How do you use the POS tagger?

[frankandrobot]

No example was given and I don't see the source code here.

[benson-basis]

package com.basistech.dp.nlp4j.impl;

import com.basistech.rosette.dm.AnnotatedText;
import com.basistech.rosette.dm.Annotator;
import com.basistech.rosette.dm.ListAttribute;
import com.basistech.rosette.dm.MorphoAnalysis;
import com.basistech.rosette.dm.Token;
import com.google.common.io.Resources;
import edu.emory.mathcs.nlp.component.template.node.NLPNode;
import edu.emory.mathcs.nlp.decode.NLPDecoder;

import java.io.IOException;
import java.io.InputStream;
import java.net.URL;
import java.util.ArrayList;
import java.util.List;
import java.util.stream.Collectors;
import java.util.stream.IntStream;

/**
 *
 */
public class Nlp4jTaggerAnnotator implements Annotator {

    // the interface here wants one parse-result per sentence.
    // so we respect sentence boundaries in the input, and nothing else.
    private final NLPDecoder decoder;

    public Nlp4jTaggerAnnotator() {
        URL configUrl = Resources.getResource("config-decode-tagger-en.xml");
        try (InputStream configInputStream = Resources.asByteSource(configUrl).openStream()) {
            decoder = new NLPDecoder(configInputStream);
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
    }
    @Override
    public AnnotatedText annotate(CharSequence input) {
        return annotate(new AnnotatedText.Builder().data(input).build());
    }

    @Override
    public AnnotatedText annotate(AnnotatedText input) {
        List<NLPNode> nodes = new ArrayList<>();
        NLPNode rootNode = new NLPNode();
        rootNode.toRoot();
        nodes.add(rootNode);
        nodes.addAll(
                IntStream.range(0, input.getTokens().size())
                        .mapToObj(i -> createNode(input.getTokens().get(i), i + 1)).collect(Collectors.toList()));
        // Nlp4J likes arrays
        NLPNode[] nodesArray = nodes.toArray(new NLPNode[nodes.size()]);
        decoder.decode(nodesArray);

        AnnotatedText.Builder builder = new AnnotatedText.Builder(input);
        ListAttribute.Builder<Token> tlBuilder = new ListAttribute.Builder<>(Token.class);
        nodes.stream().skip(1).forEach(n -> tlBuilder.add(makeToken(n)));
        builder.tokens(tlBuilder.build());
        return builder.build();
    }

    private Token makeToken(NLPNode node) {
        Token.Builder builder = new Token.Builder(node.getStartOffset(), node.getEndOffset(), node.getWordForm());
        MorphoAnalysis.Builder maBuilder = new MorphoAnalysis.Builder();
        maBuilder.partOfSpeech(node.getPartOfSpeechTag());
        if (node.getLemma() != null) {
            maBuilder.lemma(node.getLemma());
        }
        builder.addAnalysis(maBuilder.build());
        if (node.getWordClusters() != null) {
            builder.extendedProperty("word_clusters", node.getWordClusters());
        }
        return builder.build();
    }

    private NLPNode createNode(com.basistech.rosette.dm.Token token, int index) {
        NLPNode node = decoder.create();
        node.setWordForm(token.getText());
        node.setStartOffset(token.getStartOffset());
        node.setEndOffset(token.getEndOffset());
        node.setID(index);
        return node;
    }
}

[frankandrobot]

Yea, thanks I ended up figuring it out by looking at a demo:

import edu.emory.mathcs.nlp.decode.DecodeConfig
import edu.emory.mathcs.nlp.common.util.IOUtils
import edu.emory.mathcs.nlp.component.template.node.NLPNode
import edu.emory.mathcs.nlp.decode.NLPDecoder

fun posTagger(string : String) : Array<NLPNode> {
    val configUri = "src/main/resources/nlp4j/config.xml"
    val config = DecodeConfig(IOUtils.createFileInputStream(configUri))
    val decoder = NLPDecoder(config)

    return decoder.decode(string)