Update app.py

app.py

@@ -3,6 +3,7 @@ from pydantic import BaseModel
 from sentence_transformers import SentenceTransformer
 import faiss
 import pandas as pd
+import random
 from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
 
 app = FastAPI()
@@ -36,11 +37,31 @@ class SummaryRequest(BaseModel):
 
 @app.post("/get_questions")
 def get_recommended_questions(request: ChatRequest):
-    """Retrieve the most relevant diagnostic questions."""
+    """Retrieve the most relevant diagnostic questions with a conversational response."""
+    
+    # Step 1: Encode the input message for FAISS search
     input_embedding = embedding_model.encode([request.message], convert_to_numpy=True)
     distances, indices = question_index.search(input_embedding, 3)
+
+    # Step 2: Retrieve the top 3 relevant questions
     retrieved_questions = [questions_df["Questions"].iloc[i] for i in indices[0]]
-    return {"questions": retrieved_questions}
+
+    # Step 3: Define Dynamic Prompt Variations
+    empathetic_phrases = [
+        "I hear you, and I appreciate you sharing this. Let me ask:",
+        "That sounds challenging. Could you help me understand better by answering this:",
+        "I understand what you're going through. Here's something to think about:",
+        "Thank you for opening up. Let’s explore this further:",
+        "Your feelings are completely valid. Here's a question to help us understand more:"
+    ]
+
+    # Step 4: Generate Dynamic Responses
+    wrapped_responses = [
+        f"{random.choice(empathetic_phrases)} *{q}*"
+        for q in retrieved_questions
+    ]
+
+    return {"questions": wrapped_responses}
 
 @app.post("/summarize_chat")
 def summarize_chat(request: SummaryRequest):