rdblue commented on code in PR #12298: URL: https://github.com/apache/iceberg/pull/12298#discussion_r2748532516
########## core/src/main/java/org/apache/iceberg/formats/FormatModelRegistry.java: ########## @@ -0,0 +1,463 @@ +/* + * 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.iceberg.formats; + +import java.io.IOException; +import java.nio.ByteBuffer; +import java.util.List; +import java.util.Map; +import java.util.Objects; +import java.util.stream.Collectors; +import java.util.stream.IntStream; +import org.apache.iceberg.DataFile; +import org.apache.iceberg.DeleteFile; +import org.apache.iceberg.FileContent; +import org.apache.iceberg.FileFormat; +import org.apache.iceberg.Metrics; +import org.apache.iceberg.MetricsConfig; +import org.apache.iceberg.PartitionSpec; +import org.apache.iceberg.Schema; +import org.apache.iceberg.SortOrder; +import org.apache.iceberg.StructLike; +import org.apache.iceberg.common.DynMethods; +import org.apache.iceberg.deletes.EqualityDeleteWriter; +import org.apache.iceberg.deletes.PositionDelete; +import org.apache.iceberg.deletes.PositionDeleteWriter; +import org.apache.iceberg.encryption.EncryptedOutputFile; +import org.apache.iceberg.encryption.EncryptionKeyMetadata; +import org.apache.iceberg.io.DataWriter; +import org.apache.iceberg.io.FileAppender; +import org.apache.iceberg.io.FileWriter; +import org.apache.iceberg.io.InputFile; +import org.apache.iceberg.relocated.com.google.common.annotations.VisibleForTesting; +import org.apache.iceberg.relocated.com.google.common.base.Preconditions; +import org.apache.iceberg.relocated.com.google.common.collect.ImmutableList; +import org.apache.iceberg.relocated.com.google.common.collect.Maps; +import org.apache.iceberg.util.Pair; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +/** + * A registry that manages file-format-specific readers and writers through a unified object model + * factory interface. + * + * <p>This registry provides access to {@link ReadBuilder}s for data consumption and {@link + * FileWriterBuilder}s for writing various types of Iceberg content files. The appropriate builder + * is selected based on {@link FileFormat} and object model name. + * + * <p>{@link FormatModel} objects are registered through {@link #register(FormatModel)} and used for + * creating readers and writers. Read builders are returned directly from the factory. Write + * builders may be wrapped in specialized content file writer implementations depending on the + * requested builder type. + */ +public final class FormatModelRegistry { + private static final Logger LOG = LoggerFactory.getLogger(FormatModelRegistry.class); + // The list of classes which are used for registering the reader and writer builders + private static final List<String> CLASSES_TO_REGISTER = + ImmutableList.of( + "org.apache.iceberg.data.GenericFormatModels", + "org.apache.iceberg.arrow.vectorized.ArrowFormatModels", + "org.apache.iceberg.flink.data.FlinkFormatModels", + "org.apache.iceberg.spark.source.SparkFormatModels"); + + // Format models indexed by file format and object model class + private static final Map<Pair<FileFormat, Class<?>>, FormatModel<?, ?>> MODELS = + Maps.newConcurrentMap(); + + static { + registerSupportedFormats(); + } + + /** + * Registers an {@link FormatModel} in this registry. + * + * <p>The {@link FormatModel} creates readers and writers for a specific combinations of file + * format (Parquet, ORC, Avro) and object model (for example: "generic", "spark", "flink", etc.). + * Registering custom factories allows integration of new data processing engines for the + * supported file formats with Iceberg's file access mechanisms. + * + * <p>Each factory must be uniquely identified by its combination of file format and object model + * name. This uniqueness constraint prevents ambiguity when selecting factories for read and write + * operations. + * + * @param formatModel the factory implementation to register + * @throws IllegalArgumentException if a factory is already registered for the combination of + * {@link FormatModel#format()} and {@link FormatModel#type()} + */ + public static synchronized void register(FormatModel<?, ?> formatModel) { + Pair<FileFormat, Class<?