sem_three_eligify

Overview

ELIGIFY is a simple Flask-backed app with a frontend for exam eligibility checks. The backend exposes an endpoint to parse PDF files. PDF parsing now prefers the embedded text layer (PyPDF2) and falls back to OCR via Tesseract when needed.

OCR Setup (Windows)

To use the OCR-based PDF parser, install these prerequisites:

• Install Tesseract OCR: https://github.com/UB-Mannheim/tesseract/wiki

  • Default install path: C:\Program Files\Tesseract-OCR\tesseract.exe
  • Optionally set env var TESSERACT_CMD to the full path of tesseract.exe.

• Install Poppler for Windows (required by pdf2image):

  • Download from: https://github.com/oschwartz10612/poppler-windows/releases/
  • Add the bin folder to your system PATH.
  • Alternatively, set environment variable POPPLER_PATH to the Poppler bin directory and the app will use it. Example:
    • PowerShell (current session): $env:POPPLER_PATH = 'C:\\Tools\\poppler-24.08.0\\Library\\bin'
    • Permanent: setx POPPLER_PATH "C:\\Tools\\poppler-24.08.0\\Library\\bin"

Windows Quick Setup Commands

Run these in PowerShell:

# Set for current session
$env:TESSERACT_CMD = 'C:\\Program Files\\Tesseract-OCR\\tesseract.exe'
$env:POPPLER_PATH = 'C:\\Tools\\poppler-24.08.0\\Library\\bin'

# Persist for future sessions
setx TESSERACT_CMD "C:\\Program Files\\Tesseract-OCR\\tesseract.exe"
setx POPPLER_PATH "C:\\Tools\\poppler-24.08.0\\Library\\bin"

Note: method="text" does not require Poppler/Tesseract; method="ocr" requires both.

Python Dependencies

Install Python packages:

pip install -r requirements.txt

requirements.txt includes: flask, PyPDF2, pytesseract, pdf2image, Pillow, opencv-python.

Notes:

• PyPDF2 is used to extract text from PDFs that have a selectable text layer (no extra system dependencies).
• OpenCV is used to pre-process scanned pages before OCR.

Run the App

Start the Flask server:

python app.py

Visit http://localhost:3000/ in a browser.

PDF Parsing API

• Endpoint: POST /api/parse-pdf

• Form field: file (PDF)

• Optional query params:

  • method: auto (default), text, or ocr
  • dpi: integer DPI for rasterization (default 300)

• Response:

{
  "text": "...extracted text...",
  "method": "auto"
}

Example (curl)

curl -F "file=@C:\\path\\to\\your.pdf" "http://localhost:3000/api/parse-pdf?method=auto&dpi=300"

How To Use (Library)

You can import and call the parser directly in Python code.

Simple import and call

from lib.pdf_parser import extract_text_from_pdf

text = extract_text_from_pdf("lib/test_input.pdf")
print(text)

Control parsing strategy

# Prefer text layer, fallback to OCR if empty (default)
text_auto = extract_text_from_pdf("lib/test_input.pdf", method="auto")

# Force text-layer only (PyPDF2)
text_text = extract_text_from_pdf("lib/test_input.pdf", method="text")

# Force OCR-only (Tesseract)
text_ocr = extract_text_from_pdf("lib/test_input.pdf", method="ocr", dpi=300)

Using with uploaded files (Flask request.files['file'])

from flask import request
from lib.pdf_parser import extract_text_from_pdf

@app.post('/api/parse-pdf')
def parse_pdf_ep():
    f = request.files.get('file')
    if not f:
        return {"error": "No file uploaded."}, 400
    text = extract_text_from_pdf(f)  # defaults to method="auto"
    return {"text": text}