Revision: 17956
http://sourceforge.net/p/gate/code/17956
Author: ian_roberts
Date: 2014-05-11 13:08:06 +0000 (Sun, 11 May 2014)
Log Message:
-----------
First cut of a consolidated changelog for 8.0
Modified Paths:
--------------
userguide/branches/release-8.0/Makefile
userguide/branches/release-8.0/recent-changes.tex
Property Changed:
----------------
userguide/branches/release-8.0/
userguide/branches/release-8.0/recent-changes.tex
Index: userguide/branches/release-8.0
===================================================================
--- userguide/branches/release-8.0 2014-05-11 13:04:28 UTC (rev 17955)
+++ userguide/branches/release-8.0 2014-05-11 13:08:06 UTC (rev 17956)
Property changes on: userguide/branches/release-8.0
___________________________________________________________________
Modified: svn:mergeinfo
## -1,4 +1,4 ##
/userguide/branches/release-6.0:13203-13218
/userguide/branches/release-7.0:15332-15399
/userguide/tags/release-7.0:15400-15404
-/userguide/trunk:10614-10900
+/userguide/trunk:10614-10900,17945,17951-17952,17954
\ No newline at end of property
Modified: userguide/branches/release-8.0/Makefile
===================================================================
--- userguide/branches/release-8.0/Makefile 2014-05-11 13:04:28 UTC (rev
17955)
+++ userguide/branches/release-8.0/Makefile 2014-05-11 13:08:06 UTC (rev
17956)
@@ -15,7 +15,7 @@
alignment.tex parsers.tex crowdsourcing.tex uima.tex misc-creole.tex
changes.tex \
plugin-name-map.tex design.tex ant-tasks.tex negram.tex \
postag.tex mlconfig.tex iaa-kappa.tex shortcuts.tex colophon.tex \
-recent-changes.tex cloud.tex teamware.tex mimir.tex domain-creole.tex
+recent-changes.tex cloud.tex teamware.tex mimir.tex domain-creole.tex
social-media.tex
EDITOR= viw
#TEX= texify --pdf
TEX= latex
Modified: userguide/branches/release-8.0/recent-changes.tex
===================================================================
--- userguide/branches/release-8.0/recent-changes.tex 2014-05-11 13:04:28 UTC
(rev 17955)
+++ userguide/branches/release-8.0/recent-changes.tex 2014-05-11 13:08:06 UTC
(rev 17956)
@@ -19,150 +19,138 @@
\def\rcSubsubsect#1{\subsubsect{#1}}
\fi
-\rcSect[next-release]{Next Release}
+\rcSect[8.0]{Version 8.0 (May 2014)}
-\rcSubsect{April 2014}
+GATE 8.0 is a major release which brings some major new features, many new and
+updated plugins, and significant under-the-bonnet changes to GATE Embedded.
-Added `The Duplicator' to the developer tools plugin, to allow for testing of
-resource duplication in GATE Developer. See
-Section~\ref{sec:misc-creole:dev-tools} for more details.
+\rcSubsect{Major changes}
-GATE now requires Java 7. If you are stuck on Java 6 then you will need to
-continue to use GATE 7.1.
+\textbf{Java 7 required}
-The Lucene specific implementation of the Information Retrieval plugin
-(Section~\ref{sec:misc-creole:ir}) has been moved out of the core and into
-the plugin. This means that Lucene is no longer a core dependency allowing
-individual plugins to specify any version of Lucene they may require.
+GATE 8.0 requires \textbf{Java 7} or later to run.
-\rcSubsect{March 2014}
+\textbf{Tools for Twitter}
-A new plugin containing a number of developer orientated tools has been added
-to the core distribution. See Section~\ref{sec:misc-creole:dev-tools} for
-more details.
+A new ``Twitter'' plugin provides tools dedicated to Twitter data:
+\begin{itemize}
+\item format parsers to handle Tweets in the JSON formats produced by the
+ Twitter APIs
+\item Twitter-specific components such as a tokeniser and POS tagger
+\item the \emph{TwitIE} named entity annotation pipeline.
