apps/functions/utils.py

import fitz  # PyMuPDF
import numpy as np
import cv2
import tempfile
import os
from pathlib import Path

def pdf_to_images(pdf_path, dpi=200):
    doc = fitz.open(pdf_path)
    pages = []

    for page_num in range(len(doc)):
        page = doc[page_num]
        mat = fitz.Matrix(dpi / 72, dpi / 72)
        pix = page.get_pixmap(matrix=mat, alpha=False)

        img_array = np.frombuffer(pix.samples, dtype=np.uint8)
        img_array = img_array.reshape(pix.height, pix.width, 3)
        img_bgr = cv2.cvtColor(img_array, cv2.COLOR_RGB2BGR)

        pages.append((page_num + 1, img_bgr))

    doc.close()
    return pages

def ensure_output_dir(path= "outputs"):
    output_dir = Path(path)
    output_dir.mkdir(exist_ok=True)
    return output_dir

def get_temp_path(page_num):
    tmp_dir = Path("outputs/tmp")
    tmp_dir.mkdir(parents=True, exist_ok=True)
    return str(tmp_dir / f"page_{page_num}.jpg")

def cleanup_temp(path):
    try:
        if os.path.exists(path):
            os.remove(path)
    except Exception:
        pass