JEE Mains PYQS Database

This project provides a structured database of more than 14,000 previous year questions (PYQS) from JEE Mains. The questions are reverse engineered from API endpoints of a subscription site and cached for efficient use. It supports clustering, filtering, and rendering of questions into HTML for easy study.

Features

• Access to 14k+ JEE Mains PYQS
• Precomputed embeddings using the intfloat/e5-large-v2 model for efficient clustering
• Cluster similar questions together based on semantic embeddings
• Apply chainable filters (by chapter, topic, year, etc.)
• Render filtered or clustered questions into HTML using themed styles

Project Structure

The core folder contains the following modules:

• cache.py – Defines the Cache class for creating and loading internal caches. Not intended for direct user interaction.
• chapter.py – Defines the Chapter class, which is stored in the DataBaseChapters cache file. Internal use only.
• data_base.py – Defines the DataBase class. This must be initialized before any operations.
• filter.py – Defines the Filter class. Provides chainable methods to filter questions and update the current set.
• question.py – Defines the Question object.
• styles.py – Contains themed HTML styles for rendering.
• pdfy.py – Provides functions to convert clusters or sets of questions into HTML.

Installation

• Install using pip:

pip install jee_data_base

• Clone the repository:

git clone https://github.com/HostServer001/jee_mains_pyqs_data_base

Navigate into the project directory and ensure dependencies are installed.

Usage

Basic Initialization

import os
from jee_data_base import DataBase, Filter, pdfy

# Initialize database
db = DataBase()

# Initialize filter
filter = Filter(db.chapters_dict)

# Inspect available chapters
print(filter.get_possible_filter_values()["chapter"])

Its highly recommended to filter as much as possible so that your html files open smoothly in browser

Its always good to use the cluster method and render_cluster_to_html method to get your output, it provides the most efficeint way of practice

The render_cluster_to_html_skim is great if you have prepared chapter loosely and want to skin thorugh and get most out of it (use it after cluster)

Most useful feature

from jee_data_base import DataBase,Filter

path = "<path where chpater folder will be created>"
chpater = "<your example chpater>"

#Load the data base
db = DataBase()

#Initialize filter
filter = Filter(db.chapter_dict)

#Create html file
filter.render_chap_last5yrs(path,chpater,skim=False)

Filtering by Chapter and Year

# Get all questions from a specific chapter in the last 3 years
questions = filter.by_chapter("thermodynamics").by_n_last_yrs(3).get()

for q in questions:
    print(q.question)

Clustering and Rendering

# Cluster questions by topic and render to HTML
filter.current_set = filter.by_chapter("organic-compounds").by_n_last_yrs(5).get()
cluster = filter.cluster()

pdfy.render_cluster_to_html(
    cluster,
    "organic_compounds.html",
    "Organic Compounds - Last 5 Years"
)# can use render_cluster_to_html_skim() function to make a file which 
#makes a html file perfected for skiming through a chapter

Example: Render Chapter Questions by Topic

def render_chapter(chapter_name: str):
    all_q = filter.by_chapter(chapter_name).by_n_last_yrs(5).get()
    os.makedirs(chapter_name, exist_ok=True)

    for topic in filter.get_possible_filter_values()["topic"]:
        filter.current_set = all_q
        filter.by_topic(topic)
        cluster = filter.cluster()
        pdfy.render_cluster_to_html_skim(
            cluster,
            f"{chapter_name}/{topic}.html",
            topic
        )

render_chapter("alcohols-phenols-and-ethers")

Ouput

• The output will look somthing like this PDF 📄

Data Caches

• DataBaseChapters – Contains a dictionary with chapter names as keys and Chapter objects as values.
• EmbeddingsChapters – Contains precomputed embeddings of all questions to save computation time.

Contributing

Contributions are welcome. You can help by:

• Improving documentation
• Adding new filters or clustering strategies
• Enhancing rendering styles
• Reporting issues and suggesting features

Fork the repository, create a new branch for your changes, and submit a pull request.

License

This project is provided for educational purposes. Please review the repository for licensing details.