@Override
public void process(JCas cas) throws AnalysisEngineProcessException {
this.setSource(cas);
this.setTarget(cas);
AnnotationIndex<Annotation> index = cas.getAnnotationIndex(this.getSourceType());
FSIterator<Annotation> iter = index.iterator();
while (iter.hasNext()) {
Annotation annotation = iter.next();
String source = null;
if (this.getSourceFeature() == null) {
source = annotation.getCoveredText();
} else {
source = annotation.getStringValue(this.getSourceFeature());
}
if (source != null) {
String target = this.getMapping().get(source);
if (target != null) {
if (this.update().booleanValue()) {
this.update(cas, annotation, this.getTargetFeature(), target);
} else {
this.create(
cas, this.getTargetFeature(), annotation.getBegin(), annotation.getEnd(), target);
}
}
}
}
}
/**
* Metodo per il process della Cas generata con l'upload del documenti in 'Nuovo pattern'.
*
* @param cas cas da analizzare
* @throws SQLException eccezione durante l'estrazione dei dati
*/
private void processPatternAttachmentCas(final JCas cas) throws SQLException {
HpmDao dao = new HpmDao();
// il valore della property è l'hpmPatternId
String kpeopleTagPattern = getKpeopleTagPattern(cas);
// recupero e salvo su db gli attachments associati alla creazione pattern
List<Document> docs = new ArrayList<Document>();
// recupero l'email dell'utente che ha generato il pattern
String email = getUserAuthorAnnotation(cas).getEmail();
AnnotationIndex<Annotation> aaIdx = cas.getAnnotationIndex(AttachmentAnnotation.type);
FSIterator<Annotation> itAa = aaIdx.iterator();
while (itAa.hasNext()) {
AttachmentAnnotation aa = (AttachmentAnnotation) itAa.next();
Document document = new Document();
document.setAttachmentType(new AttachmentType(2));
document.setGuid(aa.getUrlAttachment());
document.setHashcode(aa.getHashcode());
document.setTemplate(false);
document.setHpmAttachmentId(aa.getId());
document.setName(aa.getAttachmentName());
docs.add(document);
}
docs = dao.savePatternDocument(docs, email, kpeopleTagPattern);
}
private void clean(JCas cas) {
AnnotationIndex<Annotation> index = cas.getAnnotationIndex(SingleWordTermAnnotation.type);
FSIterator<Annotation> iterator = index.iterator();
while (iterator.hasNext()) {
SingleWordTermAnnotation annotation = (SingleWordTermAnnotation) iterator.next();
this.clean(cas, annotation);
}
}
private void display(JCas cas) {
AnnotationIndex<Annotation> index = cas.getAnnotationIndex(SourceDocumentInformation.type);
FSIterator<Annotation> iterator = index.iterator();
if (iterator.hasNext()) {
SourceDocumentInformation sdi = (SourceDocumentInformation) iterator.next();
this.getContext().getLogger().log(Level.INFO, "Cleaning terms of " + sdi.getUri());
}
}
private void select(JCas cas) {
AnnotationIndex<Annotation> index = cas.getAnnotationIndex(TermAnnotation.type);
FSIterator<Annotation> iterator = index.iterator();
while (iterator.hasNext()) {
TermAnnotation annotation = (TermAnnotation) iterator.next();
if (annotation.getOccurrences() <= this.getThreshold().intValue()) {
this.getAnnotations().add(annotation);
}
}
}
@Override
public void process(JCas cas) throws AnalysisEngineProcessException {
try {
AnnotationIndex<Annotation> index = cas.getAnnotationIndex(WordAnnotation.type);
FSIterator<Annotation> iterator = index.iterator();
while (iterator.hasNext()) {
WordAnnotation annotation = (WordAnnotation) iterator.next();
String norm = annotation.getCoveredText();
annotation.setLemma(norm);
annotation.setStem(norm);
}
} catch (Exception e) {
throw new AnalysisEngineProcessException(e);
}
}
/**
* aggiunge le proprietà presenti.