+\end{itemize}
-Fixed a bug in the ANNIC query parser related to the escaping of reserved
-characters within feature values.
+See section~\ref{sec:social:twitter} for full details.
-Added a PR to access the TextRazor online annotation service. See
-section~\ref{sec:misc-creole:textrazor} for details.
+\textbf{ANNIE Refreshed}
-Updated GATE to support Java 8.
+The ANNIE named entity annotation pipeline which has been the mainstay of many
+GATE applications for many years has been brought up to date, with new
+gazetteers and improved JAPE grammars giving improved precision and recall on
+common test corpora.
-\rcSubsect{February 2014}
+\textbf{Tools for Crowd Sourcing}
-The Twitter JSON document format and corpus population tool are now documented
-in the Twitter plugin (Section~\ref{sec:creole:tweet}).
+A new \verb!Crowd_Sourcing! plugin provides facilities to support generation of
+manually annotated corpora via the CrowdFlower crowdsourcing
+platform\footnote{\url{http://www.crowdflower.com}}. The plugin provides
+support for two different kinds of tasks, general entity annotation (e.g.
+determining which words in a given sentence are person names) and entity
+linking (e.g. for ontology-based annotation, where the spans of the entities
+are known but not which particular ontology instance each annotation
+corresponds to). Using crowdsourcing you can generate multiply-annotated gold
+standard corpora rapidly and at relatively low cost. For full details see
+chapter~\ref{chap:crowd}.
-There has been a fairly extensive clean out of old deprecated and unsupported
-code from the main code base. In general this should not cause any problems,
-as the code has been deprecated for a long time, with sensible replacements
-provided. The one change that might cause problems is the removal of the
-\verb|annotations| parameters from the right hand side of JAPE rules. If you
try
-loading an old JAPE grammar and it fails with an error along the lines of
-\verb|Error: annotations cannot be resolved at line X| then you will need to
-fix the JAPE files before they will load. Either update the code to correctly
-use the \verb|inputAS| and \verb|outputAS| parameters or if this looks too
-complex simply add the following line to the beginning of the right hand side
-code block: \verb|AnnotationSet annotations = outputAS;|
+\rcSubsect{Other new and improved plugins}
-\rcSubsect{January 2014}
+\begin{itemize}
+\item New language plugins to support \emph{Russian} and \emph{Bulgarian}
+\item Integration of the Stanford POS Tagger
+ (section~\ref{sec:misc:creole:stanford}), which is used by TwitIE
+\item A document normalizer plugin, predominantly to normalize punctuation such
+ as Microsoft Word ``smart quotes'' (see
+ section~\ref{sec:misc-creole:doc-normalizer})
+\item Wrappers for the \emph{AlchemyAPI} keyword and entity extraction services
+ (in the \verb!AlchemyAPI! plugin)
+\item Wrapper for the \emph{TextRazor} annotation service (see
+ section~\ref{sec:misc-creole:textrazor}).
+\item New document format parser to populate a GATE corpus from one or more CSV
+ files (see section~\ref{sec:creole:csv}).
+\item Support for loading and saving GATE XML files in the binary
+ \emph{FastInfoset} format (see section~\ref{sec:creole:fastinfoset}).
+\item Various improvements to the \verb!Learning! plugin, in particular to
+ support numeric and boolean features (see
+ section~\ref{sec:ml:batch-learning-pr})
+\item Improvements to the \emph{TermRaider} term extraction plugin (see
+ section~\ref{sec:creole:termraider})
+\item The OntoRoot Gazetteer (in the \verb!Gazetteer_Ontology_Based! plugin)
+ now supports tokenisers and POS taggers other than the default ANNIE PR
+ types, making it possible to use other preprocessing tools for non-English
+ data
+\item Further improvements to the classloading model to better isolate plugins
+ from one another.
+\item A new \verb!enableDebugging! runtime parameter for JAPE grammars will add
+ additional features to every generated annotation detailing which rule was
+ responsible for creating the annotation.
