yunfengzhou-hub commented on code in PR #150: URL: https://github.com/apache/flink-ml/pull/150#discussion_r961342366
########## flink-ml-lib/src/main/java/org/apache/flink/ml/feature/idf/IDFModelData.java: ########## @@ -0,0 +1,124 @@ +/* + * 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.idf; + +import org.apache.flink.api.common.serialization.Encoder; +import org.apache.flink.api.common.typeinfo.TypeInformation; +import org.apache.flink.api.common.typeutils.base.LongSerializer; +import org.apache.flink.api.common.typeutils.base.array.LongPrimitiveArraySerializer; +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.core.memory.DataInputView; +import org.apache.flink.core.memory.DataInputViewStreamWrapper; +import org.apache.flink.core.memory.DataOutputView; +import org.apache.flink.core.memory.DataOutputViewStreamWrapper; +import org.apache.flink.ml.linalg.DenseVector; +import org.apache.flink.ml.linalg.typeinfo.DenseVectorSerializer; +import org.apache.flink.streaming.api.datastream.DataStream; +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 java.io.EOFException; +import java.io.IOException; +import java.io.OutputStream; + +/** + * Model data of {@link IDFModel}. + * + * <p>This class also provides methods to convert model data from Table to a data stream, and + * classes to save/load model data. + */ +public class IDFModelData { + /** Inverse document frequency for all terms. */ + public DenseVector idf; + /** Document frequency for all terms after filtering out infrequent terms. */ + public long[] docFreq; + /** Number of docs in the training set. */ + public long numDocs; Review Comment: Spark's `IDFModel` provides `docFreq()` and `numDocs()` methods, so that users can access these two values. Since we would not add such methods to the `IDFModel` in this PR, do we expect users to access these two values by directly reading from the model data? ########## flink-ml-lib/src/main/java/org/apache/flink/ml/feature/idf/IDF.java: ########## @@ -0,0 +1,167 @@ +/* + * 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.idf; + +import org.apache.flink.api.common.functions.AggregateFunction; +import org.apache.flink.api.common.functions.MapFunction; +import org.apache.flink.api.java.tuple.Tuple2; +import org.apache.flink.ml.api.Estimator; +import org.apache.flink.ml.common.datastream.DataStreamUtils; +import org.apache.flink.ml.linalg.BLAS; +import org.apache.flink.ml.linalg.DenseVector; +import org.apache.flink.ml.linalg.SparseVector; +import org.apache.flink.ml.linalg.Vector; +import org.apache.flink.ml.linalg.Vectors; +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.streaming.api.datastream.DataStream; +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 java.io.IOException; +import java.util.HashMap; +import java.util.Map; + +/** + * An Estimator that computes the inverse document frequency (IDF) for the input documents. IDF is + * computed following `idf = log((m + 1) / (d(t) + 1))`, where `m` is the total number of documents + * and `d(t)` is the number of documents that contains `t`. + * + * <p>Users could filter out terms that appeared in little documents by setting {@link + * IDFParams#getMinDocFreq()}. Review Comment: nit: it might be better to add a wiki link here, being consistent with other JavaDocs. ########## flink-ml-python/pyflink/examples/ml/feature/idf_example.py: ########## @@ -0,0 +1,65 @@ +################################################################################ +# 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. +################################################################################ + +# Simple program that trains a StringIndexer model and uses it for feature +# engineering. +# +# Before executing this program, please make sure you have followed Flink ML's +# quick start guideline to set up Flink ML and Flink environment. The guideline +# can be found at +# +# https://nightlies.apache.org/flink/flink-ml-docs-master/docs/try-flink-ml/quick-start/ + +from pyflink.common import Types +from pyflink.ml.core.linalg import Vectors, DenseVectorTypeInfo +from pyflink.datastream