Skip to content

nat50/SkyLearnAI

BranchesTags
 
 

Repository files navigation

SkyLearn AI — AI-Powered Learning Management System

Forked from SkyCascade/SkyLearn and extended with AI-driven lesson generation, RAG-based document retrieval, automated quiz creation, AI chatbot tutoring, and Yolo:Farm AIoT education features.

Group 23

screenshot

Table of Contents

Overview

SkyLearn AI is a full-featured Learning Management System built with Django, enhanced with AI capabilities powered by Google Gemini. It provides traditional LMS functions (course management, grading, quizzes, student/lecturer management) alongside modern AI-powered tools for automated lesson generation, intelligent quiz creation, and an AI chatbot tutor.

The project is designed as a university-level LMS with support for multiple user roles (Admin, Lecturer, Student, Parent, Department Head), course enrollment, grade tracking, payment integration (Stripe), and internationalization (English, French, Spanish, Russian).

What Changed from the Original SkyLearn

This fork introduces the following major additions and modifications on top of SkyCascade/SkyLearn:

New: ai_core Module

An entirely new Django app (ai_core/) containing:

  • Gemini LLM integration via LangChain (langchain-google-genai)
  • AI Lesson Generator — generates structured HTML lessons from any topic
  • AI Quiz Generator — auto-creates multiple-choice questions from lesson content
  • RAG (Retrieval-Augmented Generation) pipeline — parses uploaded documents (PDF, DOCX, PPTX), chunks them, embeds with Gemini Embedding API, and performs cosine-similarity semantic search to enrich lesson generation
  • AI Chatbot — a floating conversational assistant (SkyLearn AI Assistant) available on every page
  • Yolo:Farm AIoT Lesson Service — specialized lesson pipeline for IoT education with Yolo:Bit microcontroller, generating real MicroPython code that can be flashed to hardware via Web Serial API

New: Frontend Components

  • Chatbot widget (templates/snippets/chatbot_widget.html, static/css/chatbot.css, static/js/chatbot.js)
  • Yolo:Farm UI (static/css/yolo_farm.css, static/js/yolo_farm.js) — interactive lesson viewer with code blocks, serial monitor, and hardware upload functionality

Infrastructure Changes

  • Database: Switched from SQLite to PostgreSQL (Neon serverless)
  • Media storage: Added Cloudinary integration for cloud file storage
  • Dependencies: Added langchain, langchain-google-genai, PyPDF2, python-docx, python-pptx, htmldocx, numpy, psycopg2-binary, dj-database-url, cloudinary, django-cloudinary-storage
  • Font: Added Google Inter font with Vietnamese support

Modified Files

  • config/settings.py — added ai_core to installed apps, PostgreSQL database config, Cloudinary storage config
  • config/urls.py — added ai/ URL prefix for AI endpoints
  • templates/base.html — included chatbot widget, Yolo:Farm CSS/JS, Inter font
  • course/models.py — added html_content field to Upload model for storing generated lessons
  • requirements/base.txt — added AI and cloud dependencies

Features

LMS Core (from upstream SkyLearn)

  • Dashboard — School demographics and analytics (admin only)
  • News & Events — Accessible to all users
  • User Management — Admin manages students, lecturers, parents, department heads
  • Course Management — Add/drop courses, course allocation to lecturers
  • Grading System — Attendance, mid exam, final exam, assignment, quiz scores; auto-calculated total, grade (A+ to F), GPA, and CGPA
  • Grade Comments — Automatic pass/fail classification
  • Assessment & Grade Results — Dedicated pages for students
  • Session/Semester Management — Grouped assessments and grades by semester
  • Course Materials — Upload documents (PDF, DOCX, PPTX, XLS, ZIP, etc.) and videos per course
  • PDF Generator — Registration slips and grade result reports
  • Search — Full-text search across users, courses, and programs
  • Payment Integration — Stripe-based invoicing
  • Quiz System — Multiple choice, True/False, question categories, progress tracking, pass marks, score history, randomized question order
  • Internationalization — English, French, Spanish, Russian (via django-modeltranslation)
  • Page Access Restriction — Role-based permissions

