Changes for version v2.0.11 - 2016-02-22
- use new Alien::Moot for c++ library dependency
Changes for version v2.0.10_006 - 2015-09-25
- renamed distribution to 'Moot' for CPAN-friendliness
Changes for version v2.0.10_005 - 2014-04-16
- added moot-scan.perl : moot TokenReader debugging
- updated to 2.0.10_005 to jive with C++ package
Changes for version v2.0.10_004 - 2013-12-18
- TokenReader.pm,TokenWriter.pm: moot/perl fixes
- updated to 2.0.10_004 to stay in sync with trunk
Changes for version v2.0.10_003 - 2013-12-06
- waste: improved handling of negative mode selectors (e.g. -N)
- updated perl bindings
- added Moot::Waste::Annotator class
- updated Moot::TokPP to use Moot::Waste::Annotator
Changes for version v2.0.10_002 - 2013-12-02
- fixed wasteScanner choking on long utf-8-encoded characters (e.g. U+1D1A3 : MUSICAL SYMBOL ORNAMENT STROKE-9 : \xf0\x9d\x86\xa3 in bach_versuch02_1762
- wasteScanner should now handle even non-utf8 more or less gracefully
Changes for version v2.0.10_001 - 2013-11-28
- v2.0.10-1: workaround for probability underflow error propagation in mootHMM::tag_stream()
- once underflowed, no more differentiation was made, since no nodes qualified as flushable until EOF
- workaround flushes nodes whenever 'unsafe' probabilities (<-1e37) are encountered
- encoding tweaks for Moot::TokPP::analyze_buffer()
- tokpp improvements / fixes
- fixed to jive with kmw's wasteLexer changes
- wasteTrainWriter: basically working, but links are being dropped (scanner bug)
- waste training prototype in testme.perl
- added Moot::TokPP, moot-tokpp.perl : drop-in replacement for dwds_tomasotath tokenizer-supplied pseudo-morphology
- documented Waste::Lexer::dehyphenate()
- make distcheck fixes
- got Moot::Waste::Decoder working, including buffer-level access
- added Waste::Decoder to perl
- Waste::Lexer seems working
- including get/set on underlying scanner, using lexer->tr_data to hold an SV
- removed WasteLexerPerl class
- was WIP for simultaneous support of both standalone and embedded wasteLexicon objects, now abandoned
- Waste::Lexicon : now only accessible via Waste::Lexer
- avoids ref-counting madness for embedded objects
- added TokenReader, TokenWriter hierarchy wrappers
- WIP on wasteLexer, wasteLexicon
- wrapped wasteTokenScanner as Moot::Waste::Scanner
- added scanner,lexer type constants (why? they're not actually _used_ ... we should probably remove them again)
- wrapper uses PerlIO layer
- TokenReader bugfixes (check for null tr_istream in from_filename()
Changes for version v2.0.9_002 - 2013-10-22
- added re2c_ucl.py (re2c char-class generator)
- added wasteScannerScan.* templates for waste generation
- added moot(lookup|merge)-(lex|123).perl to MANIFEST
- added mootlookup-lex.perl
- fixes for weird DynaLoader bug on perl v5.14 / 32-bit i686 / debian wheezy if CCFLAGS is set in Makefile.PL
- strangely, x86_64 machine was unaffected
- bad: Linux plato 3.2.0-4-686-pae #1 SMP Debian 3.2.41-2 i686 GNU/Linux
- added command-line utils mootmerge*.perl
- updated version for 2.0.9-2
Changes for version v2.0.9_001 - 2012-03-19
- updated perl wrappers
Modules
Perl interface to the libmoot part-of-speech tagging library
libmoot : constants
libmoot : HMM
libmoot : HMM : dynamic : lexical probabilites via Boltzmann distribution
libmoot : HMM : dynamic
libmoot : HMM : dynamic : lexical probabilites
libmoot : lexical frequencies
libmoot : n-gram frequencies
libmoot : heuristic token analyzer (pseudo-morphology, wraps for Moot::Waste::Annotator)
libmoot : Token I/O
libmoot: Token I/O: reader
libmoot: Token I/O: reader: native 1 word/line format
libmoot: Token I/O: reader: built-in XML format
libmoot: Token I/O: writer
libmoot: Token I/O: writer: native 1 word/line format
libmoot: Token I/O: writer: built-in XML format
libmoot : WASTE tokenization system
libmoot : WASTE tokenizer : pattern-based token annotator
libmoot : WASTE tokenizer : post-Viterbi decoder
libmoot : WASTE tokenizer : mid-level lexer
libmoot : WASTE tokenizer : simple word-list lexicon
libmoot : WASTE tokenizer : low-level scanner