Examples of org.apache.uima.ruta.textruler.core.TextRulerExampleDocument

org.apache.uima.ruta.textruler.core.TextRulerExampleDocument

TextRulerExampleDocument stands for one document usually loaded from an XMI file. It uses the given CasCache for storing its CAS with the XMI filename as the key. It holds ArrayLists for positive and negative MLExamples which can be filled on demand for a given learning target. E.g. single slot algorithms learn rules for each slot separately, so the work-flow is to clear the current examples and create new for the next slot target. The same is with single slot boundary algorithms like LP2: It first creates all left boundary examples, learns from them, clears the examples and creates the right boundary examples and so on. This class also provides the functionality extract and created MLExmaples of a given document or test CAS for a given TextRulerTarget. Especially for boundary algorithms you can call createBoundaryAnnotationsForCas to get boundary annotations at the beginnings and endings of an example slot. Caution (this is quite a bit inconvenient at the moment!): If a CAS gets loaded from the casCache, you have to call createBoundaryAnnotationsForCas again, so your casLoader must be aware of that (see BasicLP2 for an example) ! hint: this could be renamed to MLDocument instead of TextRulerExampleDocument ?

    interTupelSeparatorsCache.clear();
    TextRulerToolkit.log("-- WIEN END");
  }


  protected boolean findRightPatterns() {
    TextRulerExampleDocument doc = exampleDocuments.getDocuments().get(0);
    boolean allFound = true;
    for (int k = 0; k < slotNames.length; k++) {
      List<TextRulerRulePattern> rightContexts = getRightContextForSlot(doc, k);
      System.out.println(rightContexts.get(0));
      int shortest = Integer.MAX_VALUE;

    }
    return allFound;
  }


  protected boolean findLeftPatterns() {
    TextRulerExampleDocument doc = exampleDocuments.getDocuments().get(0);
    // skip l 0 !
    if (slotNames.length < 2)
      return true;
    boolean allFound = true;
    for (int k = 1; k < slotNames.length; k++) {

    return allFound;
  }


  protected boolean findHeadTailAndL1Patterns() {
    List<TextRulerExampleDocument> docs = exampleDocuments.getDocuments();
    TextRulerExampleDocument doc0 = docs.get(0);
    TextRulerRulePattern head = new TextRulerRulePattern();
    TextRulerRulePattern tail = new TextRulerRulePattern();
    getPageHeadAndTailPortion(doc0, head, tail);


    final class HLCandidate {

Examples of org.apache.uima.ruta.textruler.core.TextRulerExampleDocument

Related Classes of org.apache.uima.ruta.textruler.core.TextRulerExampleDocument