The Data Pipeline

​

Overview

PyqDeck’s core value is its collection of university question papers. This page explains how papers flow into the system and how questions and solutions are managed.

​

Implementation Map

​

1. Upload Pipeline (backend/src/utils/uploadthing.js)

Users upload question paper PDFs through UploadThing.

• File: backend/src/utils/uploadthing.js
• Model: backend/src/models/Upload.js

When an upload completes, the backend receives a callback and stores the metadata in the Upload collection.

​

2. Paper Management (backend/src/controllers/paperController.js)

Paper metadata (title, year, etc.) is managed via the paperController.

• Service: backend/src/services/paperService.js
• Repository: backend/src/repositories/paperRepository.js
• Model: backend/src/models/Paper.js

Papers go through a status workflow: draft → pending → approved. Public users can only see approved papers.

​

3. Question & Syllabus Mapping

Questions are linked to papers via a join table pattern.

• Model: backend/src/models/Question.js
• Mapping: backend/src/models/QuestionPaperMap.js (links Question to Paper)
• Syllabus: backend/src/models/QuestionSyllabusMap.js (links Question to Syllabus Topic)

This allows for polymorphic questions that can belong to multiple papers or topics.

​

4. Solutions

Solutions are stored separately and linked to questions.

• Model: backend/src/models/Solution.js
• Types: teacher, student, ai

​

Academic Hierarchy

The system follows a strict hierarchy defined in the models/ directory:

1. University (University.js)
2. Branch (Branch.js)
3. Semester (Semester.js)
4. Subject (Subject.js)
5. SubjectOffering (SubjectOffering.js) - The link between a subject and a specific semester.
6. Paper (Paper.js)

​

Future: AI-Powered Parsing

The architecture is designed to support automated PDF parsing. The Solution.type: 'ai' enum is already reserved for AI-generated answers, and the Question model includes fields for normalized text to aid in AI extraction.