Revision: 16356 http://gate.svn.sourceforge.net/gate/?rev=16356&view=rev Author: ian_roberts Date: 2012-11-29 12:57:45 +0000 (Thu, 29 Nov 2012) Log Message: ----------- Consolidating changelog.
Modified Paths: -------------- userguide/branches/release-7.1/recent-changes.tex userguide/branches/release-7.1/tao_main.tex Modified: userguide/branches/release-7.1/recent-changes.tex =================================================================== --- userguide/branches/release-7.1/recent-changes.tex 2012-11-29 02:21:00 UTC (rev 16355) +++ userguide/branches/release-7.1/recent-changes.tex 2012-11-29 12:57:45 UTC (rev 16356) @@ -19,98 +19,98 @@ \def\rcSubsubsect#1{\subsubsect{#1}} \fi -\rcSect[next-release]{Next Release} +%\rcSect[next-release]{Next Release} -\rcSubsect{November 2012} +\rcSect[7.1]{Version 7.1 (November 2012)} -The method of populating a corpus from a single file has been updated to allow any mime type -to be used when creating the new documents. +\rcSubsect{New plugins} -A new document format plugin, \verb!Format_MediaWiki!, has been added to allow the easy use -of documents containing MediaWiki markup. This includes both plain text documents containing -markup, or XML dump files (such as those produced by Wikipedia). -See Section~\ref{sec:creole:mediawiki} for more details. +The \emph{TermRaider} plugin (see Section~\ref{sec:creole:termraider}) has +provides a toolkit and sample application for term extraction. -The OpenNLP plugin has been completely revised to use Apache OpenNLP 1.5.2 and -the corresponding set of models. See Section~\ref{sec:misc-creole:opennlp} for -details. +Two new plugins, \emph{Tagger\_Zemanta} (see +Section~\ref{sec:misc-creole:zemanta}) and \emph{Tagger\_Lupedia} (see +Section~\ref{sec:misc-creole:lupedia}) provide PRs that wrap online annotation +services provided by Zemanta and Ontotext. -The Stanford Parser plugin (see Section~\ref{sec:parsers:stanford}) has been -updated to parser version 2.0.4. +A new plugin named \emph{Coref\_Tools} includes a framework for fast +co-reference processing, and one PR that performs orthographical co-reference in +the style of the ANNIE Orthomatcher. See Section~\ref{sec:creole:coref-tools} +for full details. -The TermRaider plugin (see Section~\ref{sec:creole:termraider}) has been added -to provide a toolkit and sample application for term extraction. +A new \emph{Configurable Exporter} PR in the Tools plugin, allowing annotations +and features to be exported in formats specified by the user (e.g. for use with +external machine learning tools). See Section~\ref{sec:misc-creole:confexport} +for details. -Two new plugins, Tagger\_Zemanta (see Section~\ref{sec:misc-creole:zemanta}) and -Tagger\_Lupedia (see Section~\ref{sec:misc-creole:lupedia}) have been added in -GATE. The two plugins have PRs that wrap online annotation services provided by -Zemanta and Ontotext. +Support for reading a number of new document formats has also been added: +\begin{description} +\item[PubMed] and the Cochrane Library formats (see + Section~\ref{sec:creole:pubmed}). +\item[CoNLL ``IOB''] format (see Section~\ref{sec:corpora:conll}). +\item[MediaWiki] markup, both plain text and XML dump files such as those from + Wikipedia (see Section~\ref{sec:creole:mediawiki}). +\end{description} -\rcSubsect{October 2012} +In addition, ``ready-made applications'' have been added to many existing +plugins (notably the \emph{Lang\_*} non-English language plugins) to make it +easier to experiment with their PRs. -The PRs defined in the ANNIE plugin are now described by annotations on the Java classes -rather than explicitly inside creole.xml. The main reason for this change is to enable the -definitions to be inherited to any subclasses of these PRs. Creating an empty subclass is a -common way of providing a PR with a different set of default parameters (this is used extensively -in the language plugins to provide custom gazetteers and named entity transducers). This has -the added benefit of ensuring that new features also automatically percolate down to these -subclasses. +\rcSubsect{Library updates} -The corpus parameter of LanguageAnalyser (an interface most, if not all, PRs implement) is -now annotated as \verb|@Optional| as most implementations do not actually require the -parameter to be set. +Updated the Stanford Parser plugin (see Section~\ref{sec:parsers:stanford}) to +version 2.0.4 of the parser itself, and added run-time parameters to the PR to +control the parser's dependency options. -The following plugins now expose ready made applications to make experimenting with -the PRs they contain easier: Lang\_Arabic, Lang\_Cebuano, Lang\_French, Lang\_German, -Lang\_Romanian, and Tagger\_PennBio. +The Measurement and Number taggers have been upgraded to use JAPE+ instead of +JAPE. This should result in faster processing, and also allows for more memory +efficient duplication of PR instances, i.e. when a pool of applications is +created. -\rcSubsect{July 2012} +The OpenNLP plugin has been completely revised to use Apache OpenNLP 1.5.2 and +the corresponding set of models. See Section~\ref{sec:misc-creole:opennlp} for +details. -A new plugin was added to the Tools plugin, called Configurable -Exporter. This plugin allows annotations and features to be exported -in formats specified by the user. It is documented in -Section~\ref{sec:misc-creole:confexport}. +\rcSubsect{GATE Embedded API changes} -\rcSubsect{May 2012} -A new plugin was added that supports the loading of PubMed text documents into -GATE. Details are included in Section~\ref{sec:creole:pubmed}. +Some of the most significant changes in this version are ``under the bonnet'' +in GATE