Develop Reference

file.listFiles() does not list recently created file

I have a component that should do something every time a file in the watched directory is created,updated or deleted.
For the sake of simplicity, lets say it prints the file names to Sysout.
#Component
public class MyComponent {
#Autowired
MyComponentWatcher myComponentWatcher
#PostConstruct
public void init(){
myComponentWatcher.registerWatchService();
}
}
#Component
public class MyComponentWatcher {
#Value("${somePath}")
private String myPath;
private File myDir;
#PostConstruct
public void init() {
this.myDir = new File(myPath);
}
#Async
public void registerWatchService() throws IOException, InterruptedException {
WatchService watchService
= FileSystems.getDefault().newWatchService();
Path path = myDir.toPath();
path.register(
watchService,
StandardWatchEventKinds.ENTRY_CREATE,
StandardWatchEventKinds.ENTRY_DELETE,
StandardWatchEventKinds.ENTRY_MODIFY);
WatchKey key;
while ((key = watchService.take()) != null) {
for (WatchEvent<?> event : key.pollEvents()) {
System.out.println(
"Event kind:" + event.kind()
+ ". File affected: " + event.context() + ".");
}
key.reset();
File[] files = myDir.listFiles();
System.out.println(files); //!!!
}
}
}
GIVEN my dir contains 1 file
WHEN I create another file
THEN System.out.println(files); returns the old and the new file
-> I only get the old file!
I wrote a test case that uses Files.write(filePath, fileContent.getBytes(), CREATE_NEW); to create the test file and
#AfterEach
void tearDown() {
File toDelete = new File(filePath.toString());
toDelete.delete();
}
to delete the test file after the test.
In total the watcher triggers 3 times with the following lists of events:
[ENTRY_CREATE, ENTRY_MODIFY]
[ENTRY_MODIFY]
[ENTRY_MODIFY]
[ENTRY_DELETE]

Your code is working fine as expected and written.
to achieve your output, you need to recreate the object for File class.
// we are creating the object just after the constructor
#PostConstruct
public void init() {
this.myDir = new File(myPath);
}
but then you modified your path but not the actual instance variable:
please change your code File[] files = new File(myPath).listFiles();

Read two lines of a file at once in a flink streaming process

I want to process files with a flink stream in which two lines belong together. In the first line there is a header and in the second line a corresponding text.
The files are located on my local file system. I am using the readFile(fileInputFormat, path, watchType, interval, pathFilter, typeInfo) method with a custom FileInputFormat.
My streaming job class looks like this:
final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
DataStream<Read> inputStream = env.readFile(new ReadInputFormatTest("path/to/monitored/folder"), "path/to/monitored/folder", FileProcessingMode.PROCESS_CONTINUOUSLY, 100);
inputStream.print();
env.execute("Flink Streaming Java API Skeleton");
and my ReadInputFormatTest like this:
public class ReadInputFormatTest extends FileInputFormat<Read> {
private transient FileSystem fileSystem;
private transient BufferedReader reader;
private final String inputPath;
private String headerLine;
private String readLine;
public ReadInputFormatTest(String inputPath) {
this.inputPath = inputPath;
}
#Override
public void open(FileInputSplit inputSplit) throws IOException {
FileSystem fileSystem = getFileSystem();
this.reader = new BufferedReader(new InputStreamReader(fileSystem.open(inputSplit.getPath())));
this.headerLine = reader.readLine();
this.readLine = reader.readLine();
}
private FileSystem getFileSystem() {
if (fileSystem == null) {
try {
fileSystem = FileSystem.get(new URI(inputPath));
} catch (URISyntaxException | IOException e) {
throw new RuntimeException(e);
}
}
return fileSystem;
}
#Override
public boolean reachedEnd() throws IOException {
return headerLine == null;
}
#Override
public Read nextRecord(Read r) throws IOException {
r.setHeader(headerLine);
r.setSequence(readLine);
headerLine = reader.readLine();
readLine = reader.readLine();
return r;
}
}
As expected, the headers and the text are stored together in one object. However, the file is read eight times. So the problem is the parallelization. Where and how can I specify that a file is processed only once, but several files in parallel?
Or do I have to change my custom FileInputFormat even further?

