tesseract:moved from contrib and version bump

This commit is contained in:
blue-devil
2020-01-19 09:13:29 +03:00
parent 7686deaae8
commit bca11cd669
3 changed files with 156 additions and 0 deletions
+26
View File
@@ -0,0 +1,26 @@
#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# Licensed under the GNU General Public License, version 3.
# See the file http://www.gnu.org/licenses/gpl.txt
from pisi.actionsapi import shelltools
from pisi.actionsapi import autotools
from pisi.actionsapi import pisitools
from pisi.actionsapi import get
def setup():
shelltools.system("./autogen.sh")
autotools.configure("--disable-static")
# fıx unused dırect dependency analysis
pisitools.dosed("libtool", " -shared ", " -Wl,-O1,--as-needed -shared ")
def build():
autotools.make()
autotools.make("training")
def install():
autotools.rawInstall("DESTDIR=%s" % get.installDIR())
autotools.make("training-install DESTDIR=%s" % get.installDIR())
pisitools.dodoc("LICENSE", "README*")
+110
View File
@@ -0,0 +1,110 @@
<?xml version="1.0" ?>
<!DOCTYPE PISI SYSTEM "https://www.pisilinux.org/projeler/pisi/pisi-spec.dtd">
<PISI>
<Source>
<Name>tesseract</Name>
<Homepage>https://github.com/tesseract-ocr/tesseract</Homepage>
<Packager>
<Name>PisiLinux Community</Name>
<Email>admins@pisilinux.org</Email>
</Packager>
<PartOf>office.misc</PartOf>
<License>APACHE</License>
<IsA>app:console</IsA>
<Summary>Tesseract OCR Engine</Summary>
<Description>This package contains an OCR engine - libtesseract and a command line program - tesseract. Tesseract 4 adds a new neural net (LSTM) based OCR engine which is focused on line recognition, but also still supports the legacy Tesseract OCR engine of Tess</Description>
<Archive sha1sum="25318bb3f57ef72d5736730739451673f4a66f51" type="targz">https://github.com/tesseract-ocr/tesseract/archive/4.1.1.tar.gz</Archive>
<BuildDependencies>
<Dependency>leptonica-devel</Dependency>
<Dependency>cairo-devel</Dependency>
<Dependency>icu4c-devel</Dependency>
<Dependency>pango-devel</Dependency>
<Dependency>tiff-devel</Dependency>
<Dependency>libgomp</Dependency>
<!-- dependecies for documentation -->
<Dependency>libxslt-devel</Dependency>
<Dependency>asciidoc</Dependency>
</BuildDependencies>
</Source>
<Package>
<Name>tesseract</Name>
<Summary>Tesseract OCR Engine</Summary>
<RuntimeDependencies>
<Dependency>leptonica</Dependency>
<Dependency>libgomp</Dependency>
<Dependency>tiff</Dependency>
</RuntimeDependencies>
<Files>
<Path fileType="library">/usr/lib/libtesseract.so.4.0.1</Path>
<Path fileType="library">/usr/lib/libtesseract.so.4</Path>
<Path fileType="library">/usr/lib/libtesseract.so</Path>
<Path fileType="executable">/usr/bin/tesseract</Path>
<Path fileType="data">/usr/share/tessdata</Path>
<Path fileType="doc">/usr/share/doc</Path>
<Path fileType="data">/usr/share</Path>
</Files>
</Package>
<Package>
<Name>tesseract-devel</Name>
<Summary>Development files for tesseract</Summary>
<RuntimeDependencies>
<Dependency release="current">tesseract</Dependency>
</RuntimeDependencies>
<Files>
<Path fileType="library">/usr/lib/pkgconfig/tesseract.pc</Path>
<Path fileType="header">/usr/include/tesseract</Path>
</Files>
</Package>
<Package>
<Name>tesseract-tools</Name>
<Summary>Tools for tesseract</Summary>
<RuntimeDependencies>
<Dependency release="current">tesseract</Dependency>
<Dependency>fontconfig</Dependency>
<Dependency>leptonica</Dependency>
<Dependency>cairo</Dependency>
<Dependency>icu4c</Dependency>
<Dependency>pango</Dependency>
</RuntimeDependencies>
<Files>
<Path fileType="executable">/usr/bin/set_unicharset_properties</Path>
<Path fileType="executable">/usr/bin/language-specific.sh</Path>
<Path fileType="executable">/usr/bin/unicharset_extractor</Path>
<Path fileType="executable">/usr/bin/tesstrain_utils.sh</Path>
<Path fileType="executable">/usr/bin/combine_lang_model</Path>
<Path fileType="executable">/usr/bin/classifier_tester</Path>
<Path fileType="executable">/usr/bin/merge_unicharsets</Path>
<Path fileType="executable">/usr/bin/combine_tessdata</Path>
<Path fileType="executable">/usr/bin/shapeclustering</Path>
<Path fileType="executable">/usr/bin/ambiguous_words</Path>
<Path fileType="executable">/usr/bin/dawg2wordlist</Path>
<Path fileType="executable">/usr/bin/wordlist2dawg</Path>
<Path fileType="executable">/usr/bin/tesstrain.sh</Path>
<Path fileType="executable">/usr/bin/lstmtraining</Path>
<Path fileType="executable">/usr/bin/cntraining</Path>
<Path fileType="executable">/usr/bin/mftraining</Path>
<Path fileType="executable">/usr/bin/text2image</Path>
<Path fileType="executable">/usr/bin/lstmeval</Path>
</Files>
</Package>
<History>
<Update release="2">
<Date>2020-01-19</Date>
<Version>4.1.1</Version>
<Comment>Version bump.</Comment>
<Name>Blue Devil</Name>
<Email>bluedevil@sctzine.com</Email>
</Update>
<Update release="1">
<Date>2019-01-25</Date>
<Version>4.0.0</Version>
<Comment>First release</Comment>
<Name>PisiLinux Community</Name>
<Email>admins@pisilinux.org</Email>
</Update>
</History>
</PISI>
+20
View File
@@ -0,0 +1,20 @@
<?xml version="1.0" ?>
<PISI>
<Source>
<Name>tesseract</Name>
<Summary xml:lang="tr">Tesseract OCR Motoru</Summary>
<Description xml:lang="tr">tesseract, en kararlı açık kaynaklı OCR (Optik Karakter Tanıma) motorlarından biridir. Uygulama girdi olarak aldığı çeşitli resim biçimlerini analiz ederek metin biçiminde çıktı üretir.</Description>
</Source>
<Package>
<Name>tesseract-devel</Name>
<Summary xml:lang="tr">tesseract için geliştirme dosyaları</Summary>
<Description xml:lang="tr">tesseract-devel, tesseract için geliştirme dosyaları içerir.</Description>
</Package>
<Package>
<Name>tesseract-tools</Name>
<Summary xml:lang="tr">tesseract için araçlar</Summary>
<Description xml:lang="tr">tesseract-tools, tesseract için araçlar içerir.</Description>
</Package>
</PISI>