tesseract-ocr: update to 4.1.1

Also add all script languages as separate subpackages.
2019-12-28 15:10:27 +01:00 · 2019-12-28 15:10:27 +01:00 · aee0563a01
parent 56f113ec21
commit aee0563a01
34 changed files with 322 additions and 12 deletions
--- a/srcpkgs/tesseract-ocr-script-Arabic
+++ b/srcpkgs/tesseract-ocr-script-Arabic
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Armenian
+++ b/srcpkgs/tesseract-ocr-script-Armenian
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Bengali
+++ b/srcpkgs/tesseract-ocr-script-Bengali
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Canadian_Aboriginal
+++ b/srcpkgs/tesseract-ocr-script-Canadian_Aboriginal
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Cherokee
+++ b/srcpkgs/tesseract-ocr-script-Cherokee
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Cyrillic
+++ b/srcpkgs/tesseract-ocr-script-Cyrillic
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Devanagari
+++ b/srcpkgs/tesseract-ocr-script-Devanagari
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Ethiopic
+++ b/srcpkgs/tesseract-ocr-script-Ethiopic
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Fraktur
+++ b/srcpkgs/tesseract-ocr-script-Fraktur
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Georgian
+++ b/srcpkgs/tesseract-ocr-script-Georgian
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Greek
+++ b/srcpkgs/tesseract-ocr-script-Greek
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Gujarati
+++ b/srcpkgs/tesseract-ocr-script-Gujarati
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Gurmukhi
+++ b/srcpkgs/tesseract-ocr-script-Gurmukhi
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-HanS
+++ b/srcpkgs/tesseract-ocr-script-HanS
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-HanT
+++ b/srcpkgs/tesseract-ocr-script-HanT
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Hangul
+++ b/srcpkgs/tesseract-ocr-script-Hangul
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Hebrew
+++ b/srcpkgs/tesseract-ocr-script-Hebrew
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Japanese
+++ b/srcpkgs/tesseract-ocr-script-Japanese
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Kannada
+++ b/srcpkgs/tesseract-ocr-script-Kannada
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Khmer
+++ b/srcpkgs/tesseract-ocr-script-Khmer
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Lao
+++ b/srcpkgs/tesseract-ocr-script-Lao
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Latin
+++ b/srcpkgs/tesseract-ocr-script-Latin
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Malayalam
+++ b/srcpkgs/tesseract-ocr-script-Malayalam
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Myanmar
+++ b/srcpkgs/tesseract-ocr-script-Myanmar
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Oriya
+++ b/srcpkgs/tesseract-ocr-script-Oriya
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Sinhala
+++ b/srcpkgs/tesseract-ocr-script-Sinhala
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Syriac
+++ b/srcpkgs/tesseract-ocr-script-Syriac
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Tamil
+++ b/srcpkgs/tesseract-ocr-script-Tamil
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Telugu
+++ b/srcpkgs/tesseract-ocr-script-Telugu
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Thaana
+++ b/srcpkgs/tesseract-ocr-script-Thaana
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Thai
+++ b/srcpkgs/tesseract-ocr-script-Thai
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Tibetan
+++ b/srcpkgs/tesseract-ocr-script-Tibetan
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr-script-Vietnamese
+++ b/srcpkgs/tesseract-ocr-script-Vietnamese
@ -0,0 +1 @@
+tesseract-ocr
--- a/srcpkgs/tesseract-ocr/template
+++ b/srcpkgs/tesseract-ocr/template
@ -1,11 +1,12 @@
 # Template file for 'tesseract-ocr'
 pkgname=tesseract-ocr
-version=4.1.0
-revision=2
+version=4.1.1
+revision=1
 _tessdataver=4.0.0
 wrksrc="tesseract-${version}"
 build_style=gnu-configure
 configure_args="LIBLEPT_HEADERSDIR=${XBPS_CROSS_BASE}/usr/include $(vopt_enable openmp)"
+make_build_args="all training"
 hostmakedepends="automake libtool pkg-config leptonica libxslt asciidoc"
 makedepends="cairo-devel pango-devel leptonica-devel $(vopt_if openmp libgomp-devel) icu-devel"
 short_desc="Tesseract Open Source OCR engine"
@ -15,7 +16,7 @@ homepage="https://github.com/tesseract-ocr/tesseract"
 distfiles="
 https://github.com/tesseract-ocr/tesseract/archive/${version}.tar.gz>${pkgname}-${version}.tar.gz
 https://github.com/tesseract-ocr/tessdata/archive/${_tessdataver}.tar.gz>tessdata-${_tessdataver}.tar.gz"
-checksum="5c5ed5f1a76888dc57a83704f24ae02f8319849f5c4cf19d254296978a1a1961
+checksum="2a66ff0d8595bff8f04032165e6c936389b1e5727c3ce5a27b3e059d218db1cb
 38c637d3a1763f6c3d32e8f1d979f045668676ec5feb8ee1869ee77cedd31b08"

 build_options="openmp"
@ -24,12 +25,21 @@ desc_option_openmp="Enable Open MP (gomp)"