*
* @param event eventoda salvare
* @param cas oggetto Cas
*/
private void addProperties(final Event event, final JCas cas) {
AnnotationIndex<Annotation> annIdx = cas.getAnnotationIndex(PropertyAnnotation.type);
FSIterator<Annotation> it = annIdx.iterator();
while (it.hasNext()) {
PropertyAnnotation annotation = (PropertyAnnotation) it.next();
String key = annotation.getKey();
String value = annotation.getValue();
logger.debug("Aggiunta proprietà\nKey: " + key + "\nValue: " + value);
event.getProperties().put(key, value);
}
}
private void clean(JCas cas, SingleWordTermAnnotation annotation) {
Set<TermComponentAnnotation> delete = new HashSet<TermComponentAnnotation>();
AnnotationIndex<Annotation> index = cas.getAnnotationIndex(TermComponentAnnotation.type);
FSIterator<Annotation> iterator = index.subiterator(annotation);
while (iterator.hasNext()) {
TermComponentAnnotation component = (TermComponentAnnotation) iterator.next();
FSIterator<Annotation> subiterator = index.subiterator(component);
while (subiterator.hasNext()) {
delete.add((TermComponentAnnotation) subiterator.next());
}
}
for (TermComponentAnnotation del : delete) {
del.removeFromIndexes();
}
}
@Override
protected void generateAnnotations(JCas jcas)
throws AnalysisEngineProcessException, FeaturePathException {
// CAS is necessary to retrieve values
CAS currCAS = jcas.getCas();
for (String path : paths) {
// Separate Typename and featurepath
String[] segments = path.split("/", 2);
String typeName = segments[0];
// Try to get the type from the typesystem of the CAS
Type t = currCAS.getTypeSystem().getType(typeName);
if (t == null) {
throw new IllegalStateException("Type [" + typeName + "] not found in type system");
}
// get an fpi object and initialize it
// initialize the FeaturePathInfo with the corresponding part
initializeFeaturePathInfoFrom(fp, segments);
// get the annotations
AnnotationIndex<?> idx = currCAS.getAnnotationIndex(t);
FSIterator<?> iterator = idx.iterator();
while (iterator.hasNext()) {
AnnotationFS fs = (AnnotationFS) iterator.next();
try {
if (this.filterFeaturePath != null) {
// check annotation filter condition
if (this.filterFeaturePathInfo.match(fs, this.filterCondition)) {
createStemAnnotation(jcas, fs);
}
} else { // no annotation filter specified
createStemAnnotation(jcas, fs);
}
} catch (AnalysisEngineProcessException e) {
// TODO Auto-generated catch block
throw new IllegalStateException("error occured while creating a stem annotation", e);
}
}
}
}
/**
* Recupera il valore della proprietà kpeopletagpattern.
*
* @param cas cas da analizzare
* @return valore della proprietà kpeopletagpattern
*/
private String getKpeopleTagPattern(final JCas cas) {
AnnotationIndex<Annotation> annIdx = cas.getAnnotationIndex(PropertyAnnotation.type);
FSIterator<Annotation> it = annIdx.iterator();
String kpeopletagpattern = null;
while (it.hasNext()) {
PropertyAnnotation annotation = (PropertyAnnotation) it.next();
String key = annotation.getKey();
String value = annotation.getValue();
if (key.equals("kpeopletagpattern")) {
kpeopletagpattern = value;
}
}
return kpeopletagpattern;
}
private void adjust(JCas cas) {
AnnotationIndex<Annotation> index = cas.getAnnotationIndex(TermAnnotation.type);
FSIterator<Annotation> iterator = index.iterator();
while (iterator.hasNext()) {
TermAnnotation annotation = (TermAnnotation) iterator.next();
if (annotation.getVariants() != null) {
int occ = annotation.getOccurrences();
double freq = annotation.getFrequency();
double spec = annotation.getSpecificity();
for (int i = 0; i < annotation.getVariants().size(); i++) {
occ += annotation.getVariants(i).getOccurrences();
freq += annotation.getVariants(i).getFrequency();
spec += annotation.getVariants(i).getSpecificity();
}
annotation.setOccurrences(occ);
annotation.setFrequency(freq);
annotation.setSpecificity(spec);
}
}
}
/**
* aggiunge oggetti di tipo Document all'event. Nel caso di Communication associo anche la email
* ai documenti.