import StreamExecutionEnvironment +from pyflink.ml.lib.feature.idf import IDF +from pyflink.table import StreamTableEnvironment + +# Creates a new StreamExecutionEnvironment. +env = StreamExecutionEnvironment.get_execution_environment() + +# Creates a StreamTableEnvironment. +t_env = StreamTableEnvironment.create(env) + +# Generates input for training and prediction. +input_table = t_env.from_data_stream( + env.from_collection([ + (Vectors.dense(0, 1, 0, 2),), + (Vectors.dense(0, 1, 2, 3),), + (Vectors.dense(0, 1, 0, 0),), + ], + type_info=Types.ROW_NAMED( + ['input', ], + [DenseVectorTypeInfo(), ]))) + +# Creates an IDF object and initializes its parameters. +idf = IDF().set_min_doc_freq(2) + +# Trains the IDF Model. +model = idf.fit(input_table) + +# Uses the IDF Model for predictions. +output = model.transform(input_table)[0] + +# Extracts and displays the results. +field_names = output.get_schema().get_field_names() +for result in t_env.to_data_stream(output).execute_and_collect(): + print('Input Value: ' + str(result[field_names.index(idf.get_input_col())]) + + '\tOutput Value: ' + + str(result[field_names.index(idf.get_output_col())])) Review Comment: It might be better to refractor this part of code to improve its readability. ########## flink-ml-lib/src/main/java/org/apache/flink/ml/feature/idf/IDFModel.java: ########## @@ -0,0 +1,156 @@ +/* + * 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.idf; + +import org.apache.flink.api.common.functions.RichMapFunction; +import org.apache.flink.api.java.typeutils.RowTypeInfo; +import org.apache.flink.ml.api.Model; +import org.apache.flink.ml.common.broadcast.BroadcastUtils; +import org.apache.flink.ml.common.datastream.TableUtils; +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.param.Param; +import org.apache.flink.ml.util.ParamUtils; +import org.apache.flink.ml.util.ReadWriteUtils; +import org.apache.flink.streaming.api.datastream.DataStream; +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.types.Row; +import org.apache.flink.util.Preconditions; + +import org.apache.commons.lang3.ArrayUtils; + +import java.io.IOException; +import java.util.Collections; +import java.util.HashMap; +import java.util.Map; +import java.util.Objects; + +/** A Model which transforms data using the model data computed by {@link IDF}. */ +public class IDFModel implements Model<IDFModel>, IDFModelParams<IDFModel> { + private final Map<Param<?>, Object> paramMap = new HashMap<>(); + private Table modelDataTable; + + public IDFModel() { + ParamUtils.initializeMapWithDefaultValues(paramMap, this); + } + + @Override + @SuppressWarnings("unchecked") + public Table[] transform(Table... inputs) { + Preconditions.checkArgument(inputs.length == 1); + StreamTableEnvironment tEnv = + (StreamTableEnvironment) ((TableImpl) inputs[0]).getTableEnvironment(); + + DataStream<Row> data = tEnv.toDataStream(inputs[0]); + DataStream<IDFModelData> idfModelData = IDFModelData.getModelDataStream(modelDataTable); + + final String broadcastModelKey = "broadcastModelKey"; + ResolvedSchema schema = inputs[0].getResolvedSchema(); + RowTypeInfo inputTypeInfo = TableUtils.getRowTypeInfo(schema); + RowTypeInfo outputTypeInfo = + new RowTypeInfo( + ArrayUtils.addAll( + inputTypeInfo.getFieldTypes(), + TableUtils.getTypeInfoByName(schema, getInputCol())), + ArrayUtils.addAll(inputTypeInfo.getFieldNames(), getOutputCol())); + + DataStream<Row> output = + BroadcastUtils.withBroadcastStream( + Collections.singletonList(data), + Collections.singletonMap(broadcastModelKey, idfModelData), + inputList -> { + DataStream input = inputList.get(0); + return input.map( + new ComputeTfIdfFunction(broadcastModelKey, getInputCol()), + outputTypeInfo); + }); + + return new Table[] {tEnv.fromDataStream(output)}; + } + + @Override + public IDFModel setModelData(Table... inputs) { + modelDataTable = inputs[0]; + return this; + } + + @Override + public Table[] getModelData() { + return new Table[] {modelDataTable}; + } + + @Override + public Map<Param<?