与 Gate ANNIE 一起学习语法

Question

你好，我从事信息检索工作已经有一段时间了，但遇到了一些困难。 最近，我从以下链接下载了 StandAloneAnnie.java

http:// /gate.ac.uk/wiki/code-repository/src/sheffield/examples/StandAloneAnnie.java 尽管我已经能够执行它并查看输出，但我有一两个查询。

1. 此程序注释人员和位置，其中存储用于注释此类实体的语法。

2. 如何编写自己的简单语法来提取一些数据并在我的 StandAloneAnnie.java 副本中使用它？

以前的帖子 一个字符串上有数百个正则表达式 NLP 新手，关于注释的问题

Answer 1

以下是标记人身高的简单语法

Phase: Meaurements
Input: Token Number 
Options: control=appelt debug=true



Rule: Height
(
({Number})
( {Token.string=~"[Ff]t"} | {Token.string=~"[Ii]n"} | {Token.string=~"[Cc]m"})
):height
-->
:height.Height= {value= :height.Number.value, unit= :height.Token.string}

这是执行的主要代码，

    public static void main(String arg[]) {

            Gate.init();
            gate.Corpus corpus= (Corpus) Factory.createResource("gate.corpora.CorpusImpl");

//You need to register the plugin before you load it.

            Gate.getCreoleRegister().registerDirectories(new File(Gate.getPluginsHome(), ANNIEConstants.PLUGIN_DIR).toURI().toURL());
            Gate.getCreoleRegister().registerDirectories(new URL("file:///GATE_HOME/plugins/Tagger_Numbers"));//change this path


            Document doc = new DocumentImpl();
//The string to be annotated.

String str = "Height is 60 in. Weight is 150 lbs pulse rate 90 Pulse rate 90";
DocumentContentImpl impl = new DocumentContentImpl(str);
doc.setContent(impl);

//Loading processing resources. refer http://gate.ac.uk/gate/doc/plugins.html for what class the plugin belongs to

            ProcessingResource token = (ProcessingResource) Factory.createResource("gate.creole.tokeniser.DefaultTokeniser", Factory.newFeatureMap());
            ProcessingResource sspliter = (ProcessingResource) Factory.createResource("gate.creole.splitter.SentenceSplitter", Factory.newFeatureMap());
            ProcessingResource number = (ProcessingResource) Factory.createResource("gate.creole.numbers.NumbersTagger", Factory.newFeatureMap());


/*pipeline is an application that needs to be created to use resources loaded above.
Reasources must be added in a particular order eg. below the 'number' resource requires the document to be tokenised. */

corpus.add(doc);
SerialAnalyserController pipeline = (SerialAnalyserController) Factory.createResource("gate.creole.SerialAnalyserController", Factory.newFeatureMap(), Factory.newFeatureMap(), "ANNIE");
pipeline.setCorpus(corpus);
pipeline.add(token);
pipeline.add(sspliter);
pipeline.add(number);
pipeline.execute();

//Extract info from an annotated document.

AnnotationSetImpl ann=(AnnotationSetImpl)doc.getAnnotations();
Iterator<Annotation>i = ann.get(vital).iterator();
Annotation annotation = i.next();
long start = annotation.getStartNode().getOffset();
long end =  annotation.getEndNode().getOffset();
System.out.println(doc.toString().substring((int)start, (int)end));

}

注意：-
在上面的代码中，Height 的语法将写入 .jape 文件中。您需要使用 JAPE(JAPE Plus) 转换器来运行此语法。我们只需要在主代码中执行应用程序（'pipeline'）。您可以在 gateway.ac.uk/sale/tao 找到编写 jape 的教程

Answer 2

有一个 Annie 简介 幻灯片，解释了语法是如何存储的。它们位于 Jape 文件中。