+\end{itemize}
-A new plugin that allows for document normalization has been added. This
-plugin is predominately aimed at normalizing punctuation symbols (i.e.
-replacing Word style apostrophies and hypens with their ASCII equivalemts)
-to provide a common baseline for further components. See Section
-\ref{sec:misc-creole:doc-normalizer} for further details.
+\rcSubsect{Bug fixes and other improvements}
-\rcSubsect{December 2013}
+\begin{itemize}
+\item The annotation schema LR type is now available by default without the
+ need to load any plugins. The schemas that were previously loaded by default
+ by the ANNIE plugin must now be loaded explicitly if you require them
+ (section~\ref{sec:developer:schemaannotationeditor}). Annotation schemas now
+ support the \verb!include! element, so multiple schemas can be loaded by
+ loading a single master file.
+\item The segment processing PR
+ (section~\ref{sec:alignment:segment-processing}) now preserves annotation
+ IDs, allowing ID-sensitive tools such as coreference to work properly.
+\end{itemize}
-The Relations API (Section \ref{sec:api:relations}) has been updated as
-relations are now treated as ``first-class citizens'' in a similar way to
-annotations. The major changes are that relation sets are now directly
-associated with an annotation set and can only contain relations between
-members of that set or other relations within the same set, and that
-relations are now handled separately when stored in a GATE XML document
-rather than being serialized as a document feature. There is currently
-no support for directly editing relations within GATE Developer (a simple
-viewer is provided as a tab in the document viewer), so relations can only
-be created via the API.
+\rcSubsect{For developers}
-The GATE XML version number has been pushed up to 3 (it was previously 2) as
-the saving of relations into the GATE XML files mean they cannot be opened with
-previous versions of GATE. Documents without relations are still saved in a
-backwards compatible format, but the change in version number will help with
-diagnosing bug reports etc.
-
-A new plugin to support processing Bulgarian text has been added. Currently
this
-consists of a PR that integrates the
-\htlink{http://lml.bas.bg/~nakov/bulstem/index.html}{BulStem stemmer} using
-code kindly donated by Ivelina Nikolova. See Section
-\ref{sec:misc-creole:language-plugins:bulgarian} for full details.
-
-\rcSubsect{November 2013}
-
-A new plugin, \verb!Format_CSV!, provides support for populating a corpus from
-one or more CSV files. See Section \ref{sec:creole:csv} for details.
-
-There have been a number of changes to the support for Annotation Schemas:
+Changes of note for users of the GATE Embedded APIs include:
\begin{itemize}
-\item Annotation Schema is now a default language resource, which means that
- schemas can be loaded without having first loaded the ANNIE plugin.
-\item Support for using the \verb|include| element within an Annotation Schema
- XML file has been added. This allows for multiple schemas to be loaded by
- loading a single file.
-\item Loading the ANNIE plugin no longer loads a set of hidden Annotation
- Schemas. To load the ANNIE schemas you need to load the
- \verb|resources/schemas/ANNIE-Schemas.xml| file from within the ANNIE plugin.
+\item A new data model to represent relations between annotations, see
+ section~\ref{sec:api:relations} for details. Note that due to the addition
+ of these new data structures, GATE XML documents saved with version 8.0 are
+ not compatible with older versions of GATE (though GATE 8.0 can still open
+ documents saved in earlier versions)
+\item A new ``resource helper'' mechanism allows plugins to contribute
+ additional actions to existing resource types, both in the Developer GUI
+ (section~\ref{sec:creole-model:tools:resourcehelpers}) and in the Embedded
API
+ (section~\ref{sec:api:resourcehelpers})
+\item A new class \verb!gate.corpora.DocumentJsonUtils! provides methods to
+ export a GATE document in a JSON format compatible with that used by Twitter.
+ See the JavaDoc documentation for details.
+\item Many deprecated classes, fields and methods have been removed. If you
+ were previously calling any of these deprecated APIs you will need to update
+ your code accordingly. Also some classes in the GATE core that were only
+ used by one plugin have been moved into the respective plugin's source tree.