 # Create a package for one specific language $1
 pkg_lang() {
-	local f lang=$1
-	vmkdir usr/share/tessdata
+	local f script lang=$1
+	case "$1" in
+		script-*)
+		    script=/script
+		    lang=${1#script-}
+		    ;;
+		*)  script=
+		    lang=$1
+		    ;;
+	esac
+	vmkdir usr/share/tessdata${script}
 	for f in $(find ${wrksrc}/tessdata -name "${lang}.*" \
-	-o -name "${lang}_frak.*" \
-	-o -name "${lang}_vert.*" ); do
-		vinstall $f 644 usr/share/tessdata
+	    -o -name "${lang}_frak.*" \
+	    -o -name "${lang}_vert.*" ); do
+		vinstall $f 644 usr/share/tessdata${script}
 		rm $f
 	done
 }
@ -41,9 +51,6 @@ post_extract() {
 pre_configure() {
 	NOCONFIGURE=1 ./autogen.sh
 }
-post_build() {
-	make ${makejobs} training
-}
 do_check() {
 	: # submodule not in tarball
 }
@ -125,7 +132,13 @@ tesseract-ocr-all_package() {
 		ita_old jav jpn kan kat kat_old kaz khm kir kor kur kur_ara lao lat lav lit ltz mal mar \
 		mkd mlt mon mri msa mya nep nld nor oci ori pan pol por que pus ron rus san sin slk slv \
 		snd spa spa_old sqi srp srp_latn sun swa swe syr tam tat tel tgk tgl tha tir ton tur \
-		uig ukr urd uzb uzb_cyrl vie yor yid; do
+		uig ukr urd uzb uzb_cyrl vie yid yor \
+		script-Arabic script-Armenian script-Bengali script-Canadian_Aboriginal script-Cherokee \
+		script-Cyrillic script-Devanagari script-Ethiopic script-Fraktur script-Georgian \
+		script-Greek script-Gujarati script-Gurmukhi script-HanS script-HanT script-Hangul \
+		script-Hebrew script-Japanese script-Kannada script-Khmer script-Lao script-Latin \
+		script-Malayalam script-Myanmar script-Oriya script-Sinhala script-Syriac script-Tamil \
+		script-Telugu script-Thaana script-Thai script-Tibetan script-Vietnamese; do
 		depends+=" tesseract-ocr-${lang}>=${version}_${revision}"
 	done
 }
@ -1097,3 +1110,267 @@ tesseract-ocr-yor_package() {
 		$(pkg_lang ${pkgname#tesseract-ocr-})
 	}
 }
+tesseract-ocr-script-Arabic_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Arabic script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Armenian_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Armenian script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Bengali_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Bengali script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Canadian_Aboriginal_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Canadian Aboriginal script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Cherokee_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Cherokee script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Cyrillic_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Cyrillic script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Devanagari_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Devanagari script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Ethiopic_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Ethiopic script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Fraktur_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Fraktur script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Georgian_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Georgian script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Greek_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Greek script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Gujarati_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Gujarati script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Gurmukhi_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Gurmukhi script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-HanS_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - HanS script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-HanT_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - HanT script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Hangul_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Hangul script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Hebrew_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Hebrew script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Japanese_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Japanese script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Kannada_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Kannada script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Khmer_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Khmer script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Lao_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Lao script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Latin_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Latin script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Malayalam_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Malayalam script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Myanmar_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Myanmar script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Oriya_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Oriya script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Sinhala_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Sinhala script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Syriac_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Syriac script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Tamil_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Tamil script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Telugu_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Telugu script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Thaana_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Thaana script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Thai_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Thai script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Tibetan_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Tibetan script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}
+tesseract-ocr-script-Vietnamese_package() {
+	archs=noarch
+	depends="${sourcepkg}>=${version}_${revision}"
+	short_desc+=" - Vietnamese script data"
+	pkg_install() {
+		$(pkg_lang ${pkgname#tesseract-ocr-})
+	}
+}