User Guide
A complete guide to using the TFG Pedagogical Chatbot web interface.
Getting Started
Accessing the Application
Open http://localhost:3000 in your browser.
User Roles
| Role | Capabilities |
|---|---|
| Student | Chat with assistant, take tests, view own progress, configure test preferences |
| Professor | Dashboard access, manage students, upload documents, view analytics, configure test defaults for students |
| Admin | All features + user management, subject management, system configuration |
Note: The chat interface is exclusively for students. Professors access the system via their dashboard.
Authentication
Login
- Navigate to http://localhost:3000/login
- Enter your username and password
- Click Iniciar sesión
Register (if enabled)
- Click Registrarse on the login page
- Fill in:
- Email address
- Username
- Password (min 6 characters)
- Click Crear cuenta
Demo Accounts
For development, use the seeded accounts:
| Username | Password | Role |
|---|---|---|
admin | admin123 | Admin |
profesor | admin123 | Professor |
estudiante | admin123 | Student |
Chat Interface
The main chat interface is where students interact with the pedagogical assistant.
Creating a Conversation
- Click Nueva conversación button
- Enter a title (e.g., “Dudas sobre Docker”)
- Select a subject (e.g., “ISE”)
- Click Crear
Chatting
- Type your message in the input field at the bottom
- Press Enter or click the send button
- Wait for the assistant’s response
Features
- Markdown Support: Responses include formatted code, lists, and links
- Code Highlighting: Code blocks are syntax-highlighted
- Message History: Scroll up to see previous messages
- Persistent Sessions: Conversations are saved automatically
Example Interactions
Ask a question:
¿Qué es Docker y para qué se usa?
Request code examples:
Muéstrame un ejemplo de docker-compose.yml
Ask for explanations:
Explícame la diferencia entre imágenes y contenedores
Taking Tests
The chatbot can generate interactive tests to assess your knowledge.
Starting a Test
Type a request like:
Genera un test de 5 preguntas sobre contenedores Docker
During the Test
- The assistant presents questions one at a time
- Type your answer in the chat input
- Receive immediate feedback
- Continue until all questions are answered
Test Results
After completing a test, you’ll see:
- Score: Correct answers / Total questions
- Feedback: Detailed explanation for each answer
- Areas to improve: Suggestions for further study
Session Management
Switching Conversations
- Click the session selector dropdown in the header
- Select a previous conversation
- Messages load automatically
Deleting a Conversation
- Hover over a conversation in the selector
- Click the trash icon
- Confirm deletion
Professor Dashboard
Professors have access to additional features via the Dashboard.
Accessing the Dashboard
Click Dashboard in the navigation menu (visible only to professors/admins).
Overview
The dashboard shows:
- Your subjects: Cards for each subject you teach
- Quick stats: Total students, conversations, documents
Managing Students
- Click Ver estudiantes on a subject card
- View enrolled students and their activity
- Click Matricular estudiante to add new students
Viewing Student Progress
- Click Ver progreso on a subject card
- See:
- Conversation history
- Test scores over time
- Question difficulty distribution
- Areas where students struggle
Document Management
Upload teaching materials for RAG:
- Click Gestionar documentos on a subject card
- Click Subir documento
- Select a PDF, TXT, or Markdown file
- The document is indexed for semantic search
Supported formats: PDF, TXT, MD, DOCX
Admin Panel
Administrators have full system control.
Accessing Admin Panel
Click Administración in the navigation menu (visible only to admins).
User Management
View all users:
- See username, email, role, and subjects
- Filter by role
Promote a user:
- Click the shield icon next to a user
- Select new role (Student → Professor → Admin)
- Confirm promotion
Assign subjects:
- Click the book icon next to a user
- Select subjects to assign
- Save changes
Subject Management
Create a new subject:
- Click Crear asignatura
- Enter subject code (e.g., “ISE”)
- Enter subject name (e.g., “Infraestructura de Servidores”)
- Click Crear
View subject statistics:
- Total enrolled students
- Total conversations
- Document count
Settings
Access personal settings via Configuración in the menu.
Profile
View your account information:
- Username
- Role
- Assigned subjects
Test Preferences
Configure default settings for generated tests:
- Default number of questions: 1-20 questions per test
- Default difficulty: Easy, Medium, or Hard
- Language: Spanish or English for test questions
These preferences are used as defaults when generating tests. You can always override them by specifying parameters in your test request.
System Information
View the current LLM model being used by the chatbot (e.g., gemini-2.0-flash-lite, Qwen/Qwen2.5-7B-Instruct).
Note: Only students can access the chat feature. Professors should use their dashboard to view student progress and manage documents.
Keyboard Shortcuts
| Shortcut | Action |
|---|---|
Enter | Send message |
Shift + Enter | New line in message |
Esc | Close dialogs |
Tips for Effective Use
For Students
- Be specific: “¿Cómo funciona Docker networking?” vs “¿Qué es Docker?”
- Ask follow-ups: Build on previous answers
- Request examples: Ask for code samples and practical use cases
- Take tests: Regular self-assessment helps retention
For Professors
- Upload relevant materials: PDFs of lecture notes, assignments
- Monitor progress: Check dashboard weekly
- Review questions: See what students struggle with
For Admins
- Regular backups: Export MongoDB data
- Monitor logs: Check Grafana for errors
- User audits: Review user roles periodically
Troubleshooting
Can’t log in
- Check username/password are correct
- Clear browser cookies
- Try incognito mode
Chat not responding
- Check if services are running:
docker compose ps - Look for errors in browser console (F12)
- Check chatbot logs:
docker logs tfg-chatbot
Documents not appearing in search
- Wait 30 seconds for indexing
- Check RAG service logs:
docker logs tfg-rag-service - Verify Ollama is running:
docker logs ollama-service
Slow responses
- First message may be slow (model loading)
- Check Phoenix for LLM latency
- Consider using Gemini instead of local vLLM
Mobile Usage
The interface is responsive and works on mobile devices:
- Portrait mode: Sidebar collapses, tap to expand
- Landscape mode: Full interface visible
- Touch: Tap to select, swipe for navigation
Browser Support
| Browser | Support |
|---|---|
| Chrome (latest) | ✅ Full |
| Firefox (latest) | ✅ Full |
| Safari (latest) | ✅ Full |
| Edge (latest) | ✅ Full |
| IE 11 | ❌ Not supported |