Embedded: +\begin{itemize} +\item The class loading architecture underlying the loading of plugins and the + generation of code from JAPE grammars has been re-worked. The new version + allows for the complete unloading of plugins and for better memory handling of + generated classes. Different plugins can now also use different versions of the + same 3rd party libraries. There have also been a number of changes to the way + plugins are (un)loaded which should provide for more consistent behaviour. +\item The GATE XML format has been updated to handle more value types + (essentially every data type supported by XStream + (\url{http://xstream.codehaus.org/faq.html}) should be usable as feature name + or value. Files in the new format can be opened without error by older GATE + versions, but the data for the previously-unsupported types will be + interpreted as a String, containing an XML fragment. +\item The PRs defined in the ANNIE plugin are now described by annotations on + the Java classes rather than explicitly inside creole.xml. The main reason for + this change is to enable the definitions to be inherited to any subclasses of + these PRs. Creating an empty subclass is a common way of providing a PR with a + different set of default parameters (this is used extensively in the language + plugins to provide custom gazetteers and named entity transducers). This has + the added benefit of ensuring that new features also automatically percolate + down to these subclasses. If you have developed your own PR that extends one + of the ANNIE ones you may find it has acquired new parameters that were not + there previously, you may need to use the \verb!@HiddenCreoleParameter! + annotation to suppress them. +\item The corpus parameter of LanguageAnalyser (an interface most, if not all, + PRs implement) is now annotated as \verb|@Optional| as most implementations + do not actually require the parameter to be set. +\item When saving an application the plugins are now saved in the same order in + which they were originally loaded into GATE. This ensures that dependencies + between plugins are correctly maintained when applications are restored. +\item API support for working with relations between annotations was added. See + Section~\ref{sec:api:relations} for more details. +\item The method of populating a corpus from a single file has been updated to + allow any mime type to be used when creating the new documents. +\end{itemize} -The Measurement and Number taggers have been upgraded to use JAPE+ instead of JAPE. -This should result in faster processing, and also allows for more memory efficient -duplication of PR instances, i.e. when a pool of applications is created. +And numerous smaller bug fixes and performance improvements\ldots -\rcSubsect{March 2012} -The class loading architecture underlying the loading of plugins and the generation -of code from JAPE grammars has been re-worked. The new version allows for the complete -unloading of plugins and for better memory handling of generated classes. Different -plugins can now also use different versions of the same 3rd party libraries. There -have also been a number of changes to the way plugins are (un)loaded which should -provide for more consistent behaviour. - -When saving an application the plugins are now saved in the same order in which they -were originally loaded into GATE. This ensures that dependencies between plugins -are correctly maintained when applications are restored. - -A new plugin named Coref\_Tools was added. It includes a framework for fast -co-reference processing, and one PR that performs orthographical co-reference in -the style of the ANNIE Orthomatcher. See Section~\ref{sec:creole:coref-tools} -for full details. - -API support for working with relations between annotations was added. See -Section~\ref{sec:api:relations} for more details. - -The GATE XML format has been updated to handle more value types (essentially -every data type supported by XStream -(\url{http://xstream.codehaus.org/faq.html}) should be usable as feature name or -value. Files in the new format can be opened without error by older GATE -versions, but the data for the previously-unsupported types will be interpreted -as a String, containing an XML fragment. - -Updated the Stanford Parser plugin (see Section~\ref{sec:parsers:stanford}) to -version 2.0.1 of the parser itself, and added run-time parameters to the PR to -control the parser's dependency options. - -\rcSubsect{February 2012} - -We have added support for reading files in the CoNLL or ``IOB'' format into GATE -documents, described in Section~\ref{sec:corpora:conll}. +% vim:ft=tex Modified: userguide/branches/release-7.1/tao_main.tex =================================================================== --- userguide/branches/release-7.1/tao_main.tex 2012-11-29 02:21:00 UTC (rev 16355) +++ userguide/branches/release-7.1/tao_main.tex 2012-11-29 12:57:45 UTC (rev 16356) @@ -25,8 +25,8 @@ % Is this manual for a real release or for a snapshot? If it is for a release, % set \gatereleasetrue instead of \gatereleasefalse -\gatereleasefalse -%\gatereleasetrue +%\gatereleasefalse +\gatereleasetrue \newif\ifprintedbook % Is this a book to be printed? If it is to be printed, This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. ------------------------------------------------------------------------------ Keep yourself connected to Go Parallel: VERIFY Test and improve your parallel project with help from experts and peers. http://goparallel.sourceforge.net _______________________________________________ GATE-cvs mailing list GATE-cvs@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/gate-cvs