>> key = Pair.of(formatModel.format(), formatModel.type()); + + FormatModel<?, ?> existing = MODELS.get(key); + Preconditions.checkArgument( + existing == null, + "Cannot register %s: %s is registered for format=%s type=%s schemaType=%s", + formatModel.getClass(), + existing == null ? null : existing.getClass(), + key.first(), + key.second(), + existing == null ? null : existing.schemaType()); + + MODELS.put(key, formatModel); + } + + /** + * Returns a reader builder for the specified file format and object model. + * + * <p>The returned {@link ReadBuilder} provides a fluent interface for configuring how data is + * read from the input file and converted to the output objects. + * + * @param format the file format (Parquet, Avro, ORC) that determines the parsing implementation + * @param type the output type + * @param inputFile source file to read data from + * @param <D> the type of data records the reader will produce + * @param <S> the type of the output schema for the reader + * @return a configured reader builder for the specified format and object model + */ + public static <D, S> ReadBuilder<D, S> readBuilder( + FileFormat format, Class<? extends D> type, InputFile inputFile) { + FormatModel<D, S> model = modelFor(format, type); + return model.readBuilder(inputFile); + } + + /** + * Returns a writer builder for generating a {@link DataFile}. + * + * <p>The returned builder produces a writer that accepts records defined by the specified object + * model and persists them using the provided file format. Unlike basic writers, this writer + * collects file metadata during the writing process and generates a {@link DataFile} that can be + * used for table operations. + * + * @param format the file format used for writing + * @param type the input type + * @param outputFile destination for the written data + * @param <D> the type of data records the writer will accept + * @param <S> the type of the input schema for the writer + * @return a configured data write builder for creating a {@link DataWriter} + */ + public static <D, S> FileWriterBuilder<DataWriter<D>, S> dataWriteBuilder( + FileFormat format, Class<? extends D> type, EncryptedOutputFile outputFile) { + FormatModel<D, S> model = modelFor(format, type); + return forDataFile( + model.writeBuilder(outputFile), outputFile.encryptingOutputFile().location(), format); + } + + /** + * Creates a writer builder for generating a {@link DeleteFile} with equality deletes. + * + * <p>The returned builder produces a writer that accepts records defined by the specified object + * model and persists them using the given file format. The writer persists equality delete + * records that identify rows to be deleted based on the configured equality fields, producing a + * {@link DeleteFile} that can be used for table operations. + * + * @param format the file format used for writing + * @param type the input type + * @param outputFile destination for the written data + * @param equalityFieldIds the field IDs that define the equality delete columns + * @param <D> the type of data records the writer will accept + * @param <S> the type of the input schema for the writer + * @return a configured delete write builder for creating an {@link EqualityDeleteWriter} + */ + public static <D, S> FileWriterBuilder<EqualityDeleteWriter<D>, S> equalityDeleteWriteBuilder( + FileFormat format, Class<D> type, EncryptedOutputFile outputFile, int... equalityFieldIds) { + FormatModel<D, S> model = modelFor(format, type); + return forEqualityDelete( + model.writeBuilder(outputFile), + outputFile.encryptingOutputFile().location(), + format, + equalityFieldIds); + } + + /** + * Creates a writer builder for generating a {@link DeleteFile} with position-based deletes. + * + * <p>The returned builder produces a writer that accepts records defined by the specified object + * model and persists them using the given file format. The writer accepts {@link PositionDelete} + * records that identify rows to be deleted by file path and position, producing a {@link + * DeleteFile} that can be used for table operations. + * + * @param format the file format used for writing + * @param outputFile destination for the written data + * @return a configured delete write builder for creating a {@link PositionDeleteWriter} + */ + @SuppressWarnings("rawtypes") + public static <D> FileWriterBuilder<PositionDeleteWriter<D>, ?