yunfengzhou-hub commented on a change in pull request #24:
URL: https://github.com/apache/flink-ml/pull/24#discussion_r757264541
##########
File path:
flink-ml-lib/src/main/java/org/apache/flink/ml/classification/knn/KnnParams.java
##########
@@ -0,0 +1,18 @@
+package org.apache.flink.ml.classification.knn;
+
+import org.apache.flink.ml.common.param.HasFeatureColsDefaultAsNull;
+import org.apache.flink.ml.common.param.HasK;
+import org.apache.flink.ml.common.param.HasLabelCol;
+import org.apache.flink.ml.common.param.HasPredictionCol;
+import org.apache.flink.ml.common.param.HasVectorColDefaultAsNull;
+import org.apache.flink.ml.param.WithParams;
+
+/** knn parameters. */
+public interface KnnParams<T>
+ extends WithParams<T>,
Review comment:
It seems that `KnnParams` does not need to directly inherit from
`WithParams`.
##########
File path:
flink-ml-lib/src/main/java/org/apache/flink/ml/classification/knn/KnnModel.java
##########
@@ -0,0 +1,493 @@
+package org.apache.flink.ml.classification.knn;
+
+import org.apache.flink.api.common.eventtime.WatermarkStrategy;
+import org.apache.flink.api.common.functions.RichMapFunction;
+import org.apache.flink.api.common.typeinfo.TypeInformation;
+import org.apache.flink.api.connector.source.Source;
+import org.apache.flink.api.java.tuple.Tuple2;
+import org.apache.flink.api.java.typeutils.RowTypeInfo;
+import org.apache.flink.connector.file.sink.FileSink;
+import org.apache.flink.connector.file.src.FileSource;
+import org.apache.flink.core.fs.Path;
+import org.apache.flink.ml.api.core.Model;
+import org.apache.flink.ml.common.broadcast.BroadcastUtils;
+import org.apache.flink.ml.linalg.DenseMatrix;
+import org.apache.flink.ml.linalg.DenseVector;
+import org.apache.flink.ml.linalg.VectorUtils;
+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.shaded.curator4.com.google.common.base.Preconditions;
+import org.apache.flink.shaded.curator4.com.google.common.collect.ImmutableMap;
+import
org.apache.flink.shaded.jackson2.com.fasterxml.jackson.core.JsonProcessingException;
+import org.apache.flink.streaming.api.datastream.DataStream;
+import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
+import
org.apache.flink.streaming.api.functions.sink.filesystem.bucketassigners.BasePathBucketAssigner;
+import
org.apache.flink.streaming.api.functions.sink.filesystem.rollingpolicies.OnCheckpointRollingPolicy;
+import org.apache.flink.streaming.api.operators.AbstractUdfStreamOperator;
+import org.apache.flink.streaming.api.operators.OneInputStreamOperator;
+import org.apache.flink.streaming.runtime.streamrecord.StreamRecord;
+import org.apache.flink.table.api.DataTypes;
+import org.apache.flink.table.api.Schema;
+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.table.catalog.ResolvedSchema;
+import org.apache.flink.table.types.DataType;
+import org.apache.flink.table.types.logical.utils.LogicalTypeParser;
+import org.apache.flink.table.types.utils.LogicalTypeDataTypeConverter;
+import org.apache.flink.types.Row;
+
+import org.apache.commons.lang3.ArrayUtils;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.PriorityQueue;
+import java.util.TreeMap;
+import java.util.function.Function;
+
+/** Knn classification model fitted by estimator. */
+public class KnnModel implements Model<KnnModel>, KnnParams<KnnModel> {
+ protected Map<Param<?>, Object> params = new HashMap<>();
+ private Table[] modelData;
+
+ /** constructor. */
+ public KnnModel() {
+ ParamUtils.initializeMapWithDefaultValues(params, this);
+ }
+
+ /**
+ * constructor.
+ *
+ * @param params parameters for algorithm.
+ */
+ public KnnModel(Map<Param<?>, Object> params) {
+ this.params = params;
+ }
+
+ /**
+ * Set model data for knn prediction.
