/**
* abnerNER would analyze words and give confidence
*
* @param args sentences to be processed
* @param arg0 information input
* @throws AnalysisEngineProcessException
*/
public static void abnerNER(String[] args, JCas arg0) {
HashMap<String, Double> ConfMap = new HashMap<String, Double>();
/** use abner to find gene names from words */
Tagger t = new Tagger();
for (int i = 1; i < args.length; ++i) {
String s = args[i];
String[][] ents = t.getEntities(s);
/** use HashMap to store words selected by Abner */
for (int j = 0; j < ents[0].length; j++) {
ConfMap.put(ents[0][j], 1.0);
}
}
// TODO Auto-generated method stub
FSIterator<org.apache.uima.jcas.tcas.Annotation> ite =
arg0.getAnnotationIndex(NameTag.type).iterator();
while (ite.hasNext()) {
/** get the words selected by LingPipe */
String name = ((NameTag) ite.get()).getText();
/** set the confidence for words selected by both LingPipe and Abner as 1 */
if (ConfMap.containsKey(name)) {
((NameTag) ite.get()).setConfidenceAbner(1.0);
} else {
((NameTag) ite.get()).setConfidenceAbner(0.0);
}
ite.next();
}
}
private static String getResult(JCas jcas, String sourceLang, String interLang) {
FSIterator iter = jcas.getAnnotationIndex(Target.type).iterator();
String result = "";
while (iter.isValid()) {
FeatureStructure fs = iter.get();
Target transText = (Target) fs;
result = transText.getContent();
iter.moveToNext();
}
return result;
}
public static void main(String[] args)
throws IOException, InvalidXMLException, CASException, ResourceInitializationException,
SAXException {
JCas jcas = null;
// Leer el descriptor del anotador
XMLParser xmlParser = UIMAFramework.getXMLParser();
XMLInputSource in = new XMLInputSource("desc/ej4/MetricsAnnotatorPipeline.xml");
// Crear un AE en base al descriptor
AnalysisEngineDescription tsDesc = xmlParser.parseAnalysisEngineDescription(in);
// Obtener el CAS
jcas = CasCreationUtils.createCas(tsDesc).getJCas();
if (jcas != null) {
// De-serializar la anotacion de un fichero
FileInputStream inputStream = null;
inputStream = new FileInputStream("resources/annotation.xmi");
XmiCasDeserializer.deserialize(inputStream, jcas.getCas());
// Obtener el texto de la anotacion
String sofaString = jcas.getDocumentText();
System.out.println(sofaString);
// Usar las anotaciones del fichero
FSIterator it = jcas.getAnnotationIndex(Metric.type).iterator();
while (it.isValid()) {
Metric metric = (Metric) it.get();
Number number = metric.getNumber();
Unit unit = metric.getUnit();
Double value =
(number.getIsDouble())
? number.getAbsoluteDoubleValue()
: Double.valueOf(number.getAbsoluteIntegerValue());
System.out.println("===================");
System.out.println("Metric: " + metric.getCoveredText());
System.out.println("Real value: " + value * number.getSign() * unit.getMultiplier());
System.out.println("Base unit: " + unit.getBaseUnit());
it.moveToNext();
}
}
}
/**
* CasConsumer would use tags and features to write output file, evaluate and print precision,
* recall and F-1 measure.
