python3-ocrmypdf: update to 17.2.0

[slymattz]

Testing the changes

• I tested the changes in this PR: YES

Version 17 depends now on pypdfium2 as

an alternative to Ghostscript for page rendering

https://raw.githubusercontent.com/ocrmypdf/OCRmyPDF/refs/heads/main/docs/release_notes.md

EDIT 1/31/26: python3-ocrmypdf updated to 17.1.0
added changelog and shell completions
EDIT 2/9/26: python3-pypdfium2 updated to 5.4.0
EDIT 2/11/26: switched to GH for pypdfium2 as it contains tests to run
Decided to use a simple and conservative versioning convention for SONAME
https://pypdfium2.readthedocs.io/en/stable/readme.html#with-system-pdfium

I additionally tested python bindings (i.e. pypdfium2) be executing the following code:

import pypdfium2 as pdfium

pdf = pdfium.PdfDocument.new()
pdf.new_page(width=595, height=842)

output_filename = "test_output.pdf"
pdf.save(output_filename)

print(f"Saved successfully to: {output_filename}")

pdf.close()

import math
import ctypes
import os.path
import PIL.Image
import pypdfium2.raw as pdfium_c

# Load the document
filepath = os.path.abspath("./MOC-61793_Lenarex_tux_linux.pdf")
pdf = pdfium_c.FPDF_LoadDocument((filepath+"\x00").encode("utf-8"), None)

# Check page count to make sure it was loaded correctly
page_count = pdfium_c.FPDF_GetPageCount(pdf)
assert page_count >= 1

# Load the first page and get its dimensions
page = pdfium_c.FPDF_LoadPage(pdf, 0)
width  = math.ceil(pdfium_c.FPDF_GetPageWidthF(page))
height = math.ceil(pdfium_c.FPDF_GetPageHeightF(page))

# Create a bitmap
# (Note, pdfium is faster at rendering transparency if we use BGRA rather than BGRx)
use_alpha = pdfium_c.FPDFPage_HasTransparency(page)
bitmap = pdfium_c.FPDFBitmap_Create(width, height, int(use_alpha))
# Fill the whole bitmap with a white background
# The color is given as a 32-bit integer in ARGB format (8 bits per channel)
pdfium_c.FPDFBitmap_FillRect(bitmap, 0, 0, width, height, 0xFFFFFFFF)

# Store common rendering arguments
render_args = (
    bitmap,  # the bitmap
    page,    # the page
    # positions and sizes are to be given in pixels and may exceed the bitmap
    0,       # left start position
    0,       # top start position
    width,   # horizontal size
    height,  # vertical size
    0,       # rotation (as constant, not in degrees!)
    pdfium_c.FPDF_LCD_TEXT | pdfium_c.FPDF_ANNOT,  # rendering flags, combined with binary or
)

# Render the page
pdfium_c.FPDF_RenderPageBitmap(*render_args)

# Get the value of a pointer to the first item of the buffer
buffer_ptrval = pdfium_c.FPDFBitmap_GetBuffer(bitmap)
assert buffer_ptrval, "buffer pointer value must be non-null"
# Cast the pointer value to an actual pointer object so we can access .contents
buffer_ptr = ctypes.cast(buffer_ptrval, ctypes.POINTER(ctypes.c_ubyte))
# Re-interpret as array
buffer = (ctypes.c_ubyte * (width * height * 4)).from_address(ctypes.addressof(buffer_ptr.contents))

# Create a PIL image from the buffer contents
img = PIL.Image.frombuffer("RGBA", (width, height), buffer, "raw", "BGRA", 0, 1)
# Save it as file
img.save("out.png")

# Free resources
pdfium_c.FPDFBitmap_Destroy(bitmap)
pdfium_c.FPDF_ClosePage(page)
pdfium_c.FPDF_CloseDocument(pdf) 

https://pypi.org/project/pypdfium2/

Local build testing

• I built this PR locally for my native architecture: x86_64-glibc
• I built this PR locally for these architectures (if supported. mark crossbuilds):
  • aarch64-musl (crossbuild)
  • armv6l-musl (crossbuild)