Examples with zemberek.morphology._analyzer._SingleAnalysis zemberek.morphology._analyzer._SingleAnalysis used on opensource projects

Example 1 with zemberek.morphology._analyzer._SingleAnalysis

use of zemberek.morphology._analyzer._SingleAnalysis in project zemberek-nlp by ahmetaa.

the class _SingleAnalysis method fromSearchPath.

// Here we generate a _SingleAnalysis from a search path.
public static _SingleAnalysis fromSearchPath(SearchPath searchPath) {
    List<MorphemeSurface> morphemes = new ArrayList<>(searchPath.transitions.size());
    int derivationCount = 0;
    for (SurfaceTransition transition : searchPath.getTransitions()) {
        if (transition.isDerivative()) {
            derivationCount++;
        }
        Morpheme morpheme = transition.getMorpheme();
        // if empty, use the cache.
        if (transition.surface.isEmpty()) {
            MorphemeSurface suffixSurface = emptyMorphemeCache.get(morpheme);
            if (suffixSurface == null) {
                suffixSurface = new MorphemeSurface(morpheme, "");
                emptyMorphemeCache.put(morpheme, suffixSurface);
            }
            morphemes.add(suffixSurface);
            continue;
        }
        MorphemeSurface suffixSurface = new MorphemeSurface(morpheme, transition.surface);
        morphemes.add(suffixSurface);
    }
    int[] groupBoundaries = new int[derivationCount + 1];
    // we assume there is always an IG
    groupBoundaries[0] = 0;
    int morphemeCounter = 0, derivationCounter = 1;
    for (SurfaceTransition transition : searchPath.getTransitions()) {
        if (transition.isDerivative()) {
            groupBoundaries[derivationCounter] = morphemeCounter;
            derivationCounter++;
        }
        morphemeCounter++;
    }
    return new _SingleAnalysis(searchPath.getDictionaryItem(), morphemes, groupBoundaries);
}

Example 2 with zemberek.morphology._analyzer._SingleAnalysis

use of zemberek.morphology._analyzer._SingleAnalysis in project zemberek-nlp by ahmetaa.

the class _SingleAnalysisTest method morphemeGroupTest.

@Test
public void morphemeGroupTest() {
    InterpretingAnalyzer analyzer = getAnalyzer("kitap");
    _SingleAnalysis analysis = analyzer.analyze("kitaplarda").get(0);
    MorphemeGroup group = analysis.getGroup(0);
    Assert.assertEquals("kitaplarda", group.surface());
    analyzer = getAnalyzer("okumak");
    analysis = analyzer.analyze("okutmuyor").get(0);
    Assert.assertEquals(2, analysis.getMorphemeGroupCount());
    MorphemeGroup group0 = analysis.getGroup(0);
    Assert.assertEquals("oku", group0.surface());
    MorphemeGroup group1 = analysis.getGroup(1);
    Assert.assertEquals("tmuyor", group1.surface());
}

Example 3 with zemberek.morphology._analyzer._SingleAnalysis

use of zemberek.morphology._analyzer._SingleAnalysis in project zemberek-nlp by ahmetaa.

the class AnalyzerTestBase method lastMorphemeIs.

boolean lastMorphemeIs(_SingleAnalysis result, String morphemeName) {
    List<MorphemeSurface> morphemes = result.getMorphemesSurfaces();
    if (morphemes.size() == 0) {
        return false;
    }
    MorphemeSurface last = morphemes.get(morphemes.size() - 1);
    return last.morpheme.id.equalsIgnoreCase(morphemeName);
}

Example 4 with zemberek.morphology._analyzer._SingleAnalysis

use of zemberek.morphology._analyzer._SingleAnalysis in project zemberek-nlp by ahmetaa.

the class InterpretingAnalyzer method analyze.

public List<_SingleAnalysis> analyze(String input, AnalysisDebugData debugData) {
    // get stem candidates.
    List<StemTransition> candidates = Lists.newArrayListWithCapacity(3);
    for (int i = 1; i <= input.length(); i++) {
        String stem = input.substring(0, i);
        candidates.addAll(getMatchingStemTransitions(stem));
    }
    if (debugData != null) {
        debugData.input = input;
        debugData.candidateStemTransitions.addAll(candidates);
    }
    // generate initial search paths.
    List<SearchPath> paths = new ArrayList<>();
    for (StemTransition candidate : candidates) {
        int length = candidate.surface.length();
        String head = input.substring(0, length);
        String tail = input.substring(length);
        paths.add(SearchPath.initialPath(candidate, head, tail));
    }
    // search graph.
    List<SearchPath> resultPaths = search(paths, debugData);
    // generate results from successful paths.
    List<_SingleAnalysis> result = new ArrayList<>(resultPaths.size());
    for (SearchPath path : resultPaths) {
        _SingleAnalysis analysis = _SingleAnalysis.fromSearchPath(path);
        result.add(analysis);
        if (debugData != null) {
            debugData.results.add(analysis);
        }
    }
    return result;
}

Example 5 with zemberek.morphology._analyzer._SingleAnalysis

use of zemberek.morphology._analyzer._SingleAnalysis in project zemberek-nlp by ahmetaa.

the class _MorphologicalAmbiguityResolverExperiment method extracData.

public void extracData(Path p, Path outRoot, int maxAnalysisCount, int resultLimit) throws IOException {
    List<Path> files = Files.walk(p, 1).filter(s -> s.toFile().isFile() && s.toFile().getName().endsWith(".corpus")).collect(Collectors.toList());
    LinkedHashSet<SingleAnalysisSentence> result = new LinkedHashSet<>();
    int i = 0;
    for (Path file : files) {
        List<SingleAnalysisSentence> collect = collect(file, maxAnalysisCount);
        result.addAll(collect);
        i++;
        Log.info("%d of %d", i, files.size());
        if (resultLimit > 0 && result.size() > resultLimit) {
            break;
        }
    }
    String s = p.toFile().getName();
    Path out = outRoot.resolve(s + "-unambigious.txt");
    try (PrintWriter pw = new PrintWriter(out.toFile(), "utf-8")) {
        for (SingleAnalysisSentence sentence : result) {
            pw.println(sentence.sentence);
            for (Single single : sentence.tokens) {
                for (_SingleAnalysis r : single.res) {
                    pw.println(r.formatSurfaceSequence());
                }
            }
            pw.println();
        }
    }
    // saving failed words.
    failedWords.saveSortedByKeys(outRoot.resolve(s + "-failed.txt"), " ", Turkish.STRING_COMPARATOR_ASC);
    // saving failed words by frequency.
    failedWords.saveSortedByCounts(outRoot.resolve(s + "-failed.freq.txt"), " ");
}