I would modify your source to emit the available filenames (instead of the actual file contents) and then add a new processor to read a name from the input stream and then emit pairs of lines. In other words, split the current source into a source followed by a processor. The processor can be made to run at any degree of parallelism and the source would be a single instance.

Java - Runnable, lambda function, and methods of class

I'm quite new with Java (studied on University but was version 2).
Now I've developed an application that downloads files from s3 in parallel. I've used ExecutorService and Runnable to download multiple files in parallel in this way:
public class DownloaderController {
private AmazonS3 s3Client;
private ExecutorService fixedPool;
private TransferManager dlManager;
private List<MultipleFileDownload> downloads = new ArrayList<>();
public DownloaderController() {
checkForNewWork();
}
public void checkForNewWork(){
Provider1 provider = new Provider1();
fixedPool = Executors.newFixedThreadPool(4);
List<Download> providedDownloadList = provider.toBeDownloaded();
for (Download temp : providedDownloadList) {
if (!downloadData.contains(temp)) {
fixedPool.submit(download.downloadCompletedHandler(s3Client));
}
}
}
}
public void printToTextArea(String msg){
Date now = new Date();
if ( !DateUtils.isSameDay(this.lastLogged, now)){
this._doLogRotate();
}
this.lastLogged = now;
SimpleDateFormat ft = new SimpleDateFormat("dd/MM/yyyy H:mm:ss");
String output = "[ " + ft.format(now) + " ] " + msg + System.getProperty("line.separator");
Platform.runLater(() -> {
//this is a FXML object
statusTextArea.appendText(output);
});
}
}
public class Provider1 implements downloadProvider {
}
public class Download {
abstract Runnable downloadCompletedHandler(AmazonS3 s3Client);
}
public class DownloadProvider1 extends Download {
#Override
public Runnable downloadCompletedHandler(AmazonS3 s3Client){
Runnable downloadwork = () -> {
ObjectListing list = s3Client.listObjects(this.bucket,this.getFolder());
List<S3ObjectSummary> objects = list.getObjectSummaries();
AtomicLong workSize = new AtomicLong(0);
List<DeleteObjectsRequest.KeyVersion> keys = new ArrayList<>();
objects.forEach(obj -> {
workSize.getAndAdd(obj.getSize());
keys.add((new DeleteObjectsRequest.KeyVersion(obj.getKey())));
});
MultipleFileDownload fileDownload = dlManager.downloadDirectory("myBucket","folder","outputDirectory");
try {
fileDownload.waitForCompletion();
} catch (Exception e){
printToTextArea("Exception while download from AmazonS3");
}
};
return downloadwork;
}
}
In the downloadController i call every minute a function that adds some Download objects to a List that contains folders that has to be downloaded from s3. when a new Download is added it's also added to ExecutorService pool. The Download object returns the code that has to be executed to download the folder from s3 and what to do when it's download is finished.
My problem is, what is the best way to communicate between the Runnable and the DownloadController ?

Your code does not make entirely clear what the goal is. From what I understand, I would have done it something like this:
public class Download {
private AmazonS3 s3Client;
public Download(AmazonS2 client) { s3Client = client; }
public void run() { // perform download }
}
That class does nothing but download the file (cfg Separation of Concern) and is a Runnable. You can do executorService.submit(new Download(client)) and the download will be finished eventually; also, you can test it without being called concurrently.
Now, you want a callback method for logging it being finished.
public class LoggingCallback {
public void log() {
System.out.println("finished");
}
}
Also a Runnable (the method doesn't have to be run()).
And, to make sure it's triggered one after the other, maybe
class OneAfterTheOther {
private Runnable first;
private Runnable second;
public OneAfterTheOther(Runnable r1, Runnable r2) {
first = r1; second = r2;
}
public void run() { first.run(); second.run(); }
}
which if submitted like this
Download dl = new Download(client);
Logger l = new LoggingCallback();
executorService.submit(new OneAfterTheOther(dl::run, l::log));
will do what I think you're trying to do.