*
* @param event oggetto a cui associare i Document
* @param email email da salvare
* @param cas cas da elaborare
*/
private void addDocuments(final Event event, final Email email, final JCas cas) {
AnnotationIndex<Annotation> aaIdx = cas.getAnnotationIndex(AttachmentAnnotation.type);
FSIterator<Annotation> itAa = aaIdx.iterator();
while (itAa.hasNext()) {
AttachmentAnnotation aa = (AttachmentAnnotation) itAa.next();
Document document = new Document();
document.setAttachmentType(new AttachmentType(2));
document.setGuid(aa.getUrlAttachment());
document.setHashcode(aa.getHashcode());
document.setAuthor(aa.getAuthor());
document.setTemplate(false);
document.setHpmAttachmentId(aa.getId());
document.setName(aa.getAttachmentName());
event.getAttachments().add(document);
// associo il documento alla mail (per il legame EMAIL-DOCUMENT)
if (email != null) {
email.getDocuments().add(document);
}
}
}
/** Outputs Trigrams for an input Annotation. */
public void extractNgramsFromAnnotation(Annotation annotation, JCas aJCas) {
AnnotationIndex<Annotation> tokens = aJCas.getAnnotationIndex(Token.type);
Iterator<Annotation> tIterator = tokens.subiterator(annotation);
Annotation token = null;
Annotation prev = null;
Annotation prev_prev = null;
while (tIterator.hasNext()) {
prev_prev = prev;
prev = token;
token = tIterator.next();
if (tokenIsInAnnotation(annotation, prev) && tokenIsInAnnotation(annotation, prev_prev)) {
NGram ngram = new NGram(aJCas);
FSArray ngArray = new FSArray(aJCas, 3);
ngArray.set(0, prev_prev);
ngArray.set(1, prev);
ngArray.set(2, token);
ngram.setBegin(prev_prev.getBegin());
ngram.setEnd(token.getEnd());
ngram.setElements(ngArray);
ngram.setElementType("edu.cmu.deiis.types.Token");
ngram.setConfidence(1D);
ngram.setCasProcessorId(PROCESSOR_ID);
ngram.addToIndexes();
}
}
}
@Test
public void test() throws Exception {
String html = "<Parent>\n";
html += "<Child1>Some content</Child1>\n";
html += "<Child2 attribute=“someValue” />\n";
html += "<Child3>More content.</Child3>\n";
html += "</Parent>\n";
URL urlA = HtmlAnnotator.class.getClassLoader().getResource("HtmlAnnotator.xml");
if (urlA == null) {
urlA =
HtmlAnnotator.class
.getClassLoader()
.getResource("org/apache/uima/ruta/engine/HtmlAnnotator.xml");
}
URL urlC = HtmlAnnotator.class.getClassLoader().getResource("HtmlConverter.xml");
if (urlC == null) {
urlC =
HtmlAnnotator.class
.getClassLoader()
.getResource("org/apache/uima/ruta/engine/HtmlConverter.xml");
}
XMLInputSource inA = new XMLInputSource(urlA);
ResourceSpecifier specifierA = UIMAFramework.getXMLParser().parseResourceSpecifier(inA);
AnalysisEngine aeA = UIMAFramework.produceAnalysisEngine(specifierA);
aeA.setConfigParameterValue(HtmlAnnotator.PARAM_ONLY_CONTENT, false);
aeA.reconfigure();
XMLInputSource inC = new XMLInputSource(urlC);
ResourceSpecifier specifierC = UIMAFramework.getXMLParser().parseResourceSpecifier(inC);
AnalysisEngine aeC = UIMAFramework.produceAnalysisEngine(specifierC);
aeC.setConfigParameterValue(HtmlConverter.PARAM_SKIP_WHITESPACES, false);
aeC.setConfigParameterValue(HtmlConverter.PARAM_PROCESS_ALL, true);
aeC.setConfigParameterValue(
HtmlConverter.PARAM_GAP_INDUCING_TAGS, new String[] {"child1", "child2", "child3"});
aeC.setConfigParameterValue(HtmlConverter.PARAM_GAP_TEXT, "$");
aeC.reconfigure();
CAS cas = aeA.newCAS();
Type tagType = cas.getTypeSystem().getType(HtmlAnnotator.NAMESPACE + "TAG");
