Revision history for Perl extension Lingua::YaTeA.
0.626
- Remove of File::chmod, dependency in the install script (Build.PL)
0.625
- Change in the customization of the install script (Build.PL)
0.623
- Homogenisation of the output management in the print* functions (PhraseSet)
- Change of the minimal required Perl version
0.622
- Correction of the input if the format is not correct (either
incorrect lines are removed, either they are corrected)
- Correction in the processing of the testified terms (if the
option "TT-for-BioLG" is not set, yatea is usuable).
- Add a module dependency for the configuration
- Take into account the prefix for the configuration file
0.621
- Add the installation of the FR-Flemm files (which have been
forgot before:-(
- Add pod directive "=encoding utf8" to manage utf8 encoding
0.62
- better management of argument in the script yatea
- Correction in utf8 use directive to avoid failure with CPAN
tests
- correction deprecated code in IndexSet (defined @array)
0.61
- Add use utf8 in the script yatea to avoid failure with CPAN
tests
- Minor correction in the documentation of the main module
0.6
- Add of installation tests
- Add of examples for French
- Integration of the MIG corrections (Node, Testified terms)
- Correction in the forbidden structure management (split
action)
- Add NEXT.pm in the pre-required module list
- UTF-8 is used as charset for both configuration files and
texts to process
- Integration of attested terms (in a bracketed format, can be
produced by the bootstrap option)
- Option "bootstrap" to generate an output in a bracketed
format. This output can be used as a attested resource on a
other corpus
- Add a defined test in the function appendInclude (but it
seems there is still a bug on tree building).
- Monolexical term occurrences are identified as MNP when the
monolexical oprion is set
- Addition of a config set for French based on the POSTagger
Flemm (which uses the Multex tagset) and modification of the
read method in Corpus.pm (addition opf the parameter
language) to normalisze the input (correction of the output
of Flemm)
- Correction in the input normalization for Flemm
- Each printing function can now print results on stdout
- Replace ':' as tag by 'COLUMN'
- Add of some Chunking Frontiers as '('
- Only the tag is taken into account for sentence_boundary
detection
- Add Weights section to the DTD and in the XML rendering
- Correction: TF-IDF based ranking method is a DDW ranking
- Add ROOT as a reference to the term in which the current
term is nested
- Addition of several term weighting and selection measures
(C-Value and variations, iLong, ilnc, iLong, and term
autonomy)
- Add the status of the term candidate (0 : not a term, 1 :
term). By default, term candidate are terms
- Add a option for term output style to print only term
candidate having the status of term
- colors of the terms in the HTML output can be parametrized
(option file, options PARSED_COLOR and UNPARSED_COLOR)
- new function for printing only list of candidate terms
without XML header
- Addition of the option XML-corpus-raw, rendering the corpus
in XML format with terms. Documents and sentences are
identified in the XML files.
0.5
- Minor correction in the installation process
- for some OS, Config::General returns while setting variables
as yatea.rc is on read-inly mode
0.4
- use warnings in all the files
- correction of minor bugs
- control of the input type in Node.pm, NodeSet.pm, Tree.pm
- bug fix in Tree.pm (method
completeDiscontinuousNodes). Initialisation of
a variable
- carriage returns are deleted in the loaded files (module
ForbiddenStructureSet.pm, method CleanInfo, and module
ChunkingDataSet.pm, method LoadData,Corpus.pm:WrapBlocks)
- addition of the option "TC_MAXIMUM_LENGTH" to allow to limit
the length of the extracted term candidates (set in
configuration file Options, default is 12 words)
- homogeneisation of sub-trees attachement
- addition of a TF-IDF based ranking method for extracted term
candidates
- Workaround for a missing feature in Module::Info : if no
prefix is set while running "perl Build.PL", additionnal
directories 'etc' and 'conf' are not installed (or well installed)
- Modification of Makefile.PL to use the Module::Build (it
should be a CPAN.pm-compatible distribution)
- Variables CONFIG_REP, LOCALE_REP, RESULT_REP (yatea.rc) are
renamed CONFIG_DIR, LOCALE_DIR, RESULT_DIR respectively.
- Variables CONFIG_DIR, LOCALE_DIR, RESULT_DIR (yatea.rc) are
automatically set, while the configuration (in Build.PL)
- The location of the rc file (yatea.rc) is automatically set,
while the configuration (in Build.PL)
- documentation of the module TagSet.pm, File.pm and FileSet.pm
- Corrections in the LICENSE file
0.3
- removing the .config file
- bug fix: check in the hook_node is defined in Node.pm
- replacement of the use of the Parse::Lex module by a
module generated from Parse::Yapp
- addition of examples in the YaTeA PM
- correction of a bug when calculating the offset of the end
char in component of term (especially the left components)
- in parseProgressively, taking into account that the unparsed
array can be empty
0.2
- correction of bugs (or temporary workaround) in the
tree/node management
- Removing the file Config.pm (no more necessary)
and from Makefile.PL
- YaTeA.pm has been moved in Lingua (before it was in Lingua/YaTeA)
- Addition of a Build.PL file
- addition of POD tests (pod.t, pod-coverage.t)
0.1
- Version of YaTeA with a CPAN configuration