zhipeng93 commented on code in PR #196: URL: https://github.com/apache/flink-ml/pull/196#discussion_r1090392883
########## flink-ml-lib/src/main/java/org/apache/flink/ml/feature/standardscaler/OnlineStandardScalerModel.java: ########## @@ -0,0 +1,298 @@ +/* + * 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.flink.ml.feature.standardscaler; + +import org.apache.flink.api.common.state.ListState; +import org.apache.flink.api.common.state.ListStateDescriptor; +import org.apache.flink.api.common.typeinfo.TypeInformation; +import org.apache.flink.api.common.typeinfo.Types; +import org.apache.flink.api.java.typeutils.RowTypeInfo; +import org.apache.flink.metrics.Gauge; +import org.apache.flink.metrics.MetricGroup; +import org.apache.flink.ml.api.Model; +import org.apache.flink.ml.common.datastream.TableUtils; +import org.apache.flink.ml.common.metrics.MLMetrics; +import org.apache.flink.ml.linalg.BLAS; +import org.apache.flink.ml.linalg.DenseVector; +import org.apache.flink.ml.linalg.Vector; +import org.apache.flink.ml.linalg.typeinfo.VectorTypeInfo; +import org.apache.flink.ml.param.Param; +import org.apache.flink.ml.util.ParamUtils; +import org.apache.flink.ml.util.ReadWriteUtils; +import org.apache.flink.runtime.state.StateInitializationContext; +import org.apache.flink.runtime.state.StateSnapshotContext; +import org.apache.flink.streaming.api.datastream.DataStream; +import org.apache.flink.streaming.api.operators.AbstractStreamOperator; +import org.apache.flink.streaming.api.operators.TwoInputStreamOperator; +import org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer; +import org.apache.flink.streaming.runtime.streamrecord.StreamRecord; +import org.apache.flink.table.api.Table; +import org.apache.flink.table.api.bridge.java.StreamTableEnvironment; +import org.apache.flink.table.api.internal.TableImpl; +import org.apache.flink.types.Row; +import org.apache.flink.util.Preconditions; + +import org.apache.commons.lang3.ArrayUtils; + +import java.io.IOException; +import java.util.ArrayList; +import java.util.HashMap; +import java.util.List; +import java.util.Map; +import java.util.Objects; + +/** A Model which transforms data using the model data computed by {@link OnlineStandardScaler}. */ +public class OnlineStandardScalerModel + implements Model<OnlineStandardScalerModel>, + OnlineStandardScalerModelParams<OnlineStandardScalerModel> { + private final Map<Param<?>, Object> paramMap = new HashMap<>(); + private Table modelDataTable; + + public OnlineStandardScalerModel() { + ParamUtils.initializeMapWithDefaultValues(paramMap, this); + } + + @Override + public Table[] transform(Table... inputs) { + Preconditions.checkArgument(inputs.length == 1); + StreamTableEnvironment tEnv = + (StreamTableEnvironment) ((TableImpl) inputs[0]).getTableEnvironment(); + + RowTypeInfo inputTypeInfo = TableUtils.getRowTypeInfo(inputs[0].getResolvedSchema()); + String modelVersionCol = getModelVersionCol(); + + TypeInformation<?>[] outputTypes; + String[] outputNames; + if (modelVersionCol == null) { + outputTypes = ArrayUtils.addAll(inputTypeInfo.getFieldTypes(), VectorTypeInfo.INSTANCE); + outputNames = ArrayUtils.addAll(inputTypeInfo.getFieldNames(), getOutputCol()); + } else { + outputTypes = + ArrayUtils.addAll( + inputTypeInfo.getFieldTypes(), VectorTypeInfo.INSTANCE, Types.LONG); + outputNames = + ArrayUtils.addAll( + inputTypeInfo.getFieldNames(), getOutputCol(), modelVersionCol); + } + RowTypeInfo outputTypeInfo = new RowTypeInfo(outputTypes, outputNames); + + DataStream<Row> predictionResult = + tEnv.toDataStream(inputs[0]) + .connect( + StandardScalerModelData.getModelDataStream(modelDataTable) + .broadcast()) + .transform( + "PredictionOperator", + outputTypeInfo, + new PredictionOperator( + inputTypeInfo, + getInputCol(), + getWithMean(), + getWithStd(), + getMaxAllowedModelDelayMs(), + getModelVersionCol())); + + return new Table[] {tEnv.fromDataStream(predictionResult)}; + } + + /** A utility operator used for prediction. */ + @SuppressWarnings({"unchecked", "rawtypes"}) + private static class PredictionOperator extends AbstractStreamOperator<Row> + implements TwoInputStreamOperator<Row, StandardScalerModelData, Row> { + private final RowTypeInfo inputTypeInfo; + + private final String inputCol; + + private final boolean withMean; + + private final boolean withStd; + + private final long maxAllowedModelDelayMs; + + private final String modelVersionCol; + + private ListState<StreamRecord> bufferedPointsState; + + private ListState<StandardScalerModelData> modelDataState; + + /** Model data for inference. */ + private StandardScalerModelData modelData; + + private DenseVector mean; + + /** Inverse of standard deviation. */ + private DenseVector scale; + + private long modelVersion; + + private long modelTimeStamp; + + public PredictionOperator( + RowTypeInfo inputTypeInfo, + String inputCol, + boolean withMean, + boolean withStd, + long maxAllowedModelDelayMs, + String modelVersionCol) { + this.inputTypeInfo = inputTypeInfo; + this.inputCol = inputCol; + this.withMean = withMean; + this.withStd = withStd; + this.maxAllowedModelDelayMs = maxAllowedModelDelayMs; + this.modelVersionCol = modelVersionCol; + } + + @Override + public void initializeState(StateInitializationContext context) throws Exception { + super.initializeState(context); + bufferedPointsState = + context.getOperatorStateStore() + .getListState( + new ListStateDescriptor<StreamRecord>( + "bufferedPoints", + new StreamElementSerializer( + inputTypeInfo.createSerializer( + getExecutionConfig())))); + + modelDataState = + context.getOperatorStateStore() + .getListState( + new ListStateDescriptor<>( + "modelData", + TypeInformation.of(StandardScalerModelData.class))); + } + + @Override + public void snapshotState(StateSnapshotContext context) throws Exception { + super.snapshotState(context); + if (modelData != null) { + modelDataState.clear(); + modelDataState.add(modelData); + } + } + + @Override + public void open() throws Exception { + super.open(); + MetricGroup mlModelMetricGroup = + getRuntimeContext() + .getMetricGroup() + .addGroup(MLMetrics.ML_GROUP) + .addGroup(MLMetrics.ML_MODEL_GROUP); + mlModelMetricGroup.gauge(MLMetrics.TIMESTAMP, (Gauge<Long>) () -> modelTimeStamp); + mlModelMetricGroup.gauge(MLMetrics.VERSION, (Gauge<Long>) () -> modelVersion); + } + + @Override + public void processElement1(StreamRecord<Row> dataPoint) throws Exception { + if (dataPoint.getTimestamp() <= modelTimeStamp + maxAllowedModelDelayMs Review Comment: Good catch, it is indeed a bug. I have fixed the training logic as follows: - When the window strategy is based on event-time, the timestamp of model data is the maximum timestamp of the window that generates this model. - Otherwise, the timestamp of the model data is `LONG.MAX_VALUE`. Moreover, to avoid the possible overflow, the `if condition` here is updated as `if (dataPoint.getTimestamp() - maxAllowedModelDelayMs <= modelTimeStamp`. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
