[GitHub] flink pull request: Refactoring the File Monitoring Source.

[zentol <gi...@git.apache.org>]

Github user zentol commented on a diff in the pull request:

    https://github.com/apache/flink/pull/1929#discussion_r60927783
  
    --- Diff: flink-streaming-java/src/main/java/org/apache/flink/streaming/api/functions/source/FileSplitMonitoringFunction.java ---
    @@ -0,0 +1,259 @@
    +/*
    + * 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.streaming.api.functions.source;
    +
    +import org.apache.flink.api.common.io.FileInputFormat;
    +import org.apache.flink.configuration.Configuration;
    +import org.apache.flink.core.fs.FileInputSplit;
    +import org.apache.flink.core.fs.FileStatus;
    +import org.apache.flink.core.fs.FileSystem;
    +import org.apache.flink.core.fs.Path;
    +import org.apache.flink.runtime.JobException;
    +import org.apache.flink.util.Preconditions;
    +import org.slf4j.Logger;
    +import org.slf4j.LoggerFactory;
    +
    +import java.io.IOException;
    +import java.net.URI;
    +import java.util.ArrayList;
    +import java.util.HashSet;
    +import java.util.List;
    +import java.util.Set;
    +
    +/**
    + * This is the single (non-parallel) task, that monitors a user-procided path, and assigns splits
    + * to downstream tasks for further reading and processing, depending on the user-provided {@link FileSplitMonitoringFunction.WatchType}.
    + *
    + * This method keeps track of which splits have already being processed by which task, and at which point
    + * in the file we are currently processing, at the granularity of the split. In addition, it keeps track
    + * of the last modification time for each file, so that it can detect new data.
    + */
    +public class FileSplitMonitoringFunction<OUT>
    +	extends RichSourceFunction<FileInputSplit> {
    +
    +	private static final long serialVersionUID = 1L;
    +
    +	private static final Logger LOG = LoggerFactory.getLogger(FileSplitMonitoringFunction.class);
    +
    +	/**
    +	 * Specifies when computation will be triggered. This can be currently done in 3 ways.
    +	 * <li/>{@code ONLY_NEW_FILES} which implies that only new files will be processed.
    +	 * <li/>{@code REPROCESS_WITH_APPENDED} which reprocesses the whole file, as soon as new data is appended to it.
    +	 */
    +	public enum WatchType {
    +		REPROCESS_WITH_APPENDED		// Reprocesses the whole file when new data is appended.
    +	}
    +
    +	/** The path to monitor. */
    +	private final String path;
    +
    +	/** The default parallelism for the job, as this is going to be the parallelism of the downstream readers. */
    +	private final int readerParallelism;
    +
    +	/** The {@link FileInputFormat} to be read. */
    +	private FileInputFormat<OUT> format;
    +
    +	/** How often to monitor the state of the directory for new data. */
    +	private final long interval;
    +
    +	/** Which new data to process (see {@link WatchType}. */
    +	private final WatchType watchType;
    +
    +	private long globalModificationTime;
    +
    +	private FilePathFilter pathFilter;
    +
    +	private volatile boolean isRunning = true;
    +
    +	public FileSplitMonitoringFunction(
    +		FileInputFormat<OUT> format, String path,
    +		WatchType watchType, int readerParallelism, long interval) {
    +
    +		this(format, path, FilePathFilter.DefaultFilter.getInstance(), watchType, readerParallelism, interval);
    +	}
    +
    +	public FileSplitMonitoringFunction(
    +		FileInputFormat<OUT> format, String path, FilePathFilter filter,
    +		WatchType watchType, int readerParallelism, long interval) {
    +
    +		this.format = Preconditions.checkNotNull(format);
    +		this.path = Preconditions.checkNotNull(path);
    +
    +		Preconditions.checkArgument(interval >= 100,
    +			"The specified monitoring interval is smaller than the minimum allowed one (100 ms).");
    +		this.interval = interval;
    +
    +		this.watchType = watchType;
    +
    +		this.pathFilter = Preconditions.checkNotNull(filter);
    +
    +		this.readerParallelism = Math.max(readerParallelism, 1);
    +		this.globalModificationTime = Long.MIN_VALUE;
    +	}
    +
    +	@Override
    +	@SuppressWarnings("unchecked")
    +	public void open(Configuration parameters) throws Exception {
    +		super.open(parameters);
    +		format.configure(parameters);
    +	}
    +
    +	/**
    +	 * Creates the input splits for the path to be assigned to the downstream tasks.
