在GATE中JAPE规则内使用地名录作为词典

Question

我有这样的场景：

000.000.0001.000 VALUE1

000.000.0002.000 VALUE2

...

000.010.0001.000 VALUE254

SK1 | SK2 | SK3 | SK4

000  | 000 | 0001 | 000

000

000

0001

000

还有其他（更好的）方法来做我需要做的事情吗？

多谢.

Answer 1

我使用GazetteerList类找到了解决问题的方法，下面是以下代码片段：

//Gazetteer object
GazetteerList gazList = new GazetteerList() ;
//Object to map gazetteers entries and their positions in the list
//i.e.: 000.000.0001.000 -> 1,3
//This is because, in my case, the same key 
//can appear more than once in the gazetteer
HashMap<String, ArrayList<Integer>> keyMap = 
                                   new HashMap<String, ArrayList<Integer>>();
  try{
    gazList.setMode(GazetteerList.LIST_MODE);
    gazList.setSeparator("\t");
    gazList.setURL(
        new URL("file:/path/to/gazetteer/gazetteer_list_file.lst"));
    gazList.load();

    //Here is the mapping between the keys and their position
    int pos = 0;
    for( GazetteerNode gazNode : gazList){
      if(keyMap.get(gazNode.getEntry()) == null)
        keyMap.put(gazNode.getEntry(), new ArrayList<Integer>());

      keyMap.get(gazNode.getEntry()).add(pos);
      pos++;
    }

  } catch (MalformedURLException ex){
    System.out.println(ex);
  } catch (ResourceInstantiationException ex){
    System.out.println(ex);
  }

然后，可以在地图中查找匹配的键并获取其特征：

  for(Integer index : keyMap.get(key)){
      FeatureMap fmap = toFeatureMap(gazList.get(index).getFeatureMap());
      fmap.put("additionalFeature", "feature");
      outputAS.add(startOffset, endOffset, "Lookup", fmap);
  }