Java threads - waiting on all child threads in order to proceed

So a little background;
I am working on a project in which a servlet is going to release crawlers upon a lot of text files within a file system. I was thinking of dividing the load under multiple threads, for example:
a crawler enters a directory, finds 3 files and 6 directories. it will start processing the files and start a thread with a new crawler for the other directories. So from my creator class I would create a single crawler upon a base directory. The crawler would assess the workload and if deemed needed it would spawn another crawler under another thread.
My crawler class looks like this
package com.fujitsu.spider;
import java.io.BufferedReader;
import java.io.File;
import java.io.FileReader;
import java.io.IOException;
import java.io.Serializable;
import java.util.ArrayList;
public class DocumentSpider implements Runnable, Serializable {
private static final long serialVersionUID = 8401649393078703808L;
private Spidermode currentMode = null;
private String URL = null;
private String[] terms = null;
private float score = 0;
private ArrayList<SpiderDataPair> resultList = null;
public enum Spidermode {
FILE, DIRECTORY
}
public DocumentSpider(String resourceURL, Spidermode mode, ArrayList<SpiderDataPair> resultList) {
currentMode = mode;
setURL(resourceURL);
this.setResultList(resultList);
}
#Override
public void run() {
try {
if (currentMode == Spidermode.FILE) {
doCrawlFile();
} else {
doCrawlDirectory();
}
} catch (Exception e) {
e.printStackTrace();
}
System.out.println("SPIDER # " + URL + " HAS FINISHED.");
}
public Spidermode getCurrentMode() {
return currentMode;
}
public void setCurrentMode(Spidermode currentMode) {
this.currentMode = currentMode;
}
public String getURL() {
return URL;
}
public void setURL(String uRL) {
URL = uRL;
}
public void doCrawlFile() throws Exception {
File target = new File(URL);
if (target.isDirectory()) {
throw new Exception(
"This URL points to a directory while the spider is in FILE mode. Please change this spider to FILE mode.");
}
procesFile(target);
}
public void doCrawlDirectory() throws Exception {
File baseDir = new File(URL);
if (!baseDir.isDirectory()) {
throw new Exception(
"This URL points to a FILE while the spider is in DIRECTORY mode. Please change this spider to DIRECTORY mode.");
}
File[] directoryContent = baseDir.listFiles();
for (File f : directoryContent) {
if (f.isDirectory()) {
DocumentSpider spider = new DocumentSpider(f.getPath(), Spidermode.DIRECTORY, this.resultList);
spider.terms = this.terms;
(new Thread(spider)).start();
} else {
DocumentSpider spider = new DocumentSpider(f.getPath(), Spidermode.FILE, this.resultList);
spider.terms = this.terms;
(new Thread(spider)).start();
}
}
}
public void procesDirectory(String target) throws IOException {
File base = new File(target);
File[] directoryContent = base.listFiles();
for (File f : directoryContent) {
if (f.isDirectory()) {
procesDirectory(f.getPath());
} else {
procesFile(f);
}
}
}
public void procesFile(File target) throws IOException {
BufferedReader br = new BufferedReader(new FileReader(target));
String line;
while ((line = br.readLine()) != null) {
String[] words = line.split(" ");
for (String currentWord : words) {
for (String a : terms) {
if (a.toLowerCase().equalsIgnoreCase(currentWord)) {
score += 1f;
}
;
if (currentWord.toLowerCase().contains(a)) {
score += 1f;
}
;
}
}
}
br.close();
resultList.add(new SpiderDataPair(this, URL));
}
public String[] getTerms() {
return terms;
}
public void setTerms(String[] terms) {
this.terms = terms;
}
public float getScore() {
return score;
}
public void setScore(float score) {
this.score = score;
}
public ArrayList<SpiderDataPair> getResultList() {
return resultList;
}
public void setResultList(ArrayList<SpiderDataPair> resultList) {
this.resultList = resultList;
}
}
The problem I am facing is that in my root crawler I have this list of results from every crawler that I want to process further. The operation to process the data from this list is called from the servlet (or main method for this example). However the operations is always called before all of the crawlers have completed their processing. thus launching the operation to process the results too soon, which leads to incomplete data.
I tried solving this using the join methods but unfortunately I cant seems to figure this one out.
package com.fujitsu.spider;
import java.util.ArrayList;
import com.fujitsu.spider.DocumentSpider.Spidermode;
public class Main {
public static void main(String[] args) throws InterruptedException {
ArrayList<SpiderDataPair> results = new ArrayList<SpiderDataPair>();
String [] terms = {"SERVER","CHANGE","MO"};
DocumentSpider spider1 = new DocumentSpider("C:\\Users\\Mark\\workspace\\Spider\\Files", Spidermode.DIRECTORY, results);
spider1.setTerms(terms);
DocumentSpider spider2 = new DocumentSpider("C:\\Users\\Mark\\workspace\\Spider\\File2", Spidermode.DIRECTORY, results);
spider2.setTerms(terms);
Thread t1 = new Thread(spider1);
Thread t2 = new Thread(spider2);
t1.start();
t1.join();
t2.start();
t2.join();
for(SpiderDataPair d : spider1.getResultList()){
System.out.println("PATH -> " + d.getFile() + " SCORE -> " + d.getSpider().getScore());
}
for(SpiderDataPair d : spider2.getResultList()){
System.out.println("PATH -> " + d.getFile() + " SCORE -> " + d.getSpider().getScore());
}
}
}
TL:DR
I really wish to understand this subject so any help would be immensely appreciated!.

