File tesseract-ocr-traineddata.spec of Package tesseract-ocr-traineddata

#
# spec file for package tesseract-ocr-traineddata
#
# Copyright (c) 2021 SUSE LLC
#
# All modifications and additions to the file contributed by third parties
# remain the property of their copyright owners, unless otherwise agreed
# upon. The license for this file, and modifications and additions to the
# file, is the same license as for the pristine package itself (unless the
# license for the pristine package is not an Open Source License, in which
# case the license is the MIT License). An "Open Source License" is a
# license that conforms to the Open Source Definition (Version 1.9)
# published by the Open Source Initiative.

# Please submit bugfixes or comments via https://bugs.opensuse.org/
#


Name:           tesseract-ocr-traineddata
Version:        4.1.0+git.20191030.6572757
Release:        0
Summary:        Tesseract Traineddata for Various Languages
License:        Apache-2.0
URL:            https://github.com/tesseract-ocr
Source0:        tessdata_fast-%{version}.tar.xz
BuildRoot:      %{_tmppath}/%{name}-%{version}-build
BuildArch:      noarch
BuildRequires:  python3-packaging

%description
This package contains fast integer versions of trained models for the Tesseract
Open Source OCR Engine.

These models only work with the LSTM OCR engine of Tesseract 4.

%package        doc
Summary:        Documentation for %{name}

%description    doc
The %{name}-doc package contains the documentation for %{name}.

%package        osd
Summary:        Orientation & Script Detection Data for tesseract
BuildArch:      noarch
Requires:       tesseract-ocr >= %{version}

%description    osd
Orientation & Script Detection data for the Tesseract Open Source OCR Engine.

# define lang_subpkg macro
# m: 3 letter macrolanguage code
# n: language name
# -m and -n is needed for subpackages
#
%define lang_subpkg(l:m:n:) \
%define macrolang %{-m:%{-m*}}%{!-m:%{error:3 letter Language code not defined}} \
%define langname %{-n:%{-n*}}%{!-n:%{error:Language name not defined}} \
\
%package       %{macrolang}\
Summary:       %{langname} language data for %{name}\
BuildArch:     noarch\
Requires:      tesseract-ocr >= %{version}\
\
%description   %{macrolang}\
This package contains the fast integer version of the %{langname} language \
trained models for the Tesseract Open Source OCR Engine.\
\
%files %{macrolang}\
%dir %{_datadir}/tessdata\
%{_datadir}/tessdata/%{macrolang}.*

# define script_subpkg macro
# s: script name
# n: package name
#
%define script_subpkg(s:n:) \
%define scriptname %{-s:%{-s*}}%{!-s:%{error:Script name defined}} \
%define filename %{-n:%{-n*}}%{!-n:%{error:Package name not defined}} \
%define pkgname %(echo %filename | tr '[:upper:]' '[:lower:]') \
\
%package       script-%{pkgname}\
Summary:       %{scriptname} script data for %{name}\
BuildArch:     noarch\
Requires:      tesseract-ocr >= %{version}\
\
%description   script-%{pkgname}\
This package contains the fast integer version of the %{scriptname} script \
trained models for the Tesseract Open Source OCR Engine.\
\
%files script-%{pkgname}\
%dir %{_datadir}/tessdata/\
%dir %{_datadir}/tessdata/script/\
%{_datadir}/tessdata/script/%{filename}.*

