PDF rendering with python-poppler

poppler.py

# SPDX-FileCopyrightText: 2024 geisserml <[email protected]>
# SPDX-License-Identifier: MPL-2.0
# Note that Poppler is GPL-licensed, so this code is altogether affected by copyleft

import PIL.Image
import poppler  # python-poppler
from poppler.cpp.page_renderer import render_hint


def _translate_rotation(rotation):
    # Beware: this enum contained two typos with python-poppler <= 0.2.2
    # https://github.com/cbrunet/python-poppler/issues/42
    return {
        0: poppler.Rotation.rotate_0,
        90: poppler.Rotation.rotate_90,
        180: poppler.Rotation.rotate_180,
        270: poppler.Rotation.rotate_270,
    }[rotation]


def invoke_poppler_py(filepath, index, scale=4, rotation=0, password=None):
    
    # Note, as of this writing, poppler does not take /UserUnit into account.
    
    if password is None:
        pdf = poppler.load_from_file(filepath)
    else:
        pdf = poppler.load_from_file(
            filepath,
            owner_password = password,
            user_password  = password,
        )
    
    page = pdf.create_page(index)
    dpi = scale * 72
    
    # output looks poor without general anti-aliasing
    renderer = poppler.PageRenderer()
    renderer.set_render_hint(hint=render_hint.antialiasing, on=True)  # -> 1
    renderer.set_render_hint(hint=render_hint.text_antialiasing, on=True)  # -> 2
    
    image = renderer.render_page(
        page, xres=dpi, yres=dpi,
        rotate = _translate_rotation(rotation),
    )
    
    pil_image = PIL.Image.frombytes("RGBA", (image.width, image.height), image.data, "raw", str(image.format))
    
    return pil_image