+ *
+ * @param modelData knn model.
+ * @return knn classification model.
+ */
+ @Override
+ public KnnModel setModelData(Table... modelData) {
+ this.modelData = modelData;
+ return this;
+ }
+
+ /**
+ * get model data.
+ *
+ * @return list of tables.
+ */
+ @Override
+ public Table[] getModelData() {
+ return modelData;
+ }
+
+ /**
+ * @param inputs a list of tables.
+ * @return result.
+ */
+ @Override
+ public Table[] transform(Table... inputs) {
+ StreamTableEnvironment tEnv =
+ (StreamTableEnvironment) ((TableImpl)
inputs[0]).getTableEnvironment();
+ DataStream<Row> input = tEnv.toDataStream(inputs[0]);
+ DataStream<Row> model = tEnv.toDataStream(modelData[0]);
+ final String BROADCAST_STR = "broadcastModelKey";
+ Map<String, DataStream<?>> broadcastMap = new HashMap<>(1);
+ broadcastMap.put(BROADCAST_STR, model);
+ ResolvedSchema modelSchema = modelData[0].getResolvedSchema();
+ DataType idType =
+
modelSchema.getColumnDataTypes().get(modelSchema.getColumnNames().size() - 1);
+ String[] reservedCols =
inputs[0].getResolvedSchema().getColumnNames().toArray(new String[0]);
+ DataType[] reservedTypes =
inputs[0].getResolvedSchema().getColumnDataTypes().toArray(new DataType[0]);
Review comment:
Maybe we can reuse methods in `TableUtils` here.
##########
File path: flink-ml-lib/src/main/java/org/apache/flink/ml/common/param/HasK.java
##########
@@ -0,0 +1,41 @@
+/*
+ * 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.common.param;
+
+import org.apache.flink.ml.param.IntParam;
+import org.apache.flink.ml.param.Param;
+import org.apache.flink.ml.param.ParamValidators;
+import org.apache.flink.ml.param.WithParams;
+
+/** Interface for the shared K param. */
+public interface HasK<T> extends WithParams<T> {
+
+ /**
+ * topK
+ */
+ Param<Integer> K = new IntParam("k", "k", 10, ParamValidators.gt(0));
Review comment:
If an agreement has been reached that we should extract `K` as a params,
I suggest that we can directly apply this to all algorithms that have
implemented their own `getK` `setK` methods.
Besides, it might be better to set the description, default value and
validators in reference to existing practices.
##########
File path:
flink-ml-lib/src/main/java/org/apache/flink/ml/classification/knn/KnnModelData.java
##########
@@ -0,0 +1,134 @@
+package org.apache.flink.ml.classification.knn;
+
+import org.apache.flink.api.common.serialization.Encoder;
+import org.apache.flink.api.common.typeinfo.TypeInformation;
+import org.apache.flink.api.common.typeinfo.Types;
+import org.apache.flink.api.java.tuple.Tuple2;
+import org.apache.flink.api.java.typeutils.RowTypeInfo;
+import org.apache.flink.configuration.Configuration;
+import org.apache.flink.connector.file.src.reader.SimpleStreamFormat;
+import org.apache.flink.core.fs.FSDataInputStream;
+import org.apache.flink.table.types.DataType;
+import org.apache.flink.types.Row;
+
+import com.esotericsoftware.kryo.Kryo;
+import com.esotericsoftware.kryo.io.Input;
+import com.esotericsoftware.kryo.io.Output;
+
+import java.io.IOException;
+import java.io.OutputStream;
+import java.io.Serializable;
+import java.util.ArrayList;
+import java.util.Comparator;
+import java.util.List;
+
+/** knn model data, which will be used to calculate the distances between
nodes. */
+public class KnnModelData implements Serializable, Cloneable {
+ private static final long serialVersionUID = -2940551481683238630L;
Review comment:
Like the comments said in NaiveBayes PR, maybe we don't need to set this
`serialVersionUID`. And that it might be better to serialize model data into
primitives, instead of java-specific objects.
--
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
