New package: python3-ocrmypdf-16.10.1

srcpkgs/python3-ocrmypdf/patches/tesseract_binary_name_fixup.patch

@@ -0,0 +1,29 @@
+--- a/src/ocrmypdf/_exec/tesseract.py	2025-02-27 00:16:18.000000000 +0100
++++ b/src/ocrmypdf/_exec/tesseract.py	2025-03-17 11:56:25.146770854 +0100
+@@ -95,7 +95,7 @@
+ 
+ 
+ def version() -> Version:
+-    return TesseractVersion(get_version('tesseract', regex=r'tesseract\s(.+)'))
++    return TesseractVersion(get_version('tesseract-ocr', regex=r'tesseract\s(.+)'))
+ 
+ 
+ def has_thresholding() -> bool:
+@@ -113,7 +113,7 @@
+         msg += output
+         return msg
+ 
+-    args_tess = ['tesseract', '--list-langs']
++    args_tess = ['tesseract-ocr', '--list-langs']
+     try:
+         proc = run(
+             args_tess,
+@@ -135,7 +135,7 @@
+ 
+ 
+ def tess_base_args(langs: list[str], engine_mode: int | None) -> list[str]:
+-    args = ['tesseract']
++    args = ['tesseract-ocr']
+     if langs:
+         args.extend(['-l', '+'.join(langs)])
+     if engine_mode is not None:

srcpkgs/python3-ocrmypdf/patches/use_pillow_heif_instead_of_pi_heif.patch

@@ -0,0 +1,22 @@
+--- a/pyproject.toml	2025-02-27 00:16:18.000000000 +0100
++++ b/pyproject.toml	2025-03-17 15:20:09.011808694 +0100
+@@ -16,7 +16,7 @@
+   "img2pdf>=0.5",
+   "packaging>=20",
+   "pdfminer.six>=20220319",
+-  "pi-heif",                # Heif image format - maintainers: if this is removed, it will NOT break
++  "pillow-heif",                # Heif image format - maintainers: if this is removed, it will NOT break
+   "pikepdf>=8.10.1",
+   "Pillow>=10.0.1",
+   "pluggy>=1",
+--- a/src/ocrmypdf/_pipeline.py	2025-02-27 00:16:18.000000000 +0100
++++ b/src/ocrmypdf/_pipeline.py	2025-03-17 15:21:10.155848376 +0100
+@@ -42,7 +42,7 @@
+ from ocrmypdf.pluginspec import OrientationConfidence
+ 
+ try:
+-    from pi_heif import register_heif_opener
++    from pillow_heif import register_heif_opener
+ except ImportError:
+ 
+     def register_heif_opener():

srcpkgs/python3-ocrmypdf/template

@@ -0,0 +1,15 @@
+# Template file for 'python3-ocrmypdf'
+pkgname=python3-ocrmypdf
+version=16.10.1
+revision=1
+build_style=python3-pep517
+hostmakedepends="hatchling hatch-vcs"
+depends="python3-deprecation img2pdf python3-packaging python3-pdfminer.six
+ python3-pikepdf python3-Pillow python3-pluggy python3-reportlab python3-rich
+ python3-pillow_heif tesseract-ocr ghostscript unpaper pngquant jbig2enc qpdf"
+short_desc="Add OCR text layer to scanned PDF files"
+maintainer="Mateusz Sylwestrzak <slymattz@gmail.com>"
+license="MPL-2.0"
+homepage="https://github.com/ocrmypdf/OCRmyPDF"
+distfiles="${PYPI_SITE}/o/ocrmypdf/ocrmypdf-${version}.tar.gz"
+checksum=9f32059fc97e25931aaa0a8a4027b8c9faca7d9e1183089f32e0cba5631449f1