@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);
}
}
}
}
/**
* Removes all annotations of type removeAnnotationType which are contained by annotations of type
* containerAnnotationType.
*
* @param cas
* @param containerAnnotationType
* @param removeAnnotationType
*/
public static void removeAnnotations(
CAS cas, AnnotationFS containerAnnotation, Type removeAnnotationType) {
FSIndex<AnnotationFS> allRemoveAnnotations = cas.getAnnotationIndex(removeAnnotationType);
ContainingConstraint containingConstraint = new ContainingConstraint(containerAnnotation);
Iterator<AnnotationFS> containingTokens =
cas.createFilteredIterator(allRemoveAnnotations.iterator(), containingConstraint);
Collection<AnnotationFS> removeAnnotations = new LinkedList<AnnotationFS>();
while (containingTokens.hasNext()) {
removeAnnotations.add(containingTokens.next());
}
for (Iterator<AnnotationFS> it = removeAnnotations.iterator(); it.hasNext(); ) {
cas.removeFsFromIndexes(it.next());
}
}
@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();
}
@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();
}
/** Copied and modified from {@link org.apache.uima.util.CasToInlineXml} */
private static String toXML(CAS cas, AnnotationsToElements converter) throws SAXException {
ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream();
XMLSerializer sax2xml = new XMLSerializer(byteArrayOutputStream, false);
// get document text
String docText = cas.getDocumentText();
char[] docCharArray = docText.toCharArray();
// get iterator over annotations sorted by increasing start position and
// decreasing end position
FSIterator<AnnotationFS> iterator = cas.getAnnotationIndex().iterator();
// This is basically a recursive algorithm that has had the recursion
// removed through the use of an explicit Stack. We iterate over the
// annotations, and if an annotation contains other annotations, we
// push the parent annotation on the stack, process the children, and
// then come back to the parent later.
List<AnnotationFS> stack = new ArrayList<AnnotationFS>();
int pos = 0;
ContentHandler handler = sax2xml.getContentHandler();
handler.startDocument();
// write the start tag
converter.startRootElement(handler);
// now use null is a placeholder for this artificial Document annotation
AnnotationFS curAnnot = null;
while (iterator.isValid()) {
AnnotationFS nextAnnot = iterator.get();
if (curAnnot == null || nextAnnot.getBegin() < curAnnot.getEnd()) {
// nextAnnot's start point is within the span of curAnnot
if (curAnnot == null || nextAnnot.getEnd() <= curAnnot.getEnd()) // crossover span check
{
// nextAnnot is contained within curAnnot
// write text between current pos and beginning of nextAnnot
try {
handler.characters(docCharArray, pos, nextAnnot.getBegin() - pos);
pos = nextAnnot.getBegin();
converter.startAnnotationElement(nextAnnot, handler);
// push parent annotation on stack
stack.add(curAnnot);
// move on to next annotation
curAnnot = nextAnnot;
} catch (StringIndexOutOfBoundsException e) {
System.err.println(
"Invalid annotation range: "
+ nextAnnot.getBegin()
+ ","
+ nextAnnot.getEnd()
+ " in document of length "
+ docText.length());
}
}
iterator.moveToNext();
} else {
// nextAnnot begins after curAnnot ends
// write text between current pos and end of curAnnot
try {
handler.characters(docCharArray, pos, curAnnot.getEnd() - pos);
pos = curAnnot.getEnd();
} catch (StringIndexOutOfBoundsException e) {
System.err.println(
"Invalid annotation range: "
+ curAnnot.getBegin()
+ ","
+ curAnnot.getEnd()
+ " in document of length "
+ docText.length());
}
converter.endAnnotationElement(curAnnot, handler);
// pop next containing annotation off stack
curAnnot = stack.remove(stack.size() - 1);
}
}
// finished writing all start tags, now finish up
if (curAnnot != null) {
try {
handler.characters(docCharArray, pos, curAnnot.getEnd() - pos);
pos = curAnnot.getEnd();
} catch (StringIndexOutOfBoundsException e) {
System.err.println(
"Invalid annotation range: "
+ curAnnot.getBegin()
+ ","
+ curAnnot.getEnd()
+ "in document of length "
+ docText.length());
}
converter.endAnnotationElement(curAnnot, handler);
while (!stack.isEmpty()) {
curAnnot = stack.remove(stack.size() - 1); // pop
if (curAnnot == null) {
break;
}
try {
handler.characters(docCharArray, pos, curAnnot.getEnd() - pos);
pos = curAnnot.getEnd();
} catch (StringIndexOutOfBoundsException e) {
System.err.println(
"Invalid annotation range: "
+ curAnnot.getBegin()
+ ","
+ curAnnot.getEnd()
+ "in document of length "
+ docText.length());
}
converter.endAnnotationElement(curAnnot, handler);
}
}
if (pos < docCharArray.length) {
handler.characters(docCharArray, pos, docCharArray.length - pos);
}
converter.endRootElement(handler);
handler.endDocument();
// return XML string
return new String(byteArrayOutputStream.toByteArray());
}