Skip to content

Project introduces an innovative AI-assisted chemistry learning platform designed to fundamentally shift the pedagogical approach to organic reaction mechanisms from rote memorization to deep, reasoning-based understanding

Notifications You must be signed in to change notification settings

chadvik88/JEEChemSage.AI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

⌬ JeeChemSage.AI

A Visual Chemistry Tutor for IIT-JEE Aspirants

JeeChemSage.AI is an interactive chemistry learning tool built to help students master organic reaction mechanisms for competitive exams like the IIT-JEE. Designed with high school learners in mind, the app visually breaks down complex reactions, identifies functional groups, predicts reaction classes, and even tests your understanding with MCQs — all within a clean Streamlit interface.


📸 Live Demo

🧬 Try the app (Streamlit link)

Application Screenshots

Features

1. Smart Reaction Parsing

  • Accepts custom or sample organic reactions like ethyl bromide + KOH -> ethene + HBr
  • Automatically translates IUPAC/common names to SMILES using RDKit and PubChem fallback
  • Parses reactants/products using custom jee_parser.py

2. Reaction Classification

  • Identifies reaction mechanism (e.g., E2, SN1, SN2, etc.)
  • Detects key functional groups (alcohols, alkyl halides, amines, ethers, etc.)
  • Displays both ML-predicted reaction class and rule-based interpretation

3. Molecular Properties (Reactants & Products)

  • Molecular weight
  • LogP
  • Hydrogen donors/acceptors
  • Topological Polar Surface Area (TPSA)
  • Rotatable bonds

4. Mechanism Explanation (ELI15 Mode)

  • Option to "Explain like I’m 15" for each predicted mechanism
  • Helps beginners understand the why behind each transformation
  • Example:

    “Imagine a Lego model where you pull off one brick (hydrogen) and another (Br) pops off at the same time...”

5. Visual Mechanism Diagrams

  • Loads mechanism-specific PNG diagrams from assets/mechanisms/ (like e2.png, sn2.png)
  • Visual walkthrough of how bonds break and form

6. "Test Yourself" MCQs

  • Auto-generates multiple-choice questions based on the input reaction
  • Provides immediate feedback on correctness
  • Covers mechanism identification, byproducts, nucleophile/electrophile roles, etc.

7. Reaction History Sidebar

  • Easily re-run or revisit previous inputs via clickable buttons in the sidebar

🛠 Tech Stack

Tool Use
Python 3.10+ Core application logic
Streamlit Frontend + UI rendering
RDKit Chemistry parsing, SMILES, and drawings
PubChemPy Backup molecular name resolution
Pillow/3Dmol.js 2D and 3D Molecule Visualization

Project Structure

JeeChemSage.AI/
│
├── app/
│   └── streamlit_app.py         # Main app interface
│
├── core/                        # Core logic modules
│   ├── jee_parser.py            # Reaction parser
│   ├── name_to_smiles.py        # Name to SMILES resolver (RDKit + PubChem fallback)
│   ├── properties.py            # Molecular property calculator
│   ├── reaction_classifier.py   # Rule-based reaction classification
│   ├── functional_groups.py     # SMARTS pattern detection
│   ├── mechanism_visuals.py     # Mechanism image resolver
│   └── mcq_generator.py         # MCQ generator
│
├── assets/
│   └── mechanisms/              # PNG diagrams (e.g., e2.png, sn1.png)
│
└── README.md

Sample Reactions:

  • reaction: "ethyl bromide + KOH -> ethene + HBr" type: "E2 Elimination"
  • reaction: "tert-butyl bromide + H2O -> tert-butyl alcohol + HBr" type: "SN1"
  • reaction: "ethanol + H2SO4 -> ethene + H2O" type: "E1"
  • reaction: "benzamide + Br2 + KOH -> aniline + KBr + CO2" type: "Hoffmann Rearrangement"

Future Ideas:

  • "ML-powered reaction classification (currently rule-based fallback)"
  • "Intelligent MCQ difficulty levels"
  • "Integration with NCERT & Allen chemistry syllabus"
  • "Voice input (e.g., via Whisper API)"
  • "Online hosting via Streamlit Cloud or Hugging Face Spaces"

Installation:

  - step: Clone the repository
    command: "git clone https://github.com/yourusername/JeeChemSage.AI.git"
  - step: Navigate into the project directory
    command: "cd JeeChemSage.AI"
  - step: Install dependencies (recommended inside a virtual environment)
    command: "pip install -r requirements.txt"
  - step: Run the Streamlit app
    command: "streamlit run app/streamlit_app.py"
  - tip: "If you see RDKit or SMILES parsing errors, make sure you’ve installed RDKit correctly"
    command: "conda install -c rdkit rdkit"

Tagline: "For students, by a student."

About

Project introduces an innovative AI-assisted chemistry learning platform designed to fundamentally shift the pedagogical approach to organic reaction mechanisms from rote memorization to deep, reasoning-based understanding

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages