Examples with NerSentence zemberek.ner.NerSentence used on opensource projects

Example 1 with NerSentence

use of zemberek.ner.NerSentence in project zemberek-nlp by ahmetaa.

the class FindNamedEntities method run.

@Override
public void run() throws Exception {
    initializeOutputDir();
    IOUtil.checkDirectoryArgument(modelRoot, "Model Root");
    IOUtil.checkFileArgument(inputPath, "Input File");
    Path out = outDir.resolve(inputPath.toFile().getName() + ".ne");
    List<String> lines = Files.readAllLines(inputPath, StandardCharsets.UTF_8);
    List<String> sentences = TurkishSentenceExtractor.DEFAULT.fromParagraphs(lines);
    Log.info("There are %d lines and about %d sentences", lines.size(), sentences.size());
    TurkishMorphology morphology = TurkishMorphology.createWithDefaults();
    PerceptronNer ner = PerceptronNer.loadModel(modelRoot, morphology);
    Stopwatch sw = Stopwatch.createStarted();
    int tokenCount = 0;
    try (PrintWriter pw = new PrintWriter(out.toFile(), "UTF-8")) {
        for (String sentence : sentences) {
            sentence = TextUtil.normalizeApostrophes(sentence);
            sentence = TextUtil.normalizeQuotesHyphens(sentence);
            sentence = TextUtil.normalizeSpacesAndSoftHyphens(sentence);
            List<String> words = TurkishTokenizer.DEFAULT.tokenizeToStrings(sentence);
            tokenCount += words.size();
            NerSentence result = ner.findNamedEntities(sentence, words);
            pw.println(result.getAsTrainingSentence(annotationStyle));
        }
    }
    double secs = sw.elapsed(TimeUnit.MILLISECONDS) / 1000d;
    Log.info("Token count = %s", tokenCount);
    Log.info("File processed in %.4f seconds.", secs);
    Log.info("Speed = %.2f tokens/sec", tokenCount / secs);
    Log.info("Result is written in %s", out);
}

Example 2 with NerSentence

use of zemberek.ner.NerSentence in project zemberek-nlp by ahmetaa.

the class UseNer method main.

public static void main(String[] args) throws IOException {
    // assumes you generated a model in my-model directory.
    Path modelRoot = Paths.get("my-model");
    TurkishMorphology morphology = TurkishMorphology.createWithDefaults();
    PerceptronNer ner = PerceptronNer.loadModel(modelRoot, morphology);
    String sentence = "Ali Kaan yarın İstanbul'a gidecek.";
    NerSentence result = ner.findNamedEntities(sentence);
    List<NamedEntity> namedEntities = result.getNamedEntities();
    for (NamedEntity namedEntity : namedEntities) {
        System.out.println(namedEntity);
    }
}