📘 SGPA — Study Guide & Personal Assistant

AI-Powered Study Assistant

---

---

🧠 Project Overview

Students often struggle to grasp difficult topics or summarize lengthy notes.
SGPA is an AI-powered web app that acts as a personal academic assistant, capable of:

• 🧩 Explaining complex concepts in simple terms
• 📄 Summarizing notes or uploaded PDFs
• ❓ Generating quizzes, solving exam questions, and evaluating answers

Combines Streamlit for UI and Gemini 2.5 Flash API for fast, intelligent AI responses — all in a clean chat-based interface.

---

Recent updates (v1.2.0):

• Explainer now suggests simple sketchable diagrams for visual understanding.
• Summarizer handles very short or scanned PDFs more safely with clearer warnings.
• Added lightweight, privacy-friendly usage analytics (anonymous sessions and mode usage).

---

⚙️ System Design

🏗️ Architecture

A lightweight Streamlit frontend interacts with Google Gemini 2.5 Flash backend via secure API calls.
All secrets managed safely via .env and st.secrets.

🧩 Core Features

Mode	Function	Example
🧠 Explainer	Simplifies academic concepts	“Explain Deadlock in OS”
📄 Summarizer	Condenses notes or PDFs	Upload 20-page PDF → summary
🧩 Quizzer	Quiz generator, solver, evaluator (multi-mode workflow)	MCQs, solve/evaluate Q&As

Other Features:

• 📂 PDF upload (PyPDF2 extraction, with guards for very short/empty text)
• 📊 Lightweight usage logging for sessions and mode usage (CSV-based, privacy-friendly)
• 💬 Real-time chat interface
• 🔄 New chat/reset option
• ☁️ Deployed on Streamlit Cloud

---

🧙‍♂️ Quizzer Mode — Three Powerful Sub-modes

1. 📝 Generate Questions
   Enter a topic/chapter/passage. Get a variety of questions (MCQ, T/F, fill-in, descriptive) — answers listed together as an answer key for self-testing.
2. 📖 Solve Questions
   Paste your exam questions (optionally add word limits or marks). Get concise, exam-ready answers formatted per input.
3. ✅ Evaluate Answers
   Submit questions and your answers (with '---' separator, or sequential prompts). Get detailed feedback, correction, and scoring.

---

🧱 Project Structure

SGPA/
├── main.py
├── requirements.txt
├── assets/
│ └── PROBLEM STATEMENTS.pdf
├── components/
│ ├── chat_ui.py
│ ├── pdf_handler.py
│ └── sidebar.py
├── core/
│ ├── ai_utils.py
│ ├── explainer.py
│ ├── pdf_handler.py
│ ├── quizzer.py
│ └── summarizer.py
└── utils/
└── gemini_helper.py

---

🪜 Workflow

©️🖼️ Diagram Credits: https://gitdiagram.com/

---

📚 In-Repo User Guide (Quick Start)

You can keep the PDF as the detailed reference and mirror a concise, in-README guide here.

1️⃣ Getting Started

• Open the deployed app: https://sgpai-study-buddy.streamlit.app/
• Select a mode from the sidebar: Explainer, Summarizer, or Quizzer
• Provide input (topic, notes, PDF, or questions) in the main chat area

2️⃣ Mode Usage

• Explainer:
  Type your concept or question (e.g., “Explain paging in OS for exams”).
  SGPA returns a simple, exam-oriented explanation, plus an idea for a quick diagram you can sketch.

• Summarizer:
  Upload a PDF or paste notes.
  SGPA returns an exam-ready summary with headings, bullets, and practice questions.

• Quizzer:

  • Use “Generate Questions” for practice questions with an answer key.
  • Use “Solve Questions” to get answers to your questions.
  • Use “Evaluate Answers” to paste both question and your answer to receive feedback and scoring.

3️⃣ Tips for Best Results

• Mention exam context (e.g., “for B.Tech 3rd sem OS viva”) for sharper responses.
• Use follow-up prompts in the same chat to refine or extend answers.
• Reset the chat using the “New Chat” / reset option before switching topics heavily.

For full details, screenshots, and troubleshooting, refer to the PDF:

---

💡 Tech Stack

Category	Technologies
Frontend	Streamlit
Backend / AI	Google Gemini 2.5 Flash API
Language	Python
Libraries	PyPDF2, google-generativeai, streamlit, dotenv
Deployment	Streamlit Community Cloud
Security	.env + st.secrets key handling

---

🧾 Results

• 🎯 Simple, modern, and interactive chat-based UI
• 📑 Smart summarization, quiz generation, and answer evaluation
• ⚡ Fast, context-aware AI with Gemini 2.5 Flash
• 🧩 Smooth multi-mode workflow for study and revision

---

🚀 Future Scope

• 🗣️ Speech-to-text / text-to-speech interaction
• 🌐 Multi-language explanations
• 🧠 Flashcard & spaced-repetition support
• 👤 Memory-based user personalization
• ☁️ Drive/Notion integration for notes & sessions

---

🧩 “Integrating AI with Education — Making Learning Simpler, Smarter, and Accessible for All.”

---

📜 Usage & Attribution

• You are welcome to fork this repository to learn from it or build your own version of SGPA.
• If you deploy this project publicly or create a derivative version:
  • Keep the existing license file.
  • Credit “SGPA by Ammaar Ahmad Khan (GPA95)”.
  • Include a link back to the original repo:
    • https://github.com/GPA95/SGPA

For contributions, please see CONTRIBUTING.md for guidelines on forking, branching, and opening pull requests.

---

👨‍💻 Author

Ammaar Ahmad Khan

• GitHub: @GPA95

🌟 If you find this repository useful, please give it a star! 🌟

---