# see https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes
# and https://en.wikipedia.org/wiki/List_of_ISO_639-2_codes
%lang_subpkg -m afr -n Afrikaans
%lang_subpkg -m amh -n Amharic
%lang_subpkg -m ara -n Arabic
%lang_subpkg -m asm -n Assamese
%lang_subpkg -m aze -n Azerbaijani
%lang_subpkg -m aze_cyrl -n %{quote:Azerbaijani (Cyrillic)}
%lang_subpkg -m bel -n Belarusian
%lang_subpkg -m ben -n Bengali
%lang_subpkg -m bod -n %{quote:Tibetan (Standard)}
%lang_subpkg -m bos -n Bosnian
%lang_subpkg -m bre -n Breton
%lang_subpkg -m bul -n Bulgarian
%lang_subpkg -m cat -n Catalan
%lang_subpkg -m ceb -n Cebuano
%lang_subpkg -m ces -n Czech
%lang_subpkg -m chi_sim -n %{quote:Chinese (Simplified)}
%lang_subpkg -m chi_sim_vert -n %{quote:Chinese (Simplified, Vertical)}
%lang_subpkg -m chi_tra -n %{quote:Chinese (Traditional)}
%lang_subpkg -m chi_tra_vert -n %{quote:Chinese (Traditional, Vertical)}
%lang_subpkg -m chr -n Cherokee
%lang_subpkg -m cos -n Corsican
%lang_subpkg -m cym -n Welsh
%lang_subpkg -m dan -n Danish
%lang_subpkg -m deu -n German
%lang_subpkg -m div -n %{quote:Dhivehi; Maldivian}
%lang_subpkg -m dzo -n Dzongkha
%lang_subpkg -m ell -n Greek
%lang_subpkg -m eng -n English
%lang_subpkg -m enm -n %{quote:Middle English (1100-1500)}
%lang_subpkg -m epo -n Esperanto
%lang_subpkg -m est -n Estonian
%lang_subpkg -m eus -n Basque
%lang_subpkg -m fao -n %{quote:Faroese}
%lang_subpkg -m fas -n %{quote:Persian (Farsi)}
%lang_subpkg -m fil -n %{quote:Filipino; Pilipino}
%lang_subpkg -m fin -n Finnish
%lang_subpkg -m fra -n French
%lang_subpkg -m frk -n Fraktur
%lang_subpkg -m frm -n %{quote:Middle French (ca. 1400-1600)}
%lang_subpkg -m fry -n %{quote:Western Frisian}
%lang_subpkg -m gla -n %{quote:Gaelic; Scottish Gaelic}
%lang_subpkg -m gle -n Irish
%lang_subpkg -m glg -n Galician
%lang_subpkg -m grc -n %{quote:Ancient Greek}
%lang_subpkg -m guj -n Gujarati
%lang_subpkg -m hat -n Haitian
%lang_subpkg -m heb -n Hebrew
%lang_subpkg -m hin -n Hindi
%lang_subpkg -m hrv -n Croatian
%lang_subpkg -m hun -n Hungarian
%lang_subpkg -m hye -n Armenian
%lang_subpkg -m iku -n Inuktitut
%lang_subpkg -m ind -n Indonesian
%lang_subpkg -m isl -n Icelandic
%lang_subpkg -m ita -n Italian
%lang_subpkg -m ita_old -n %{quote:Italian (Old)}
%lang_subpkg -m jav -n Javanese
%lang_subpkg -m jpn -n Japanese
%lang_subpkg -m jpn_vert -n "Japanese (Vertical)"
%lang_subpkg -m kan -n Kannada
%lang_subpkg -m kat -n Georgian
%lang_subpkg -m kat_old -n %{quote:Georgian (Old)}
%lang_subpkg -m kaz -n Kazakh
%lang_subpkg -m khm -n Khmer
%lang_subpkg -m kir -n Kyrgyz
%lang_subpkg -m kor -n Korean
%lang_subpkg -m kor_vert -n "Korean (Vertical)"
%lang_subpkg -m kmr -n Kurmanji
%lang_subpkg -m lao -n Lao
%lang_subpkg -m lat -n Latin
%lang_subpkg -m lav -n Latvian
%lang_subpkg -m lit -n Lithuanian
%lang_subpkg -m ltz -n Luxembourgish
%lang_subpkg -m mal -n Malayalam
%lang_subpkg -m mar -n Marathi
%lang_subpkg -m mkd -n Macedonian
%lang_subpkg -m mlt -n Maltese
%lang_subpkg -m mon -n Mongolian
%lang_subpkg -m mri -n Maori
%lang_subpkg -m msa -n Malay
%lang_subpkg -m mya -n Burmese
%lang_subpkg -m nep -n Nepali
%lang_subpkg -m nld -n Dutch
%lang_subpkg -m nor -n Norwegian
%lang_subpkg -m oci -n Occitan
%lang_subpkg -m ori -n Oriya
%lang_subpkg -m pan -n Panjabi
%lang_subpkg -m pol -n Polish
%lang_subpkg -m por -n Portuguese
%lang_subpkg -m pus -n Pashto
%lang_subpkg -m que -n Quechuan
%lang_subpkg -m ron -n Romanian
%lang_subpkg -m rus -n Russian
%lang_subpkg -m san -n Sanskrit
%lang_subpkg -m sin -n Sinhala
%lang_subpkg -m slk -n Slovakian
%lang_subpkg -m slv -n Slovenian
%lang_subpkg -m snd -n Sindhi
%lang_subpkg -m spa -n Spanish
%lang_subpkg -m spa_old -n %{quote:Spanish (Old)}
%lang_subpkg -m sqi -n Albanian
%lang_subpkg -m srp -n Serbian
%lang_subpkg -m srp_latn -n %{quote:Serbian (Latin)}
%lang_subpkg -m sun -n Sundanese
%lang_subpkg -m swa -n Swahili
%lang_subpkg -m swe -n Swedish
%lang_subpkg -m syr -n Syriac
%lang_subpkg -m tam -n Tamil
%lang_subpkg -m tat -n Tatar
%lang_subpkg -m tel -n Telugu
%lang_subpkg -m tgk -n Tajik
%lang_subpkg -m tha -n Thai
%lang_subpkg -m tir -n Tigrinya
%lang_subpkg -m ton -n Tongan
%lang_subpkg -m tur -n Turkish
%lang_subpkg -m uig -n Uyghur
%lang_subpkg -m ukr -n Ukrainian
%lang_subpkg -m urd -n Urdu
%lang_subpkg -m uzb -n Uzbek
%lang_subpkg -m uzb_cyrl -n %{quote:Uzbek (Cyrillic)}
%lang_subpkg -m vie -n Vietnamese
%lang_subpkg -m yid -n Yiddish
%lang_subpkg -m yor -n Yoruba

