PK
OOA7`#x
x
desc/BNCCPE.xml
InputDirectory/data/BNCXML/corpus/A_AllOutputDirectory/home/pebble/data/BNCXML/corpus/A-done-1immediate
PK
G|7m= = desc/BNCTypeSystemDescriptor.xml
BNCTypeSystemDescriptor1.0com.digitalpebble.bncxml.SourceDocumentInformationuima.tcas.Annotationuriuima.cas.String
PK
G|7*[; ) desc/BNCXMLCollectionReaderDescriptor.xml
org.apache.uima.javacom.digitalpebble.bncxml.BNCXmlCollectionReaderBNCXMLCollectionReaderDescriptorReads BNC XML files from a directory1.0InputDirectoryStringfalsetruecom.digitalpebble.rasp.Tokencom.digitalpebble.rasp.Sentencecom.digitalpebble.bncxml.SourceDocumentInformationtruefalsetrue
PK
G|7* ! desc/BNCXMLConsumerDescriptor.xml
org.apache.uima.javacom.digitalpebble.bncxml.BNCXMLConsumerBNCXMLConsumerDescriptorGenerates a XML representation of BNC documents. Needs to have Sentences and Tokens in the default View and the original XML of the file in a view called 'xml'1.0DigitalPebbleOutputDirectoryStringfalsetrueOutputDirectorytemp-uima-output/bncxmlcom.digitalpebble.rasp.WordFormcom.digitalpebble.rasp.Tokencom.digitalpebble.rasp.Sentencecom.digitalpebble.rasp.Dependencycom.digitalpebble.bncxml.SourceDocumentInformationxmlfalsefalsefalse
PK
G|7}t7 7 desc/CombinedRASPDescriptor.xml
org.apache.uima.javafalseCombinedRASPDescriptor1.0SentenceSplitterTokenizerPosTaggerMorpherParsertruetruefalse
PK
G|78! ! desc/Morpher.xml
org.apache.uima.javatruecom.digitalpebble.rasp.morph.MorphoAnnotatorMorpher1.0com.digitalpebble.rasp.Tokencom.digitalpebble.rasp.Sentencetruetruefalse
PK
G|7,W desc/Parser.xml
org.apache.uima.javatruecom.digitalpebble.rasp.parser.ParserAnnotatorParser1.0outputspecifies the type and content of annotations generated by the Parser
-oa : trees labelled with grammar aliases
-ot : trees labelled with grammar rule names;
-otg : rule-labelled trees and grammatical relations;
-og : grammatical relations
-ogio : grammatical relations weighted using a variant of the inside-outside algorithm
-ogw : weighted grammatical relations computed from the top n trees (only useful if the number of parses option -n is set to >1)
-otgioStringfalsetrueparseNumGive the maximum number of parses that should be produced for each sentence. The default is 1; a value of zero indicates all parses.IntegerfalsefalsetimeSet a CPU time limit (in seconds) for the processing of each sentence (default 20).IntegerfalsefalsesubcategorisationTurn on the use of verb subcategorisation frame probabilities; there is built-in information for around 500 verbs.BooleanfalsefalsephrasalVerbsuse of a list of phrasal verbs that normally allows more accurate identification of verb-particle constructions.BooleanfalsefalseparseNum1time20subcategorisationfalsephrasalVerbstrueoutput-otgcom.digitalpebble.rasp.WordFormcom.digitalpebble.rasp.Tokencom.digitalpebble.rasp.Sentencecom.digitalpebble.rasp.Dependencytruetruefalse
PK
t7Z[ desc/PosTagger.xml
org.apache.uima.javatruecom.digitalpebble.rasp.tagger.PosTaggerPosTagger1.0parametersStringParameters for the POS executable. See RASP documentation for more details. Only the parameter specifying the format (i.e multiple tags) is set implicitlyStringfalsetruegenerateMultipleTagsIf true, a Token will get one or more WordForm with a probability otherwise there will be only one WordForm per Token with a probability set to 1.0BooleanfalsetrueparametersStringB1 b C1 N t auxiliary_files/slb.trn d auxiliary_files/seclarge.lex j auxiliary_files/unkstats-seclarge m auxiliary_files/tags.mapgenerateMultipleTagstruecom.digitalpebble.rasp.Sentencecom.digitalpebble.rasp.Tokencom.digitalpebble.rasp.Tokencom.digitalpebble.rasp.WordFormentruetruefalse
PK
LE,7# # desc/RASPModulesCPE.xml
InputDirectory/usr/local/bin/apache-uima/examples/data/xmlLanguageenTokenizerPath/usr/local/bin/RASP/token/token.ix86_linux-1immediate
PK
G~;7nI desc/RASPTypes.xml
RASPTypes1.1com.digitalpebble.rasp.TokenA token for Raspuima.tcas.AnnotationwordFormsA Token is related to one or more WordFormuima.cas.FSArraycom.digitalpebble.rasp.WordFormcom.digitalpebble.rasp.SentenceAnnotation for a Sentenceuima.tcas.Annotationcom.digitalpebble.rasp.WordFormA WordForm consists of a POS tag, a lemma and possibly a probability. There is one or more WordForm per Token (as in the MAF ISO Norm)uima.tcas.Annotationlemmalemma of the Formuima.cas.StringPOSPOS tag for a given formuima.cas.Stringprobabilityuima.cas.Doublesuffixuima.cas.Stringcom.digitalpebble.rasp.DependencyA dependency between two word formsuima.tcas.Annotationdeptypeuima.cas.Stringsubtypeuima.cas.Stringheadcom.digitalpebble.rasp.WordFormdepcom.digitalpebble.rasp.WordFormcom.digitalpebble.rasp.ClauseA clause as returned by the RASP analyser. It can contain one or more word forms or clausesuima.tcas.Annotationruleuima.cas.Stringsubclausesarray of subelements. contains WordForms or Clausesuima.cas.FSArrayuima.tcas.Annotation
PK
G|7k> desc/SentenceSplitter.xml
org.apache.uima.javatruecom.digitalpebble.rasp.splitter.SentenceSplitterSentenceSplitterSimple sentence splitter which calls an external command1.0com.digitalpebble.rasp.Sentencetruetruefalse
PK
G|7+
> > desc/Tokenizer.xml
org.apache.uima.javatruecom.digitalpebble.rasp.tokenizer.NativeTokenAnnotatorTokenizer1.0com.digitalpebble.rasp.Sentencecom.digitalpebble.rasp.Tokentruetruefalse
PK
G|7D" D" doc/BNC.html
RASP4UIMA
RASP4UIMA contains two additional resources for the XML Edition of theBritish National Corpus. The British National Corpus (BNC) is a 100 million word collection of samples of written
and spoken language from a wide range of sources, designed to represent a wide
cross-section of British English from the later part of the 20th century, both spoken and
written. The latest edition is the BNC XML Edition, released in 2007.
These resources are provided as-is and are meant essentially to
illsutrate the use of RASP4UIMA. The corresponding XML descriptors can
be found in the /desc directory of RASP4UIMA.
BNC XML Collection Reader
This resource takes as input a directory containing XML documents
at the BNC DTD and convert them into UIMA CASes. The original XML
is kept in a separate SOFA while the <s> and <w> elements
are converted into annotations of type com.digitalpebble.rasp.Sentence and com.digitalpebble.rasp.Token.
The POS tagger, Morpher and Parser of RASP4UIMA can then be used on these documents.
BNC XML Consumer
This resource takes a CAS generated by the BNC XML Collection Reader and annotated by RASP4UIMA.
It regenerates the XML content of the original document but
replaces the original <s> and <w> elements with the
corresponding information generated by the RASP modules. In
short, the POS tags and lemmas are replaced with the RASP equivalents.
In addition the grammatical relations found by RASP are added to the XML.
The example below shows a sentence from a BNC document as generated by the BNC XML Consumer.
RASP is a
domain-independent, robust parsing system for English. For
ease of installation, the system is distributed in the form of binaries
for 3 widespread unix architectures (Intel-32bit and -64bit/Linux, and
Sparc/Solaris). It is free for
research purposes. RASP is described in :
Briscoe, E., J. Carroll and R. Watson (2006) The Second Release of the
RASP System. In Proceedings of the COLING/ACL 2006 Interactive
Presentation Sessions, Sydney, Australia.
UIMA is an
Apache project in incubation which provides a component
framework for analysing unstructured content such as text, audio and
video. It comprises an SDK and tooling for composing and running
analytic components written in Java and C++.