>, Object> getParamMap() { + return paramMap; + } + + @Override + public void save(String path) throws IOException { + ReadWriteUtils.saveMetadata(this, path); + ReadWriteUtils.saveModelData( + IDFModelData.getModelDataStream(modelDataTable), + path, + new IDFModelData.ModelDataEncoder()); + } + + /** + * Loads model data from path. Review Comment: nit: It might be better to keep consistent with other `Stage` subclasses, and do not add a JavaDoc for `load` method for now. ########## flink-ml-lib/src/main/java/org/apache/flink/ml/feature/idf/IDFModel.java: ########## @@ -0,0 +1,156 @@ +/* + * 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.idf; + +import org.apache.flink.api.common.functions.RichMapFunction; +import org.apache.flink.api.java.typeutils.RowTypeInfo; +import org.apache.flink.ml.api.Model; +import org.apache.flink.ml.common.broadcast.BroadcastUtils; +import org.apache.flink.ml.common.datastream.TableUtils; +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.param.Param; +import org.apache.flink.ml.util.ParamUtils; +import org.apache.flink.ml.util.ReadWriteUtils; +import org.apache.flink.streaming.api.datastream.DataStream; +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.types.Row; +import org.apache.flink.util.Preconditions; + +import org.apache.commons.lang3.ArrayUtils; + +import java.io.IOException; +import java.util.Collections; +import java.util.HashMap; +import java.util.Map; +import java.util.Objects; + +/** A Model which transforms data using the model data computed by {@link IDF}. */ +public class IDFModel implements Model<IDFModel>, IDFModelParams<IDFModel> { + private final Map<Param<?>, Object> paramMap = new HashMap<>(); + private Table modelDataTable; + + public IDFModel() { + ParamUtils.initializeMapWithDefaultValues(paramMap, this); + } + + @Override + @SuppressWarnings("unchecked") + public Table[] transform(Table... inputs) { + Preconditions.checkArgument(inputs.length == 1); + StreamTableEnvironment tEnv = + (StreamTableEnvironment) ((TableImpl) inputs[0]).getTableEnvironment(); + + DataStream<Row> data = tEnv.toDataStream(inputs[0]); + DataStream<IDFModelData> idfModelData = IDFModelData.getModelDataStream(modelDataTable); + + final String broadcastModelKey = "broadcastModelKey"; + ResolvedSchema schema = inputs[0].getResolvedSchema(); + RowTypeInfo inputTypeInfo = TableUtils.getRowTypeInfo(schema); + RowTypeInfo outputTypeInfo = + new RowTypeInfo( + ArrayUtils.addAll( + inputTypeInfo.getFieldTypes(), + TableUtils.getTypeInfoByName(schema, getInputCol())), + ArrayUtils.addAll(inputTypeInfo.getFieldNames(), getOutputCol())); + + DataStream<Row> output = + BroadcastUtils.withBroadcastStream( + Collections.singletonList(data), + Collections.singletonMap(broadcastModelKey, idfModelData), + inputList -> { + DataStream input = inputList.get(0); + return input.map( + new ComputeTfIdfFunction(broadcastModelKey, getInputCol()), + outputTypeInfo); + }); + + return new Table[] {tEnv.fromDataStream(output)}; + } + + @Override + public IDFModel setModelData(Table... inputs) { + modelDataTable = inputs[0]; + return this; + } + + @Override + public Table[] getModelData() { + return new Table[] {modelDataTable}; + } + + @Override + public Map<Param<?>, Object> getParamMap() { + return paramMap; + } + + @Override + public void save(String path) throws IOException { + ReadWriteUtils.saveMetadata(this, path); + ReadWriteUtils.saveModelData( + IDFModelData.getModelDataStream(modelDataTable), + path, + new IDFModelData.ModelDataEncoder()); + } + + /** + * Loads model data from path. + * + * @param tEnv Stream table environment. + * @param path Model path. + * @return IDFModel model. + */ + public static IDFModel load(StreamTableEnvironment tEnv, String path) throws IOException { + IDFModel model = ReadWriteUtils.loadStageParam(path); + + Table modelDataTable = + ReadWriteUtils.loadModelData(tEnv, path, new IDFModelData.ModelDataDecoder()); + return model.setModelData(modelDataTable); + } + + /** Computes the tf-idf for each term in the input document. */ Review Comment: It might be better to explain what "tf" stands for somewhere in this class or `IDF`. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
