     progress.progress(40)
 
-    # Keyword extraction
+    # Extract keywords
     progress.text("Extracting keywords...")
     keywords = kw_model.extract_keywords(
         review,
@@ -77,34 +75,29 @@ def main():
     )
     progress.progress(60)
 
-    # Display results
+    # Display scores and keywords side by side
     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 Keywords")
+        st.subheader("Top 3 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}\"
@@ -114,24 +107,23 @@ 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).
 """
-    # 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']
 
+    response = openai_client.chat.completions.create(
+        model="gpt-35-turbo",
+        messages=[
+            {"role": "system", "content": "You are a product-feedback analyst."},
+            {"role": "user", "content": prompt}
+        ],
+        temperature=0.7,
+        max_tokens=200
+    )
+    gpt_reply = response.choices[0].message.content.strip()
     st.markdown(gpt_reply)
+
+    # Complete
     progress.progress(100)
     progress.text("Done!")
 
 if __name__ == "__main__":
     main()
 
-