Github user arunmahadevan commented on a diff in the pull request:
https://github.com/apache/storm/pull/1816#discussion_r92331432
--- Diff:
external/storm-pmml/src/main/java/org/apache/storm/pmml/runner/jpmml/JPmmlModelRunner.java
---
@@ -0,0 +1,134 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.storm.pmml.runner.jpmml;
+
+import org.apache.storm.pmml.model.ModelOutputFields;
+import org.apache.storm.pmml.model.Stream;
+import org.apache.storm.pmml.runner.PmmlModelRunner;
+import org.apache.storm.tuple.Fields;
+import org.apache.storm.tuple.Tuple;
+import org.dmg.pmml.FieldName;
+import org.jpmml.evaluator.Evaluator;
+import org.jpmml.evaluator.EvaluatorUtil;
+import org.jpmml.evaluator.FieldValue;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import java.util.ArrayList;
+import java.util.HashMap;
+import java.util.LinkedHashMap;
+import java.util.List;
+import java.util.Map;
+
+/**
+ * JPMML implementation of {@link PmmlModelRunner}. It extracts the raw
inputs from the tuple for all
+ * 'active fields', and builds a tuple with the predicted scores for the
'predicted fields' and 'output fields'.
+ * In this implementation all the declared streams will have the same
scored tuple.
+ *
+ * The 'predicted', 'active', and 'output' fields are extracted from the
PMML model.
+ */
+public class JPmmlModelRunner implements PmmlModelRunner<Map<FieldName,
Object>,
+ Map<FieldName, FieldValue>,
+ Map<FieldName, ?>> {
+
+ private static final Logger LOG =
LoggerFactory.getLogger(JPmmlModelRunner.class);
+
+ private final Evaluator eval; // Jpmml evaluator
+ private final List<FieldName> activeFields;
+ private final List<FieldName> predictedFields;
+ private final List<FieldName> outputFields;
+ private ModelOutputFields modelOutputFields;
+
+ public JPmmlModelRunner(Evaluator evaluator, ModelOutputFields
modelOutputFields) {
+ this.eval = evaluator;
+ activeFields = evaluator.getActiveFields();
+ this.modelOutputFields = modelOutputFields;
+ predictedFields = eval.getPredictedFields();
+ outputFields = eval.getOutputFields();
+ }
+
+ /**
+ * @return The raw inputs extracted from the tuple for all 'active
fields'
+ */
+ @Override
+ public Map<FieldName, Object> extractRawInputs(Tuple tuple) {
+ LOG.debug("Extracting raw inputs from tuple: = [{}]", tuple);
+ final Map<FieldName, Object> rawInputs = new LinkedHashMap<>();
+ for (FieldName activeField : activeFields) {
+ rawInputs.put(activeField,
tuple.getValueByField(activeField.getValue()));
+ }
+ LOG.debug("Raw inputs = [{}]", rawInputs);
+ return rawInputs;
+ }
+
+ @Override
+ public Map<FieldName, FieldValue> preProcessInputs(Map<FieldName,
Object> rawInputs) {
+ LOG.debug("Pre processing raw inputs: = [{}]", rawInputs);
+ final Map<FieldName, FieldValue> preProcInputs = new
LinkedHashMap<>();
+ for (Map.Entry<FieldName, Object> rawEntry : rawInputs.entrySet())
{
+ preProcInputs.putIfAbsent(rawEntry.getKey(),
EvaluatorUtil.prepare(eval, rawEntry.getKey(), rawEntry.getValue()));
+ }
+ LOG.debug("Pre processed inputs = [{}]", preProcInputs);
+ return preProcInputs;
+ }
+
+ @Override
+ public Map<FieldName, ?> predictScores(Map<FieldName, FieldValue>
preProcInputs) {
+ LOG.debug("Predicting scores for pre processed inputs: = [{}]",
preProcInputs);
+ Map<FieldName, ?> predictedScores = eval.evaluate(preProcInputs);
+ LOG.debug("Predicted scores = [{}]", predictedScores);
+ return predictedScores;
+ }
+
+ /**
+ * @return the predicted scores for the 'predicted fields' and 'output
fields'.
+ * All the declared streams will have the same scored tuple.
+ */
+ @Override
+ public Map<Stream, List<Object>> scoredTuplePerStream(Tuple input) {
+ final Map<FieldName, Object> rawInputs = extractRawInputs(input);
+ final Map<FieldName, FieldValue> preProcInputs =
preProcessInputs(rawInputs);
+ final Map<FieldName, ?> predScores = predictScores(preProcInputs);
+
+ return toValuesMap(predScores);
+ }
+
+ // Sends the same tuple (list of scored/predicted values) to all the
declared streams
+ private Map<Stream, List<Object>> toValuesMap(Map<FieldName, ?>
predScores) {
+ final List<Object> scoredVals = new ArrayList<>();
+
+ for (FieldName predictedField : predictedFields) {
+ Object targetValue = predScores.get(predictedField);
+ scoredVals.add(EvaluatorUtil.decode(targetValue));
+ }
+
+ for (FieldName outputField : outputFields) {
+ Object targetValue = predScores.get(outputField);
+ scoredVals.add(EvaluatorUtil.decode(targetValue));
+ }
+
+ final Map<Stream, List<Object>> valuesMap = new HashMap<>();
+
+ for (Map.Entry<? extends Stream, ? extends Fields> entry :
modelOutputFields.toMap().entrySet()) {
--- End diff --
use keySet. It may be better if ModelOuputFields itself has a method to
returns the keys, right now its more of map wrapper.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
