BE-18: Feedback Panel Research (Rating, Correctness, Length)

[tecnodeveloper]

Description:
Research how to build a feedback system for chatbot responses. Users will rate each bot reply on quality, correctness, and length. This data will be stored in MongoDB and used for analytics and model improvement.

---

User Story

Given user receives a chatbot response
When user interacts with feedback panel
Then system should capture and store structured feedback

---

Tasks

---

Understand Feedback System

1. Define Purpose of Feedback

   • Improve chatbot accuracy
   • Track response quality
   • Collect user satisfaction

2. Identify Feedback Types

   • Rating (1–5 stars)
   • Correctness (Yes / No / Partial)
   • Length (Too short / Good / Too long)

---

UI/UX Research

3. Study Feedback UI Patterns

   • Star rating systems
   • Thumbs up/down systems
   • Dropdown selection systems

4. Decide UI Style

   • Simple panel below message
   • Inline feedback buttons
   • Expandable feedback section

---

Frontend Design

5. Create Feedback Panel Component

   • Rating selector
   • Correctness selector
   • Length selector

6. Attach to Each Message

   • Only bot messages
   • Unique message_id tracking

---

Backend Design

7. Create Feedback API

   • POST /feedback/submit
   • Accept message_id + rating data

8. Validate Input

   • Rating range check (1–5)
   • Required fields validation

---

MongoDB Storage

9. Create Feedback Collection

   • message_id
   • user_id
   • rating
   • correctness
   • length_feedback
   • timestamp

10. Ensure Linking

• Connect feedback → message → session

---

Data Flow Design

11. Frontend Flow

• User clicks feedback
• Send to backend API
• Show confirmation

12. Backend Flow

• Receive feedback
• Store in MongoDB
• Link to message

---

Feedback Analytics Preparation

13. Define Metrics

• Average rating per session
• Accuracy score
• Response quality trends

14. Prepare for Dashboard

• Store structured data
• Enable aggregation queries

---

Postman Testing 🧪

15. Setup Postman

• Create POST request /feedback/submit

16. Test Payload

{
  "message_id": "abc123",
  "user_id": "user1",
  "rating": 4,
  "correctness": "yes",
  "length": "good"
}

17. Validate Responses

• Success response
• Error handling
• Missing fields

---

Edge Cases

18. Handle Invalid Data

• Rating out of range
• Missing message_id
• Duplicate feedback

19. Prevent Spam

• One feedback per message
• User restriction

---

Performance Considerations

20. Efficient Storage

• Lightweight documents
• Indexed message_id

21. Scalability

• High feedback volume handling
• Fast aggregation queries

---

Integration with Chat UI

22. Attach to Chat Messages

• Show after bot response
• Enable quick rating

23. Update UI State

• Show submitted status
• Disable duplicate feedback

---

Acceptance Criteria

• Feedback system defined clearly
• UI structure designed
• Backend API planned
• MongoDB schema finalized
• Postman testing completed
• Linked to chat messages

---

Testing Steps

1. Simulate chat response
2. Submit feedback
3. Check MongoDB storage
4. Validate linking with message
5. Test duplicate prevention

---

Definition of Done

• Feedback system fully designed
• Data structure finalized
• API ready for implementation
• Connected to chatbot flow