> positionDeleteWriteBuilder( + FileFormat format, EncryptedOutputFile outputFile) { + FormatModel<PositionDelete, ?> model = modelFor(format, PositionDelete.class); + return forPositionDelete( + model.writeBuilder(outputFile), outputFile.encryptingOutputFile().location(), format); + } + + @VisibleForTesting + static Map<Pair<FileFormat, Class<?>>, FormatModel<?, ?>> models() { + return MODELS; + } + + @SuppressWarnings("unchecked") + private static <D, S> FormatModel<D, S> modelFor(FileFormat format, Class<? extends D> type) { + FormatModel<D, S> model = (FormatModel<D, S>) MODELS.get(Pair.of(format, type)); + Preconditions.checkArgument( + model != null, "Format model is not registered for format %s and type %s", format, type); + return model; + } + + @SuppressWarnings("CatchBlockLogException") + private static void registerSupportedFormats() { + // Uses dynamic methods to call the `register` for the listed classes + for (String classToRegister : CLASSES_TO_REGISTER) { + try { + DynMethods.builder("register").impl(classToRegister).buildStaticChecked().invoke(); + } catch (NoSuchMethodException e) { + // failing to register a factory is normal and does not require a stack trace + LOG.info( + "Skip registration of {}. Likely the jar is not in the classpath", classToRegister); + } + } + } + + private static <D, S> FileWriterBuilder<DataWriter<D>, S> forDataFile( + WriteBuilder<D, S> writeBuilder, String location, FileFormat format) { + return new FileWriterBuilderImpl<>( + writeBuilder.content(FileContent.DATA), + location, + format, + builder -> { + Preconditions.checkArgument( + builder.spec != null, "Cannot create data writer without spec"); + Preconditions.checkArgument( + builder.spec.isUnpartitioned() || builder.partition != null, + "Partition must not be null when creating data writer for partitioned spec"); + + return new DataWriter<>( + builder.writeBuilder.build(), + builder.format, + builder.location, + builder.spec, + builder.partition, + builder.keyMetadata, + builder.sortOrder); + }); + } + + private static <D, S> FileWriterBuilder<EqualityDeleteWriter<D>, S> forEqualityDelete( + WriteBuilder<D, S> writeBuilder, String location, FileFormat format, int[] equalityFieldIds) { + return new FileWriterBuilderImpl<>( + writeBuilder.content(FileContent.EQUALITY_DELETES), + location, + format, + builder -> { + Preconditions.checkState( + builder.schema != null, "Cannot create equality delete file without a schema"); + Preconditions.checkState( + equalityFieldIds != null, + "Cannot create equality delete file without delete field ids"); + Preconditions.checkArgument( + builder.spec != null, "Spec must not be null when creating equality delete writer"); + Preconditions.checkArgument( + builder.spec.isUnpartitioned() || builder.partition != null, + "Partition must not be null for partitioned writes"); + + return new EqualityDeleteWriter<>( + builder + .writeBuilder + .schema(builder.schema) + .meta("delete-type", "equality") + .meta( + "delete-field-ids", + IntStream.of(equalityFieldIds) + .mapToObj(Objects::toString) + .collect(Collectors.joining(", "))) + .build(), + builder.format, + builder.location, + builder.spec, + builder.partition, + builder.keyMetadata, + builder.sortOrder, + equalityFieldIds); + }); + } + + @SuppressWarnings({"unchecked", "rawtypes"}) + private static <D> FileWriterBuilder<PositionDeleteWriter<D>, ?> forPositionDelete( + WriteBuilder<PositionDelete, ?> writeBuilder, String location, FileFormat format) { + return new FileWriterBuilderImpl<>( + (WriteBuilder<PositionDelete, Object>) writeBuilder.content(FileContent.POSITION_DELETES), + location, + format, + builder -> { + Preconditions.checkArgument( + builder.spec != null, "Spec must not be null when creating position delete writer"); Review Comment: Minor: Iceberg error messages should be short and clear. In cases like this, we typically use "Invalid partition spec for position delete writer: null". Also, "spec" is not specific enough so we should always use "partition spec". -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
