Examples with Feature org.jpmml.converter.Feature used on opensource projects

Example 46 with Feature

use of org.jpmml.converter.Feature in project jpmml-sparkml by jpmml.

the class CountVectorizerModelConverter method encodeFeatures.

@Override
public List<Feature> encodeFeatures(SparkMLEncoder encoder) {
    CountVectorizerModel transformer = getTransformer();
    DocumentFeature documentFeature = (DocumentFeature) encoder.getOnlyFeature(transformer.getInputCol());
    ParameterField documentField = new ParameterField(FieldName.create("document"));
    ParameterField termField = new ParameterField(FieldName.create("term"));
    TextIndex textIndex = new TextIndex(documentField.getName()).setTokenize(Boolean.TRUE).setWordSeparatorCharacterRE(documentFeature.getWordSeparatorRE()).setLocalTermWeights(transformer.getBinary() ? TextIndex.LocalTermWeights.BINARY : null).setExpression(new FieldRef(termField.getName()));
    Set<DocumentFeature.StopWordSet> stopWordSets = documentFeature.getStopWordSets();
    for (DocumentFeature.StopWordSet stopWordSet : stopWordSets) {
        if (stopWordSet.isEmpty()) {
            continue;
        }
        DocumentBuilder documentBuilder = DOMUtil.createDocumentBuilder();
        String tokenRE;
        String wordSeparatorRE = documentFeature.getWordSeparatorRE();
        switch(wordSeparatorRE) {
            case "\\s+":
                tokenRE = "(^|\\s+)\\p{Punct}*(" + JOINER.join(stopWordSet) + ")\\p{Punct}*(\\s+|$)";
                break;
            case "\\W+":
                tokenRE = "(\\W+)(" + JOINER.join(stopWordSet) + ")(\\W+)";
                break;
            default:
                throw new IllegalArgumentException("Expected \"\\s+\" or \"\\W+\" as splitter regex pattern, got \"" + wordSeparatorRE + "\"");
        }
        InlineTable inlineTable = new InlineTable().addRows(DOMUtil.createRow(documentBuilder, Arrays.asList("string", "stem", "regex"), Arrays.asList(tokenRE, " ", "true")));
        TextIndexNormalization textIndexNormalization = new TextIndexNormalization().setCaseSensitive(stopWordSet.isCaseSensitive()).setRecursive(// Handles consecutive matches. See http://stackoverflow.com/a/25085385
        Boolean.TRUE).setInlineTable(inlineTable);
        textIndex.addTextIndexNormalizations(textIndexNormalization);
    }
    DefineFunction defineFunction = new DefineFunction("tf" + "@" + String.valueOf(CountVectorizerModelConverter.SEQUENCE.getAndIncrement()), OpType.CONTINUOUS, null).setDataType(DataType.INTEGER).addParameterFields(documentField, termField).setExpression(textIndex);
    encoder.addDefineFunction(defineFunction);
    List<Feature> result = new ArrayList<>();
    String[] vocabulary = transformer.vocabulary();
    for (int i = 0; i < vocabulary.length; i++) {
        String term = vocabulary[i];
        if (TermUtil.hasPunctuation(term)) {
            throw new IllegalArgumentException(term);
        }
        result.add(new TermFeature(encoder, defineFunction, documentFeature, term));
    }
    return result;
}

Example 47 with Feature

use of org.jpmml.converter.Feature in project jpmml-sparkml by jpmml.

the class IDFModelConverter method encodeFeatures.

@Override
public List<Feature> encodeFeatures(SparkMLEncoder encoder) {
    IDFModel transformer = getTransformer();
    List<Feature> features = encoder.getFeatures(transformer.getInputCol());
    Vector idf = transformer.idf();
    if (idf.size() != features.size()) {
        throw new IllegalArgumentException();
    }
    List<Feature> result = new ArrayList<>();
    for (int i = 0; i < features.size(); i++) {
        Feature feature = features.get(i);
        TermFeature termFeature = (TermFeature) feature;
        result.add(termFeature.toWeightedTermFeature(idf.apply(i)));
    }
    return result;
}

Example 48 with Feature

use of org.jpmml.converter.Feature in project jpmml-sparkml by jpmml.

the class ClassificationModelConverter method registerOutputFields.

@Override
public List<OutputField> registerOutputFields(Label label, SparkMLEncoder encoder) {
    T model = getTransformer();
    CategoricalLabel categoricalLabel = (CategoricalLabel) label;
    List<OutputField> result = new ArrayList<>();
    String predictionCol = model.getPredictionCol();
    OutputField pmmlPredictedField = ModelUtil.createPredictedField(FieldName.create("pmml(" + predictionCol + ")"), categoricalLabel.getDataType(), OpType.CATEGORICAL);
    result.add(pmmlPredictedField);
    List<String> categories = new ArrayList<>();
    DocumentBuilder documentBuilder = DOMUtil.createDocumentBuilder();
    InlineTable inlineTable = new InlineTable();
    List<String> columns = Arrays.asList("input", "output");
    for (int i = 0; i < categoricalLabel.size(); i++) {
        String value = categoricalLabel.getValue(i);
        String category = String.valueOf(i);
        categories.add(category);
        Row row = DOMUtil.createRow(documentBuilder, columns, Arrays.asList(value, category));
        inlineTable.addRows(row);
    }
    MapValues mapValues = new MapValues().addFieldColumnPairs(new FieldColumnPair(pmmlPredictedField.getName(), columns.get(0))).setOutputColumn(columns.get(1)).setInlineTable(inlineTable);
    final OutputField predictedField = new OutputField(FieldName.create(predictionCol), DataType.DOUBLE).setOpType(OpType.CATEGORICAL).setResultFeature(ResultFeature.TRANSFORMED_VALUE).setExpression(mapValues);
    result.add(predictedField);
    Feature feature = new CategoricalFeature(encoder, predictedField.getName(), predictedField.getDataType(), categories) {