    +	 * Those are going to read their contents for further processing.
    +	 */
    +	private FileInputSplit[] createInputSplits() throws JobException {
    +		FileInputSplit[] inputSplits;
    +		try {
    +			inputSplits = format.createInputSplits(readerParallelism);
    +		} catch (Throwable t) {
    +			throw new JobException("Creating the input splits caused an error: " + t.getMessage(), t);
    +		}
    +		return inputSplits;
    +	}
    +
    +	@Override
    +	public void run(SourceFunction.SourceContext<FileInputSplit> context) throws Exception {
    +		FileSystem fileSystem = FileSystem.get(new URI(path));
    +		while (isRunning) {
    +			monitor(fileSystem, context);
    +			Thread.sleep(interval);
    +		}
    +	}
    +
    +	private void monitor(FileSystem fs, SourceContext<FileInputSplit> context) throws IOException, JobException {
    +		List<FileStatus> files = listEligibleFiles(fs);
    +		for (FileInputSplit split : getInputSplitsToForward(files)) {
    +			processSplit(split, context);
    +		}
    +	}
    +
    +	private void processSplit(FileInputSplit split, SourceContext<FileInputSplit> context) {
    +		if (watchType == WatchType.REPROCESS_WITH_APPENDED) {
    +			context.collect(split);
    +		} else {
    +			throw new RuntimeException("Unknown WatchType" + watchType);
    +		}
    +	}
    +
    +	private Set<FileInputSplit> getInputSplitsToForward(List<FileStatus> files) throws JobException {
    +		if(files.isEmpty()) {
    +			return new HashSet<>();
    +		}
    +
    +		FileInputSplit[] inputSplits = createInputSplits();
    +
    +		Set<FileInputSplit> splitsToForward = new HashSet<>();
    +		for (FileInputSplit split: inputSplits) {
    +			for(FileStatus file: files) {
    +				if (file.getPath().equals(split.getPath())) {
    +					splitsToForward.add(split);
    +					break;
    +				}
    +			}
    +		}
    +		return splitsToForward;
    +	}
    +
    +	/**
    +	 * Returns the files that have data to be processed. This method returns the
    +	 * Paths to the aforementioned files. It is up to the {@link #processSplit(FileInputSplit, SourceContext)}
    +	 * method to decide which parts of the file to be processed, and forward them downstream.
    +	 */
    +	private List<FileStatus> listEligibleFiles(FileSystem fileSystem) throws IOException {
    +		List<FileStatus> files = new ArrayList<>();
    +
    +		FileStatus[] statuses = fileSystem.listStatus(new Path(path));
    +		long maxModificationTime = Long.MIN_VALUE;
    +		if (statuses == null) {
    +			LOG.warn("Path does not exist: {}", path);
    +		} else {
    +			// handle the new files
    +			for (FileStatus status : statuses) {
    +				Path filePath = status.getPath();
    +				long modificationTime = status.getModificationTime();
    +				if (!shouldIgnore(filePath, modificationTime)) {
    +					files.add(status);
    +					if(modificationTime > maxModificationTime) {
    +						maxModificationTime = modificationTime;
    +					}
    +				}
    +			}
    +		}
    +
    +		// after finding the eligible files to process, update the
    +		// global max modification time seen so far to reflect the
    +		// latest monitoring round.
    +		if(maxModificationTime > globalModificationTime) {
    +			globalModificationTime = maxModificationTime;
    +		}
    +
    +		return files;
    +	}
    +
    +	/**
    +	 * Returns {@code true} if the file is NOT to be processed further.
    +	 * This happens in the following cases:
    +	 *
    +	 * If the user-specified path filtering method returns {@code true} for the file,
    +	 * or if the modification time of the file is smaller than the {@link #globalModificationTime}, which
    +	 * signals the point in time where we last checked the path for new data.
    +	 */
    +	private boolean shouldIgnore(Path filePath, long modificationTime) {
    +		return (pathFilter != null && pathFilter.filterPaths(filePath)) || modificationTime <= globalModificationTime;
    +	}
    +
    +	@Override
    +	public void close() throws Exception {
    +		super.close();
    +		format.close();
    +	}
    +
    +	@Override
    +	public void cancel() {
    +		isRunning = false;
    +	}
    +//
    +//	/**
    +//	 * The default file path filtering method and is used
    +//	 * if no other such function is provided. This filter leaves out
    +//	 * files starting with ".", "_", and "_COPYING_".
    --- End diff --
    
    leftover code?


---
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 infrastructure@apache.org or file a JIRA ticket
with INFRA.
---