%script_subpkg -n Arabic -s Arabic
%script_subpkg -n Armenian -s Armenian
%script_subpkg -n Bengali -s Bengali
%script_subpkg -n Canadian_Aboriginal -s %{quote:Canadian (Aboriginal)}
%script_subpkg -n Cherokee -s Cherokee
%script_subpkg -n Cyrillic -s Cyrillic
%script_subpkg -n Devanagari -s Devanagari
%script_subpkg -n Ethiopic -s Ethiopic
%script_subpkg -n Fraktur -s Fraktur
%script_subpkg -n Georgian -s Georgian
%script_subpkg -n Greek -s Greek
%script_subpkg -n Gujarati -s Gujarati
%script_subpkg -n Gurmukhi -s Gurmukhi
%script_subpkg -n HanS -s %{quote:Han (Simplified)}
%script_subpkg -n HanS_vert -s %{quote:Han (Simplified, Vertical)}
%script_subpkg -n HanT -s %{quote:Han (Traditional)}
%script_subpkg -n HanT_vert -s %{quote:Han (Traditional, Vertical)}
%script_subpkg -n Hangul -s Hangul
%script_subpkg -n Hangul_vert -s %{quote:Hangul (Vertical)}
%script_subpkg -n Hebrew -s Hebrew
%script_subpkg -n Japanese -s Japanese
%script_subpkg -n Japanese_vert -s %{quote:Japanese (Vertical)}
%script_subpkg -n Kannada -s Kannada
%script_subpkg -n Khmer -s Khmer
%script_subpkg -n Lao -s Lao
%script_subpkg -n Latin -s Latin
%script_subpkg -n Malayalam -s Malayalam
%script_subpkg -n Myanmar -s Myanmar
%script_subpkg -n Oriya -s Oriya
%script_subpkg -n Sinhala -s Sinhala
%script_subpkg -n Syriac -s Syriac
%script_subpkg -n Tamil -s Tamil
%script_subpkg -n Telugu -s Telugu
%script_subpkg -n Thaana -s Thaana
%script_subpkg -n Thai -s Thai
%script_subpkg -n Tibetan -s Tibetan
%script_subpkg -n Vietnamese -s Vietnamese

%prep
%autosetup -n tessdata_fast-%{version}

%install
mkdir -p %{buildroot}/%{_datadir}/tessdata/
cp -a *.traineddata %{buildroot}/%{_datadir}/tessdata/
cp -r script %{buildroot}/%{_datadir}/tessdata

%files doc
%license LICENSE
%doc README.md

%files osd
%{_datadir}/tessdata/equ.traineddata
%{_datadir}/tessdata/osd.traineddata

%changelog