AI-Powered Features (new in this fork)

  • 🤖 AI Lesson Generation — Generate complete HTML lessons from a topic using Gemini LLM
  • 📚 RAG-Enhanced Lessons — Select uploaded course documents as context; the system parses, chunks, embeds, and semantically retrieves relevant content to produce more accurate lessons
  • 📝 AI Quiz Generation — Auto-create multiple-choice quizzes from any lesson content with configurable difficulty (easy/medium/hard) and question count
  • 💬 AI Chatbot Tutor — Floating chat widget on every page; supports multi-turn conversation, answers academic questions, provides study tips
  • 🌾 Yolo:Farm AIoT Lessons — Specialized lesson pipeline for IoT education with the Yolo:Bit microcontroller (OhStem platform); generates real MicroPython code blocks
  • 🔌 Web Serial API — Upload generated MicroPython code directly to physical Yolo:Bit hardware from the browser
  • 📄 Lesson to DOCX Export — Save AI-generated lessons as Word documents in course materials

Tech Stack

Layer Technology
Backend Django 4.2 (LTS), Python 3.8+
AI / LLM Google Gemini (via LangChain), Gemini Embedding API
RAG Custom pipeline: PyPDF2, python-docx, python-pptx → chunking → Gemini embeddings → cosine similarity (NumPy)
Database PostgreSQL (Neon serverless)
Media Storage Cloudinary
Frontend Bootstrap 5.3, jQuery 3.7, FontAwesome 6.5, SCSS
Payments Stripe
PDF Generation xhtml2pdf, ReportLab
DOCX Generation python-docx, htmldocx
Admin Panel Django JET (reboot)
i18n django-modeltranslation (EN, FR, ES, RU)
Static Files WhiteNoise

Project Structure

SkyLearnAI/
├── ai_core/                    # 🆕 AI module (lesson, quiz, chatbot, RAG)
│   ├── models.py               #    AIGeneration, DocumentChunk models
│   ├── views.py                #    API endpoints for AI features
│   ├── urls.py                 #    /ai/ URL routes
│   └── services/
│       ├── base.py             #    Abstract LLM service (LangChain)
│       ├── gemini.py           #    Google Gemini provider
│       ├── lesson.py           #    General lesson generation
│       ├── yolo_farm_lesson.py #    Yolo:Farm IoT lesson generation
│       ├── quiz.py             #    Quiz generation with JSON parsing
│       ├── chatbot.py          #    Conversational AI assistant
│       ├── rag.py              #    Embedding, indexing, semantic search
│       └── document_parser.py  #    PDF/DOCX/PPTX text extraction + chunking
│
├── accounts/                   # User management (Student, Lecturer, Parent, etc.)
├── config/                     # Django settings, root URLs, WSGI/ASGI
├── core/                       # News & Events, Session, Semester, ActivityLog
├── course/                     # Programs, Courses, Uploads, Videos
├── quiz/                       # Quiz system (MCQ, True/False, categories)
├── result/                     # Grading: TakenCourse, Result (GPA/CGPA)
├── search/                     # Full-text search
├── payments/                   # Stripe payment & invoicing
│
├── templates/                  # Django HTML templates
│   └── snippets/
│       └── chatbot_widget.html # 🆕 Floating chatbot UI
├── static/
│   ├── css/
│   │   ├── chatbot.css         # 🆕 Chatbot styles
│   │   └── yolo_farm.css       # 🆕 Yolo:Farm lesson UI styles
│   └── js/
│       ├── chatbot.js          # 🆕 Chatbot client logic
│       └── yolo_farm.js        # 🆕 Yolo:Farm + Web Serial API logic
│
├── requirements/
│   ├── base.txt                # All dependencies
│   ├── local.txt               # Development extras (black, factory-boy)
│   └── production.txt          # Production extras
├── manage.py
└── .env.example                # Environment variable template

Requirements

Installation

  1. Clone the repository
git clone https://github.com/nat50/SkyLearnAI.git
cd SkyLearnAI
  1. Create and activate a virtual environment
python -m venv .venv

# Windows
.venv\Scripts\activate

# macOS / Linux
source .venv/bin/activate
  1. Install dependencies
pip install -r requirements.txt
  1. Configure environment variables
cp .env.example .env

Edit .env with your actual credentials (see Environment Variables below).

