malayalam-ocr / ocr /pdf_utils.py
afueth
Malayalam OCR app
835eeaa
Raw History Blame Contribute Delete
696 Bytes
from __future__ import annotations
import io
import fitz # PyMuPDF
from PIL import Image
def pdf_has_native_text(page: fitz.Page, min_chars: int = 50) -> bool:
txt = (page.get_text("text") or "").strip()
return len(txt) >= min_chars
def extract_native_text(page: fitz.Page) -> str:
return (page.get_text("text") or "").strip()
def render_page_to_pil(page: fitz.Page, zoom: float = 2.0) -> Image.Image:
"""
Render PDF page to image for OCR. zoom ~2.0 is a good balance.
Higher zoom -> better OCR but slower.
"""
mat = fitz.Matrix(zoom, zoom)
pix = page.get_pixmap(matrix=mat, alpha=False)
img = Image.open(io.BytesIO(pix.tobytes("png")))
return img