+ In particular, Java RHS actions in JAPE rules no longer provide the
+ long-deprecated \verb!annotations! variable -- use \verb!inputAS! or
+ \verb!outputAS! as appropriate.
+\item Many library dependencies have been updated to more recent versions.
+\item The GATE APIs make much wider use of generics than previously -- many
+ places in the code that previously used raw types are now properly generic
+\item A new \verb!Developer_Tools! plugin
+ (section~\ref{sec:misc-creole:dev-tools}) provides utilities to assist in
+ debugging applications in GATE Developer.
\end{itemize}
-The GATE source code has been restructured slightly to make the separation of
-the main and test code explicit. This ensures that no code can ever end up
-relying upon test code accidentally. Changes have also been made to the Eclipse
-project files to separate the core code from the plugins.
-
-\rcSubsect{September 2013}
-
-A number of changes have been made to the Groovy plugin:
+If you are working on the core GATE source code, note that:
\begin{itemize}
-\item Upgraded the \emph{Groovy} plugin to use Groovy version 2.0.
-\item The implicit imports available when using Groovy have been reduced to
- just \verb|gate| and \verb|gate.util|. If you were using classes from other
- GATE packages you will need to explcitly import them.
-\item Each Groovy script used via the plugin (i.e. instances of the ScriptPR,
- Scriptable Controllers, or the Groovy Console) are now isolated from each
- other meaning that classes created in once cannot refer to those created in
- another instance (it is unlikely that this was possible before, but now it
- definately isn't).
+\item the source tree has been split into ``main'' and ``test'', isolating the
+ test classes from the rest of the source
+\item each plugin is now a separate Eclipse ``project'', and the main project
+ is just the core sources, which makes it easier to control dependencies among
+ the different parts
+\item dependencies are no longer checked in to subversion, instead they are
+ fetched at build time from the Maven central repository by Apache Ivy.
\end{itemize}
-A new plugin that wraps the Stanford Part-of-Speech tagger. See Section
-\ref{sec:misc:creole:stanford} for details.
-
-\rcSubsect{August 2013}
-
-Added support for Resource Helpers which can added new features to exisiting
-resource types without requiring a Visual Resource and which can also be
accessiable
-via the embedded API. See Sections \ref{sec:creole-model:tools:resourcehelpers}
-and \ref{sec:api:resourcehelpers} for details.
-
-Added support for populating a corpus from a MediaWiki XML dump file.
Previously
-loading an XML file containing multiple pages resulted in only the last page
being
-used to create the document. See Section \ref{sec:creole:mediawiki} for full
details.
-
-A new plugin to support the reading and writing of compressed XML files in the
Fast
-Infoset format. This format gives space savings of around 80\% when used to
store
-GATE XML documents. For full details see Section \ref{sec:creole:fastinfoset}.
-
-New support for processing Russian including a part-of-speech tagger,
morphological
-analyser and a gazetteer. See Section
\ref{sec:misc-creole:language-plugins:russian}
-for more details.
-
-\rcSubsect{March 2013}
-
-Fixed a bug which caused the 7.1 version of the OntoRootGazetteer to produce no
-Lookup annotations in its default configuration.
-
\rcSect[7.1]{Version 7.1 (November 2012)}
\rcSubsect{New plugins}
Property changes on: userguide/branches/release-8.0/recent-changes.tex
___________________________________________________________________
Modified: svn:mergeinfo
## -3,4 +3,4 ##
/userguide/branches/release-7.0/recent-changes.tex:15332-15399
/userguide/branches/release-7.1/recent-changes.tex:16356-16357
/userguide/tags/release-7.0/recent-changes.tex:15400-15404
-/userguide/trunk/recent-changes.tex:10614-10900
+/userguide/trunk/recent-changes.tex:10614-10900,17951-17952
\ No newline at end of property
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Is your legacy SCM system holding you back? Join Perforce May 7 to find out:
• 3 signs your SCM is hindering your productivity
• Requirements for releasing software faster
• Expert tips and advice for migrating your SCM now
http://p.sf.net/sfu/perforce
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs
