edu.nyu.jet.aceJet

Class Ace

java.lang.Object

edu.nyu.jet.aceJet.Ace

public class Ace
extends Object

procedures for generating ACE output for a Jet document.

Field Summary

Fields

Modifier and Type	Field and Description
static boolean	entityTrace
static edu.nyu.jet.aceJet.RelationPatternSet	eve
static Gazetteer	gazetteer
static boolean	monocase if true, case information is not considered significant (for finding names or sentence boundaries).
static boolean	perfectEntities
static boolean	perfectMentions
static boolean	preferEntities
static boolean	preferRelations
static boolean	useParser

Constructor Summary

Constructors

Constructor and Description
Ace()

Method Summary

Methods

Modifier and Type	Method and Description
static boolean	allLowerCase(Document doc)
static boolean	allLowerCase(Document doc, Span span) return true if either all the letters in span are lower case, or the fraction of letters which are upper case exceeds MAX_UPPER.
static void	buildAceEntities(Document doc, String docId, AceDocument aceDoc) create ACE entities from entity annotations produced by refres.
static String	getDocId(Document doc) returns the document ID of Document doc, if found, else returns null.
static void	init(String propertyFile) initialize ACE: read property file and load all resources specified by the properties file.
static void	initForFileIO(String[] args) 'initForFileIO' provides the initialization component of the 'main' method.
static void	main(String[] args) generate ACE annotation files (in APF) format for a list of documents.
static AceDocument	processDocument(Document doc, String sourceId, String sourceFile, String docPathBase) process a (Jet) document and create a corresponding AceDocument.
static void	processFile(String currentDocPath) process a file containing a source document and generate an Ace APF file.
static void	processFileAndCatchError(String currentDocPath) process a single document to generate an Ace APF file, catching any exceptions which occur.
static void	setPatternSet(String fileName)
static void	tagReciprocalRelations(Document doc) assigns reciprocal relations subject-1 and object-1
static boolean	titleCase(Document doc, Span span) returns true if Span span of Document doc appears to be capitalized as a title: if there are no words beginning with a lower-case letter except for a small list of function words (articles, possessive pronouns, prepositions, ...).

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

useParser

public static boolean useParser

perfectMentions

public static boolean perfectMentions

perfectEntities

public static boolean perfectEntities

preferRelations

public static boolean preferRelations

preferEntities

public static boolean preferEntities

entityTrace

public static boolean entityTrace

monocase

public static boolean monocase

if true, case information is not considered significant (for finding names or sentence boundaries).

gazetteer

public static Gazetteer gazetteer

eve

public static edu.nyu.jet.aceJet.RelationPatternSet eve

Constructor Detail

Ace

public Ace()

Method Detail

main

public static void main(String[] args)
                 throws IOException

generate ACE annotation files (in APF) format for a list of documents. Takes four to eight command line parameters:
propertyFile: Jet properties file
filelist: a list of the files to be processed
docDirectory: the path of the directory containing the input files
outDirectory: the path of the directory in which APF files are to be written
parseDir: the path of the directory containing parse trees [optional]
glarfDir: the path of the directory containing GLARF tuples [optional]
glarfSuffix: the file suffix for GLARF files
parseSuffix: the file suffix for parse files
For each file in filelist, the document is read from docDirectory.file.sgm and the APF file is written to outDirectory.file.sgm.apf

Throws:

IOException

initForFileIO

public static void initForFileIO(String[] args)
                          throws IOException

'initForFileIO' provides the initialization component of the 'main' method.

Throws:

IOException

init

public static void init(String propertyFile)
                 throws IOException

initialize ACE: read property file and load all resources specified by the properties file.

Throws:

IOException

processFileAndCatchError

public static void processFileAndCatchError(String currentDocPath)
                                     throws IOException

process a single document to generate an Ace APF file, catching any exceptions which occur.

Throws:

IOException

processFile

public static void processFile(String currentDocPath)
                        throws IOException

process a file containing a source document and generate an Ace APF file.

Parameters:

currentDocPath - The full path of the file containing the source document

Throws:

IOException

processDocument

public static AceDocument processDocument(Document doc,
                          String sourceId,
                          String sourceFile,
                          String docPathBase)
                                   throws IOException

process a (Jet) document and create a corresponding AceDocument.

Parameters:

doc - The (Jet) source document

sourceId - The source document id; typically the 'sourceFile', but without any file extension. If no DOCID is specified within the document text) as the base for the names of all entities, relations, events, etc.

sourceFile - The name of the source document. This is used to fill the source file field of the AceDocument. If 'doc' is read from a file, should be the file name, including extension, but not including directory path.

docPathBase - The full path to the source document, with the file extension removed; used to access corresponding parse and GLARF files. Should be null if these files are not being used

Returns:

The AceDocument

Throws:

IOException

getDocId

public static String getDocId(Document doc)

returns the document ID of Document doc, if found, else returns null. It looks for
the text of a DOCID annotation (as provided with ACE documents);
the text of a DOCNO annotation (as provided with TDT4 and TDT5 documents);
the DOCID feature of a document annotation (for GALE ASR transcripts); or
the id feature of a doc annotation (for newer LDC documents).

allLowerCase

public static boolean allLowerCase(Document doc)

allLowerCase

public static boolean allLowerCase(Document doc,
                   Span span)

return true if either all the letters in span are lower case, or the fraction of letters which are upper case exceeds MAX_UPPER. Either condition is an indication that case is not significant in detecting names or sentence boundaries.

titleCase

public static boolean titleCase(Document doc,
                Span span)

returns true if Span span of Document doc appears to be capitalized as a title: if there are no words beginning with a lower-case letter except for a small list of function words (articles, possessive pronouns, prepositions, ...).

buildAceEntities

public static void buildAceEntities(Document doc,
                    String docId,
                    AceDocument aceDoc)

create ACE entities from entity annotations produced by refres.

tagReciprocalRelations

public static void tagReciprocalRelations(Document doc)

assigns reciprocal relations subject-1 and object-1

setPatternSet

public static void setPatternSet(String fileName)
                          throws IOException

Throws:

IOException