        @Override
        public ContinuousFeature toContinuousFeature() {
            PMMLEncoder encoder = ensureEncoder();
            return new ContinuousFeature(encoder, getName(), getDataType());
        }
    };
    encoder.putOnlyFeature(predictionCol, feature);
    if (model instanceof HasProbabilityCol) {
        HasProbabilityCol hasProbabilityCol = (HasProbabilityCol) model;
        String probabilityCol = hasProbabilityCol.getProbabilityCol();
        List<Feature> features = new ArrayList<>();
        for (int i = 0; i < categoricalLabel.size(); i++) {
            String value = categoricalLabel.getValue(i);
            OutputField probabilityField = ModelUtil.createProbabilityField(FieldName.create(probabilityCol + "(" + value + ")"), DataType.DOUBLE, value);
            result.add(probabilityField);
            features.add(new ContinuousFeature(encoder, probabilityField.getName(), probabilityField.getDataType()));
        }
        encoder.putFeatures(probabilityCol, features);
    }
    return result;
}

Example 49 with Feature

use of org.jpmml.converter.Feature in project jpmml-sparkml by jpmml.

the class ConverterUtil method toPMML.

public static PMML toPMML(StructType schema, PipelineModel pipelineModel) {
    checkVersion();
    SparkMLEncoder encoder = new SparkMLEncoder(schema);
    List<org.dmg.pmml.Model> models = new ArrayList<>();
    Iterable<Transformer> transformers = getTransformers(pipelineModel);
    for (Transformer transformer : transformers) {
        TransformerConverter<?> converter = ConverterUtil.createConverter(transformer);
        if (converter instanceof FeatureConverter) {
            FeatureConverter<?> featureConverter = (FeatureConverter<?>) converter;
            featureConverter.registerFeatures(encoder);
        } else if (converter instanceof ModelConverter) {
            ModelConverter<?> modelConverter = (ModelConverter<?>) converter;
            org.dmg.pmml.Model model = modelConverter.registerModel(encoder);
            models.add(model);
        } else {
            throw new IllegalArgumentException("Expected a " + FeatureConverter.class.getName() + " or " + ModelConverter.class.getName() + " instance, got " + converter);
        }
    }
    org.dmg.pmml.Model rootModel;
    if (models.size() == 1) {
        rootModel = Iterables.getOnlyElement(models);
    } else if (models.size() > 1) {
        List<MiningField> targetMiningFields = new ArrayList<>();
        for (org.dmg.pmml.Model model : models) {
            MiningSchema miningSchema = model.getMiningSchema();
            List<MiningField> miningFields = miningSchema.getMiningFields();
            for (MiningField miningField : miningFields) {
                MiningField.UsageType usageType = miningField.getUsageType();
                switch(usageType) {
                    case PREDICTED:
                    case TARGET:
                        targetMiningFields.add(miningField);
                        break;
                    default:
                        break;
                }
            }
        }
        MiningSchema miningSchema = new MiningSchema(targetMiningFields);
        MiningModel miningModel = MiningModelUtil.createModelChain(models, new Schema(null, Collections.<Feature>emptyList())).setMiningSchema(miningSchema);
        rootModel = miningModel;
    } else {
        throw new IllegalArgumentException("Expected a pipeline with one or more models, got a pipeline with zero models");
    }
    PMML pmml = encoder.encodePMML(rootModel);
    return pmml;
}

Example 50 with Feature

use of org.jpmml.converter.Feature in project jpmml-sparkml by jpmml.

the class FeatureConverter method registerFeatures.

public void registerFeatures(SparkMLEncoder encoder) {
    Transformer transformer = getTransformer();
    if (transformer instanceof HasOutputCol) {
        HasOutputCol hasOutputCol = (HasOutputCol) transformer;
        String outputCol = hasOutputCol.getOutputCol();
        List<Feature> features = encodeFeatures(encoder);
        encoder.putFeatures(outputCol, features);
    } else if (transformer instanceof HasOutputCols) {
        HasOutputCols hasOutputCols = (HasOutputCols) transformer;
        String[] outputCols = hasOutputCols.getOutputCols();
        List<Feature> features = encodeFeatures(encoder);
        if (outputCols.length != features.size()) {
            throw new IllegalArgumentException("Expected " + outputCols.length + " features, got " + features.size() + " features");
        }
        for (int i = 0; i < outputCols.length; i++) {
            String outputCol = outputCols[i];
            Feature feature = features.get(i);
            if (feature instanceof BinarizedCategoricalFeature) {
                BinarizedCategoricalFeature binarizedCategoricalFeature = (BinarizedCategoricalFeature) feature;
                encoder.putFeatures(outputCol, (List) binarizedCategoricalFeature.getBinaryFeatures());
            } else {
                encoder.putOnlyFeature(outputCol, feature);
            }
        }
    }
}