* v0.3.4 - Mon Dec 17 17:08:09 CET 2012
- fixed several issues with installation routines
- added config files for preparing OPUS data
- changed handling of Uplug environment variables and standard locations
- fixed call to Grok tagger/chunker
- fixed word alignment test
- fixed POS-tagging for Swedish with Grok
* v0.3.3 - Mon Dec 10 19:14:00 CET 2012
- fixed some problems with documentation
- prepared language packs and Uplug add-ons
- added documentation for language packs
* v0.3.2 - Fri Dec 7 20:33:14 CET 2012
- added documentation
- global installation
- makefile for generating documentation in different formats
* v0.3.1 - Thu Dec 6 20:33:14 CET 2012
- install now compiles external tools on systems without pre-compiled binaries
- improved uplug-treetagger package (automatic download/install)
- added flag for fallback modules
- added test if modules exist to uplug startup script
- added help information to uplug via command line flags
- added documentation to uplug startup script
- list available modules
- fixed problem with uplug-giza
* v0.3 - Mon Nov 26 11:57:26 CET 2012
- re-arranging modules, scripts, config-files
- include installation routines
- divide Uplug in several packages (language packs etc)
- integrate MaltParser and pre-trained models
- new pre-compiled binaries (also for Mac OSX)
- improved pre-processing pipelines for several languages
- pre-trained hunpos models
- integration of Europarl/Moses sentence splitter/tokenizer
- (some) unit tests
- moved to git and bitbucket
- added Moses-style word alignment format for clue alignment output
* unversioned
- toktag.pl: better handling of already tokenized documents
- added wrapper for alpino (dutch) tokenizer)
- tools/subtitles/srt2xml: handling of formating tags (<i> etc) is fixed!
- tools/subtitles/srtalign.pl: added check for illegal subtitle
synchronization scaling factor (<=0) and made it possible to use cognates
and dictionaries together for automatic synchronization
- ICA: show only non-empty sentence alignments (x:0 sentence alignments don't
make sense for word alignment), changed in include/xmldoc.inc
- added -adj to all word alignment config files for the 'link' step
* v0.2.0c and v0.2.0d
- added scripts to tools/subtitles to handle (pre-process and sentence align)
movie subtitle in srt format
- robust conversion of encodings in tag.pl/toktag.pl/chunk.pl
- added treetagger startup scripts for es and nl, replace "nbsp" to " "
- robust conversion between encodings in bitext-indexer.pl/opus-indexer.pl
- added startup scripts for spanish and dutch tree-tagger models
- updated startup scripts for other treetagger models according to latest
TreeTagger distribution
- fixed hunalign (bug in converting alignment output to xml)
- added missing ';' at line 40 in Uplug.pm
* 13-10-2006 joerg
- ISA & ICA with simple user management (multiple users, password protection)
- integrated hunalign sentence aligner
- integrated GMA sentence aligner
- alternative sentence aligner with cognate filter: uplugalign
* 26-05-2006 joerg
- improved ISA & ICA (some fixes, allow empty alignments in ISA ...)
- bin/wordalign.pl can bee run in server mode (hard-coded port, localhost) to
serve ICA
* 16-03-2006 joerg
- nothing really new, serveral bug-fixes done since last time
- ISA & ICA - two web interfaces for interactive alignment (in web/php)
* 28-07-2005 joerg
- fixed a bug in learning bootstrapped clues with linkclue.pl
(target language feature strings were not produced correctly)
- took away Makefile.PL (isn't used anyway and seems to confuse)
(there should be a proper installation script but there is none ...)
* July 2005 joerg
- added web interfaces for interactive alignment (web/php)
- fixed some bugs in word alignment (problems with bitext segments
other than 1:1 sentence)
* 21-06-2005 joerg
- fixed some stupid bugs in bin/markup.pl
* 20-06-2005 joerg
- fixed some more bugs in word alignment configuration files
- added subtree indexing for XML files (enables jumping to certain positions
in the corpus)
- added experimental align.php file for alignment visualization
* 24-03-2005 joerg
- found&fixed some bugs in systems/align/word/link and
systems/align/word/test/link (shortcut definitions for dynamic clue
weights)
- fixed a bug in Uplug/Config.pm in CheckParam
* 03-03-2005 joerg
- re-implementation of Uplug.pm, Uplug/Config.pm, uplug
- some minor changes in Uplug/IO* (mainly cleaning up, it's still messy though)
- some minor changes in the configuration files for pre-processing and
* 15-02-2005 joerg
- new version of ini/DataStreams.ini
- new/changed configuration files for word-alignment
(basic, default, tagged, advanced)
* 15-02-2005 lars
- started manual
- started installation script
- changed Uplug.pm to call converter.pl from $UPLUGHOME
* 26-10-2004 joerg
- Grok source code files removed from Uplug release
- added Giza.pm for running GIZA++
* 15-02-2005 joerg
- giza.pl creates 2 clue-dbms when running in both alignment directions
- standard word alignment settings: basic, default, tagged, advanced
- some minor bug-fixes