added rule based feedback system

app.py

@@ -172,22 +172,62 @@ def create_rag_pipeline(file_paths, model, temperature, max_tokens):
         logger.debug("Exception details:", exc_info=True)
         return None, f"Error creating RAG pipeline: {e}"
 
-# Function to handle feedback (Optional Enhancement)
+# Define positive and negative words for rule-based sentiment analysis
+POSITIVE_WORDS = {
+    "good", "great", "excellent", "amazing", "wonderful", "fantastic", "positive",
+    "helpful", "satisfied", "happy", "love", "liked", "enjoyed", "beneficial",
+    "superb", "awesome", "nice", "brilliant", "favorable", "pleased"
+}
+
+NEGATIVE_WORDS = {
+    "bad", "terrible", "awful", "poor", "disappointed", "unsatisfied", "hate",
+    "hated", "dislike", "dislikes", "worst", "negative", "not helpful", "frustrated",
+    "unhappy", "dissatisfied", "unfortunate", "horrible", "annoyed", "problem", "issues"
+}
+
+# Function to handle feedback with rule-based sentiment analysis
 def handle_feedback(feedback_text):
     """
-    Handles user feedback by logging it.
-    In a production environment, consider storing feedback in a database or external service.
+    Handles user feedback by analyzing its sentiment and providing a dynamic response.
+    Stores the feedback in a temporary file for persistence during the session.
 
     Parameters:
     - feedback_text (str): The feedback provided by the user.
 
     Returns:
-    - str: Acknowledgment message.
+    - str: Acknowledgment message based on feedback sentiment.
     """
     if feedback_text and feedback_text.strip() != "":
-        # For demonstration, we'll log the feedback. Replace this with database storage if needed.
-        logger.info(f"User Feedback: {feedback_text}")
-        return "Thank you for your feedback!"
+        # Normalize feedback text to lowercase for comparison
+        feedback_lower = feedback_text.lower()
+
+        # Count positive and negative words
+        positive_count = sum(word in feedback_lower for word in POSITIVE_WORDS)
+        negative_count = sum(word in feedback_lower for word in NEGATIVE_WORDS)
+
+        # Determine sentiment based on counts
+        if positive_count > negative_count:
+            sentiment = "positive"
+            acknowledgment = "Thank you for your positive feedback! We're glad to hear that you found our service helpful."
+        elif negative_count > positive_count:
+            sentiment = "negative"
+            acknowledgment = "We're sorry to hear that you're not satisfied. Your feedback is valuable to us, and we'll strive to improve."
+        else:
+            sentiment = "neutral"
+            acknowledgment = "Thank you for your feedback. We appreciate your input."
+
+        # Log the feedback with sentiment
+        logger.info(f"User Feedback: {feedback_text} | Sentiment: {sentiment}")
+
+        # Optionally, store feedback in a temporary file
+        try:
+            with open("/tmp/user_feedback.txt", "a") as f:
+                f.write(f"{feedback_text} | Sentiment: {sentiment}\n")
+            logger.debug("Feedback stored successfully in /tmp/user_feedback.txt.")
+        except Exception as e:
+            logger.error(f"Error storing feedback: {e}")
+
+        return acknowledgment
     else:
         return "No feedback provided."
 
@@ -210,8 +250,10 @@ def answer_question(model, temperature, max_tokens, question, feedback):
         logger.info("Received invalid input from user.")
         return "Please provide a valid question or input containing meaningful text.", ""
 
-    # Check if the RAG pipeline needs to be re-initialized (e.g., if model or parameters have changed)
-    # For simplicity, we'll assume the pipeline remains the same. For dynamic models, implement re-initialization here.
+    # Check if the RAG pipeline is initialized
+    if rag_chain is None:
+        logger.error("RAG pipeline is not initialized.")
+        return "The system is currently unavailable. Please try again later.", ""
 
     try:
         answer = rag_chain.run(question)