  1. Run database migrations
python manage.py migrate
  1. Create a superuser (admin)
python manage.py createsuperuser
  1. Start the development server
python manage.py runserver
  1. Open http://127.0.0.1:8000 in your browser.

Environment Variables

Create a .env file in the project root with the following variables:

# Django
DEBUG=True
SECRET_KEY=your-random-secret-key

# Email
EMAIL_BACKEND=django.core.mail.backends.console.EmailBackend
EMAIL_HOST=smtp.gmail.com
EMAIL_PORT=587
EMAIL_USE_TLS=True
EMAIL_FROM_ADDRESS=SkyLearn <youremail@example.com>
EMAIL_HOST_USER=youremail@example.com
EMAIL_HOST_PASSWORD=yourpassword

# PostgreSQL Database
DB_NAME=your_db_name
DB_USER=your_db_user
DB_PASSWORD=your_db_password
DB_HOST=your_db_host
DB_PORT=5432

# Cloudinary (media storage)
CLOUDINARY_CLOUD_NAME=your_cloud_name
CLOUDINARY_API_KEY=your_api_key
CLOUDINARY_API_SECRET=your_api_secret

# Google Gemini AI (required for AI features)
GOOGLE_API_KEY=your_gemini_api_key

Usage

Role Access
Admin Full dashboard, manage students/lecturers/courses, view analytics
Lecturer Submit grades, view allocated courses, upload materials, generate AI lessons
Student Enroll in courses, view grades/results, take quizzes, use AI chatbot
Parent View connected student's information

AI Endpoints

Endpoint Method Description
/ai/lesson/generate/ POST Generate an AI lesson (with optional RAG context)
/ai/lesson/save-doc/ POST Save generated lesson as DOCX to course materials
/ai/lesson/<id>/html/ GET Retrieve stored HTML lesson content
/ai/quiz/generate/ POST Generate MCQ quiz from lesson content
/ai/chatbot/ POST Send message to AI chatbot

AI Features in Detail

RAG Pipeline

  1. Document Upload — Lecturers upload PDF/DOCX/PPTX files to a course
  2. Text Extraction — document_parser.py extracts text using PyPDF2, python-docx, or python-pptx
  3. Chunking — Text is split into 4000-character overlapping chunks (200-char overlap)
  4. Embedding — Each chunk is embedded using gemini-embedding-001 via LangChain
  5. Storage — Chunks and embeddings are stored in the DocumentChunk model (PostgreSQL JSON field)
  6. Retrieval — At lesson generation time, the user's topic is embedded and compared against stored chunks using cosine similarity (NumPy)
  7. Generation — Top-K relevant chunks are injected into the Gemini prompt as context

Yolo:Farm AIoT

  • Generates lessons specifically for the Yolo:Bit microcontroller (OhStem platform)
  • Code blocks contain real MicroPython code using OhStem libraries (from yolobit import *)
  • Lessons can use the Yolo:Farm textbook PDF as RAG context (auto-detected)
  • Generated code can be uploaded to physical hardware via the Web Serial API directly from the browser

References

License

This project is licensed under the MIT License.

Original SkyLearn © 2023 Adil Mohak.

About

The AI version of SkyLearn lms

Resources

Readme

License

MIT license

Code of conduct

Code of conduct

Contributing

Contributing
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Python 49.9%
  • HTML 38.3%
  • JavaScript 4.9%
  • SCSS 3.7%
  • CSS 3.2%