*
* @param arg0
* @throws ResourceProcessException
*/
@Override
public void processCas(CAS arg0) throws ResourceProcessException {
/** convert type of arg0 */
JCas jcas = null;
try {
jcas = arg0.getJCas();
} catch (CASException e1) {
// TODO Auto-generated catch block
e1.printStackTrace();
}
// TODO Auto-generated method stub
FSIterator<Annotation> ite = jcas.getAnnotationIndex(WordTag.type).iterator();
while (ite.hasNext()) {
/** collect features */
String id = ((WordTag) ite.get()).getId();
int begin = ((WordTag) ite.get()).getBegin0();
int end = ((WordTag) ite.get()).getEnd0();
String name = ((WordTag) ite.get()).getName();
/** organize string for output */
report.append(id);
report.append("|");
report.append(begin);
report.append(" ");
report.append(end);
report.append("|");
report.append(name);
report.append("\n");
/** count the length of output string */
count++;
ite.next();
}
result = report.toString();
File sampleOut = new File("src/main/resources/data/sample.out");
try {
testRecall = FileUtils.file2String(sampleOut);
} catch (IOException e1) {
// TODO Auto-generated catch block
e1.printStackTrace();
}
/** split strings from file into sentences */
String[] resultSplit = result.split("\n");
String[] recallSplit = testRecall.split("\n");
PrecisionRecallCalculator(recallSplit, resultSplit);
/** write the output file to the project root */
String path = "hw1-longh.out";
File dirFile = new File(path);
/** make sure no conflict */
if (dirFile.exists()) {
dirFile.delete();
}
try {
/** write file */
BufferedWriter bw1 = new BufferedWriter(new FileWriter(path, true));
bw1.write(report.toString());
bw1.flush();
bw1.close();
} catch (IOException e) {
e.printStackTrace();
}
}
/** 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());
}
/**
* Called when the processing of a Document is completed. <br>
* The process status can be looked at and corresponding actions taken.
*
* @param aCas CAS corresponding to the completed processing
* @param aStatus EntityProcessStatus that holds the status of all the events for aEntity
*/
public void entityProcessComplete(CAS aCas, EntityProcessStatus aStatus) {
if (aStatus != null) {
if (aStatus.isException()) {
System.err.println("Error on process CAS call to remote service:");
List exceptions = aStatus.getExceptions();
for (int i = 0; i < exceptions.size(); i++) {
((Throwable) exceptions.get(i)).printStackTrace();
}
if (!ignoreErrors) {
System.err.println("Terminating Client...");
stop();
}
}
if (logCas) {
String ip = "no IP";
List eList = aStatus.getProcessTrace().getEventsByComponentName("UimaEE", false);
for (int e = 0; e < eList.size(); e++) {
ProcessTraceEvent event = (ProcessTraceEvent) eList.get(e);
if (event.getDescription().equals("Service IP")) {
ip = event.getResultMessage();
}
}
String casId = ((UimaASProcessStatus) aStatus).getCasReferenceId();
if (casId != null) {
long current = System.nanoTime() / 1000000 - mStartTime;
if (casMap.containsKey(casId)) {
Object value = casMap.get(casId);
if (value != null && value instanceof Long) {
long start = ((Long) value).longValue();
System.out.println(ip + "\t" + start + "\t" + (current - start));
}
}
}
} else {
System.out.print(".");
if (0 == (entityCount + 1) % 50) {
System.out.print((entityCount + 1) + " processed\n");
}
}
}
// if output dir specified, dump CAS to XMI
if (outputDir != null) {
// try to retrieve the filename of the input file from the CAS
File outFile = null;
Type srcDocInfoType =
aCas.getTypeSystem().getType("org.apache.uima.examples.SourceDocumentInformation");
if (srcDocInfoType != null) {
FSIterator it = aCas.getIndexRepository().getAllIndexedFS(srcDocInfoType);
if (it.hasNext()) {
FeatureStructure srcDocInfoFs = it.get();
Feature uriFeat = srcDocInfoType.getFeatureByBaseName("uri");
Feature offsetInSourceFeat = srcDocInfoType.getFeatureByBaseName("offsetInSource");
String uri = srcDocInfoFs.getStringValue(uriFeat);
int offsetInSource = srcDocInfoFs.getIntValue(offsetInSourceFeat);
File inFile;
try {
inFile = new File(new URL(uri).getPath());
String outFileName = inFile.getName();
if (offsetInSource > 0) {
outFileName += ("_" + offsetInSource);
}
outFileName += ".xmi";
outFile = new File((String) outputDir, outFileName);
} catch (MalformedURLException e1) {
// invalid URI, use default processing below
}
}
}
if (outFile == null) {
outFile = new File((String) outputDir, "doc" + entityCount);
}
try {
FileOutputStream outStream = new FileOutputStream(outFile);
try {
XmiCasSerializer.serialize(aCas, outStream);
} finally {
outStream.close();
}
} catch (Exception e) {
System.err.println("Could not save CAS to XMI file");
e.printStackTrace();
}
}
// update stats
entityCount++;
String docText = aCas.getDocumentText();
if (docText != null) {
size += docText.length();
}
// Called just before sendCas with next CAS from collection reader
}