AnnotationIndex<AnnotationFS> ai = null;
FSIterator<AnnotationFS> iterator = null;
cas.setDocumentText(html);
aeA.process(cas);
aeC.process(cas);
CAS plainTextCas = cas.getView(HtmlConverter.DEFAULT_MODIFIED_VIEW);
assertEquals("$Some content$$More content.", plainTextCas.getDocumentText());
ai = plainTextCas.getAnnotationIndex(tagType);
iterator = ai.iterator();
assertEquals(4, ai.size());
assertEquals("$Some content$$More content.", iterator.next().getCoveredText());
assertEquals("$Some content", iterator.next().getCoveredText());
assertEquals("$", iterator.next().getCoveredText());
assertEquals("$More content.", iterator.next().getCoveredText());
cas.release();
}
@Test
public void testExpandOffsets() throws Exception {
String html = "<Parent>\n";
html += "<Child1>Some content</Child1>\n";
html += "<Child2 attribute=“someValue” />\n";
html += "<Child3>More content.</Child3>\n";
html += "</Parent>\n";
URL urlA = HtmlAnnotator.class.getClassLoader().getResource("HtmlAnnotator.xml");
if (urlA == null) {
urlA =
HtmlAnnotator.class
.getClassLoader()
.getResource("org/apache/uima/ruta/engine/HtmlAnnotator.xml");
}
URL urlC = HtmlAnnotator.class.getClassLoader().getResource("HtmlConverter.xml");
if (urlC == null) {
urlC =
HtmlAnnotator.class
.getClassLoader()
.getResource("org/apache/uima/ruta/engine/HtmlConverter.xml");
}
XMLInputSource inA = new XMLInputSource(urlA);
ResourceSpecifier specifierA = UIMAFramework.getXMLParser().parseResourceSpecifier(inA);
AnalysisEngine aeA = UIMAFramework.produceAnalysisEngine(specifierA);
aeA.setConfigParameterValue(HtmlAnnotator.PARAM_ONLY_CONTENT, false);
aeA.reconfigure();
XMLInputSource inC = new XMLInputSource(urlC);
ResourceSpecifier specifierC = UIMAFramework.getXMLParser().parseResourceSpecifier(inC);
AnalysisEngine aeC = UIMAFramework.produceAnalysisEngine(specifierC);
aeC.setConfigParameterValue(HtmlConverter.PARAM_SKIP_WHITESPACES, false);
aeC.setConfigParameterValue(HtmlConverter.PARAM_PROCESS_ALL, true);
aeC.setConfigParameterValue(HtmlConverter.PARAM_EXPAND_OFFSETS, true);
aeC.reconfigure();
CAS cas = aeA.newCAS();
Type tagType = cas.getTypeSystem().getType(HtmlAnnotator.NAMESPACE + "TAG");
Feature expandedFeature = tagType.getFeatureByBaseName("expandedOffsets");
AnnotationIndex<AnnotationFS> ai = null;
FSIterator<AnnotationFS> iterator = null;
cas.setDocumentText(html);
aeA.process(cas);
aeC.process(cas);
CAS plainTextCas = cas.getView(HtmlConverter.DEFAULT_MODIFIED_VIEW);
assertEquals("Some contentMore content.", plainTextCas.getDocumentText());
ai = plainTextCas.getAnnotationIndex(tagType);
iterator = ai.iterator();
assertEquals(4, ai.size());
AnnotationFS next = null;
next = iterator.next();
assertEquals(false, next.getBooleanValue(expandedFeature));
assertEquals("Some contentMore content.", next.getCoveredText());
next = iterator.next();
assertEquals(false, next.getBooleanValue(expandedFeature));
assertEquals("Some content", next.getCoveredText());
next = iterator.next();
boolean b1 = next.getBooleanValue(expandedFeature);
assertEquals("More content.", next.getCoveredText());
next = iterator.next();
boolean b2 = next.getBooleanValue(expandedFeature);
assertEquals("More content.", next.getCoveredText());
// for one of these two annotation (with same offsets) the feature must be set to true
assertEquals(true, b1 || b2);
cas.release();
}