Suomi-malaga - Voikko edition ============================= General information ------------------- Suomi-malaga is a description of Finnish morphology written in Malaga (http://home.arcor.de/bjoern-beutel/malaga/). You should use malaga version 7.8 or later. Currently Suomi-malaga is used in two different applications: text indexer Sukija and spellchecker/hyphenator Voikko. Version 1.0 and later will work with both applications. This release creates Voikko morphology with version 2 dictionary format. All of the documentation about Finnish morphology is in Finnish (see README.fi and subdirectory doc). This README contains only build and usage instructions for distribution packagers. Build and installation ---------------------- Building Suomi-malaga for from this package requires malaga, python and make. No configuration is required: to build the code for Voikko, you only need to run make voikko Installation can be done by running make voikko-install DESTDIR=/usr/lib/voikko (Replace /usr/lib/voikko with the directory you want to install the files to. Installing to ~/.voikko will cause libvoikko to use this version of Suomi-malaga only for the user who does the installation.) Building the code for Sukija can be done by running make sukija Supported Make targets ---------------------- - sukija Builds the binary files needed by text indexer Sukija. - voikko Builds the binary files needed by libvoikko. - voikko-sukija Builds the binary files needed by libvoikko with a big dictionary that can be used by Sukija. - voikko-install DESTDIR=/usr/lib/voikko Installs the binary files needed by libvoikko to the directory specified by DESTDIR. DESTDIR is optional and defaults to /usr/lib/voikko - dist-gzip Builds the full source package. - clean Removes all files generated by other targets. - update-vocabulary Updates the XML vocabulary from the nightly snapshot at joukahainen.puimula.org. This target requires wget to be available. - TAGS Builds an Emacs tag table file from the vocabulary database. Variables for tuning the build process -------------------------------------- - make voikko: * VOIKKO_BUILDDIR=path/to/directory Specifies the directory where build files are written to while building for Voikko. Default: voikko (build within source directory) * GENLEX_OPTS="--option1=xxx --option2=yyy ..." Sets options string for generate_lex.py. Available options for generate_lex.py are + --min-frequency=n Limits the words to be included in the .lex files to the specified or higher frequency class. Default is 9. + --extra-usage=usage1,usage2,... If a word has usage flags (it belongs to a special vocabulary), it is included in the vocabulary only if at least one of the usage flags is listed here. Available usage flags are listed in file vocabulary/flags.txt. Listing "sukija" here causes application specific exclusions to be ignored (words marked with not_voikko will also be included). By default, no special vocabularies are included. + --style=style1,style2,... If a word has style flags (such as old, foreign or dialect), it is included in the vocabulary only if all of the style flags are listed here. Available style flags are listed in file vocabulary/flags.txt. Default: old,international,inappropriate + --sourceid Insert word identifiers from Joukahainen to lexicon and return them during morphological analysis. This option has no effect unless VOIKKO_DEBUG=yes is set. By default source ids are not preserved. * EXTRA_LEX="path/to/file1.lex path/to/file2.lex ..." Adds extra malaga lexicon files to the vocabulary. By default, no extra lexicons are added. * VANHAHKOT_MUODOT=yes|no See voikko/doc/liput.txt. Default: yes * VANHAT_MUODOT=yes|no See voikko/doc/liput.txt. Default: no * SUKIJAN_MUODOT=yes|no Include words that exist just for Sukija. Default: no * VOIKKO_DEBUG=yes|no Include information that is not needed by libvoikko but may be needed for debugging or by external applications (full morphological analysis). Default: no * VOIKKO_VARIANT=variant Set the short name for the language variant of this vocabulary. The name should match the regular expression [a-z][a-z0-9_]* Default: standard * VOIKKO_DESCRIPTION="Description of the vocabulary" Set the long description for the language variant of this vocabulary. * SM_PATCHINFO="Information about applied patches" If you have modified the source code or are distributing prerelease versions, describe any modifications made to the released version here. It may be best to change this directly in the Makefile. Copyright and license information --------------------------------- This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. See file COPYING for details. Copyright (©) 2006 - 2015 Hannu Väisänen (Email: Hannu.Vaisanen@uef.fi) and 2006 - 2015 Harri Pitkänen (email@example.com). Contributors listed in file CONTRIBUTORS hold copyrights to the vocabulary data.