1

我将 GATE API 与 java 代码一起使用，并尝试在文档文本上运行已知的 JAPE 规则之一，但不幸的是我无法获得适当的结果。我的代码如下：

public void initAnnie() throws GateException, IOException {
    Out.prln("Initialising ANNIE...");

    // load the ANNIE application from the saved state in plugins/ANNIE
    File pluginsHome = Gate.getPluginsHome();
    File anniePlugin = new File(pluginsHome, "ANNIE");
    File annieGapp = new File(anniePlugin, "ANNIE_with_defaults.gapp");
    annieController = (CorpusController) PersistenceManager
            .loadObjectFromFile(annieGapp);

    Out.prln("...ANNIE loaded");
} // initAnnie()

/** Tell ANNIE's controller about the corpus you want to run on */
public void setCorpus(Corpus corpus) {
    annieController.setCorpus(corpus);
} // setCorpus

/** Run ANNIE */
public void execute() throws GateException {
    Out.prln("Running ANNIE...");
    annieController.execute();
    Out.prln("...ANNIE complete");
} // execute()

/**
 * Run from the command-line, with a list of URLs as argument.
 * <P>
 * <B>NOTE:</B><BR>
 * This code will run with all the documents in memory - if you want to
 * unload each from memory after use, add code to store the corpus in a
 * DataStore.
 */
public static void main(String args[]) throws GateException, IOException {
// initialise the GATE library
Out.prln("Initialising GATE...");
Gate.init();
Out.prln("...GATE initialised");
// load ANNIE plugin - you must do this before you can create tokeniser
// or JAPE transducer resources.
Gate.getCreoleRegister().registerDirectories(
new File(Gate.getPluginsHome(), "ANNIE").toURI().toURL());

 // Build the pipeline
  SerialAnalyserController pipeline =
 (SerialAnalyserController)Factory.createResource(
   "gate.creole.SerialAnalyserController");
  LanguageAnalyser tokeniser = (LanguageAnalyser)Factory.createResource(
  "gate.creole.tokeniser.DefaultTokeniser");
LanguageAnalyser jape = (LanguageAnalyser)Factory.createResource(
 "gate.creole.Transducer", gate.Utils.featureMap(
     "grammarURL", new     
 File("C:path/to/univerity_rules.jape").toURI().toURL(),
   "encoding", "UTF-8")); // ensure this matches the file
pipeline.add(tokeniser);
pipeline.add(jape);

// create document and corpus
// create a GATE corpus and add a document for each command-line
// argument
Corpus corpus = Factory.newCorpus("JAPE corpus");

 URL u = new URL("file:/path/to/Document.txt");
 FeatureMap params = Factory.newFeatureMap();
 params.put("sourceUrl", u);
 params.put("preserveOriginalContent", new Boolean(true));
 params.put("collectRepositioningInfo", new Boolean(true));
 Out.prln("Creating doc for " + u);
 Document doc = (Document)
   Factory.createResource("gate.corpora.DocumentImpl", params);
 corpus.add(doc);
 pipeline.setCorpus(corpus);

// run it
pipeline.execute();

// extract results
System.out.println("Found annotations of the following types: " +
  doc.getAnnotations().getAllTypes());


} // main

 }

使用的 JAPE 规则如下：

Phase:firstpass 
Input: Lookup Token 

//note that we are using Lookup and Token both inside our rules. 
Options: control = appelt


Rule: University1 
Priority: 20
(
  {Token.string == "University"} 
  {Token.string == "of"}
  {Lookup.minorType == city} 
):orgName 
-->
:orgName.Organisation = 
  {kind = "university", rule = "University1"}

最后我得到的结果如下：

 Initialising GATE...
 log4j:WARN No appenders could be found for logger (gate.Gate).
 log4j:WARN Please initialize the log4j system properly.
 log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
  ...GATE initialised
 Creating doc for file:path/to/Document.txt
 Found annotations of the following types: [SpaceToken, Token]

请任何帮助

4

1 回答 1

2

问题是您没有尝试在 JAPE 程序中使用的“查找”注释。

您需要添加 2 个额外的资源：

    LanguageAnalyser gazetter = (LanguageAnalyser)Factory.createResource(
            "gate.creole.gazetteer.DefaultGazetteer");
    LanguageAnalyser splitter = (LanguageAnalyser)Factory.createResource(
            "gate.creole.splitter.SentenceSplitter");

您的处理资源应按以下顺序运行：

    pipeline.add(tokeniser);
    pipeline.add(gazetter);
    pipeline.add(splitter); 
    pipeline.add(jape);

Gazetterr 将创建“查找”注释。

句子拆分器将停止创建跨越两个句子的“组织”注释。

已经过测试，它对我有用。

...GATE initialised
Creating doc for file:/Users/andreyshafirin/tmp/testdoc.txt
Found annotations of the following types: [Lookup, Organisation, Token, Split, SpaceToken, Sentence]

PS：

我认为有一种更好的方法可以从 Java 代码中使用 GATE。您可以在 GATE Developer 中创建应用程序，对其进行自定义并将其保存到文件中（您将在此处找到方法）。然后，您可以从您的 java 代码中加载 GATE 应用程序（请参阅此示例，以及更多其他示例以了解如何操作）。这样您就不必担心与处理资源属性相关的大量细节和特性（您将在 GUI 中定义和更改它们）。

祝你好运。

于 2014-04-25T13:59:32.943 回答