Please contact the respective projects for any question related to RASP or UIMA. You can use the DigitalPebble user group for any question specific to RASP4UIMA.
Installation
This version of RASP4UIMA has been tested on Apache UIMA 2.1.0. It is
available as a PEAR package and can be installed with the PEAR installer. You
will also need to download and install RASP2 from the RASP
project page.
Run the PearInstaller
(e.g. /usr/local/bin/apache-uima/bin/runPearInstaller.sh). Select
the RASP4UIMA pear file and a target directory for the installation. In
this manual we assume RASP4UIMA has been
installed in /usr/local/bin/RASP4UIMA.
Note: RASP4UIMA relies on a system environment (rasp.home) to determine where the original RASP executables are located. See $RASP4UIMA/metadata/setenv.txt formore details.
Make sure you specify the location of RASP with -Drasp.home when
you call the UIMA executables. For instance, if you want to run your
component in the Collection Processing Engine Configurator GUI
application, you need to add the environment variables settings from
the component's setenv.txt file to the cpeGui.bat (cpeGui.sh) script
file in the <UIMA_HOME>/bin directory.
Test
Once RASP4UIMA has been installed with the PEAR installer, you
can test the installation with the Collection Processing Engine (CPE).
Please refer to the UIMA documentation for more details on the use of
this tool. Don't forget to add the variables from setenv.txt to the script (see above).
For this test we'll use the Collection Reader
and Xmi Writer
CAS Consumer
available in the UIMA examples. These two components are
respectively in charge of converting a collection of documents into
CASes and serialize the CASes into XML files. Their descriptors can be found in the /examples directory of UIMA.
Click on the button Add
of the section Analysis
Engines and select the file SentenceSplitter.xml located
in the /desc
directory of the the RASP4UIMA installation (e.g. /usr/local/bin/RASP4UIMA/desc).
Repeat the procedure for the files Tokenizer.xml, POStagger.xml, Morpher.xml and Parser.xml.
You should get something similar to the screenshot below:
Click on the Play button, after a while you should get a summary of the
process. You can use the AnnotationViewer
of
UIMA which takes as input a directory containing XML files at the xmi
format and a TypeSystem file. The TypeSystem description for
RASP4UIMA is in the file /desc/RASPTypes.xml.
Once you've specified both the input directory and the TypeSystem,
click on View and
double click on one of the documents of the list. You should get
something similar to the screenshot below. More details about the
Annotation Types generated by RASP4UIMA can be found in the Modules
section.
Takes as input the text of a document and creates annotations of Type Sentence. The sentence splitter of RASP identifies acronyms and common abbreviations (e.g. Dr.)
Tokenizer
Creates annotations of type Token, using the information about Sentences. A Token is a simple annotation which contains only a list of WordForms. This separation between Tokens and WForms is based on the MAF ISO proposal
POS Tagger
The WordForms are created by the Part of Speech tagger. The tagset is close to CLAWS C7 (see e.g. Appendix C of Jurafsky,
D. and Martin, J. Speech and Language Processing, Prentice-Hall, 2000
for more details), although it is in fact a cut down version of the
CLAWS C2 tagset.
A WordForm gets a POS attribute, which is a simple String and a probability.
The POS Tagger takes two
parameters: a String which corresponds to the parameters used in
the original RASP and a boolean parameter indicating whether or not to
generate different WordForms for a Token.
Morpher
Next the tagger output is lemmatized, based on the tags
assigned to word tokens. See Briscoe and Carroll (2002) for further details
and a reference to a detailed paper describing this module. In RASP4UIMA the Morpher adds an attribute lemma to the WordForms.
Parser
The probabilistic parser analyses the PoS tag sequence or
chart of initial more probable tags and generates a parse forest
representation containing all possible subanalyses with associated
probabilities. From this representation it is able to construct
the n-best syntactic trees and / or (weighted) grammatical relations.
The parser generates annotations of the type Clause and Dependency. A Clause has an attribute rule and contains an array of subclauses. A subclause can be a WordForm or another Clause element. Dependencies have a type and subtype and link two WordForms as head and dependency.
The main parameter of the Parser is the string specifying its output.
Please refer to the documentation of RASP for more details about the
values allowed for this parameter and the other options available.
Note that the Parser requires a recent machine with minimum 1.5G of RAM.
RASP4UIMA Type System
The illustration below summarizes the Types used in RASP4UIMA.