You need a couple of changes in your code:
In the spider:
List<Thread> threads = new LinkedList<Thread>();
for (File f : directoryContent) {
if (f.isDirectory()) {
DocumentSpider spider = new DocumentSpider(f.getPath(), Spidermode.DIRECTORY, this.resultList);
spider.terms = this.terms;
Thread thread = new Thread(spider);
threads.add(thread)
thread.start();
} else {
DocumentSpider spider = new DocumentSpider(f.getPath(), Spidermode.FILE, this.resultList);
spider.terms = this.terms;
Thread thread = new Thread(spider);
threads.add(thread)
thread.start();
}
}
for (Thread thread: threads) thread.join()
The idea is to create a new thread for each spider and start it. Once they are all running, you wait until each on is done before the Spider itself finishes. This way each spider thread keeps running until all of its work is done (thus the top thread runs until all children and their children are finished).
You also need to change your runner so that it runs the two spiders in parallel instead of one after another like this:
Thread t1 = new Thread(spider1);
Thread t2 = new Thread(spider2);
t1.start();
t2.start();
t1.join();
t2.join();

You should use a higher-level library than bare Thread for this task. I would suggest looking into ExecutorService in particular and all of java.util.concurrent generally. There are abstractions there that can manage all of the threading issues while providing well-formed tasks a properly protected environment in which to run.
For your specific problem, I would recommend some sort of blocking queue of tasks and a standard producer-consumer architecture. Each task knows how to determine if its path is a file or directory. If it is a file, process the file; if it is a directory, crawl the directory's immediate contents and enqueue new tasks for each sub-path. You could also use some properly-synchronized shared state to cap the number of files processed, depth, etc. Also, the service provides the ability to await termination of its tasks, making the "join" simpler.
With this architecture, you decouple the notion of threads and thread management (handled by the ExecutorService) with your business logic of tasks (typically a Runnable or Callable). The service itself has the ability to tune how to instantiate, such as a fixed maximum number of threads or a scalable number depending on how many concurrent tasks exist (See factory methods on java.util.concurrent.Executors). Threads, which are more expensive than the Runnables they execute, are re-used to conserve resources.
If your objective is primarily something functional that works in production quality, then the library is the way to go. However, if your objective is to understand the lower-level details of thread management, then you may want to investigate the use of latches and perhaps thread groups to manage them at a lower level, exposing the details of the implementation so you can work with the details.

Can I watch for single file change with WatchService (not the whole directory)?

When I'm trying to register a file instead of a directory java.nio.file.NotDirectoryException is thrown. Can I listen for a single file change, not the whole directory?

Just filter the events for the file you want in the directory:
final Path path = FileSystems.getDefault().getPath(System.getProperty("user.home"), "Desktop");
System.out.println(path);
try (final WatchService watchService = FileSystems.getDefault().newWatchService()) {
final WatchKey watchKey = path.register(watchService, StandardWatchEventKinds.ENTRY_MODIFY);
while (true) {
final WatchKey wk = watchService.take();
for (WatchEvent<?> event : wk.pollEvents()) {
//we only register "ENTRY_MODIFY" so the context is always a Path.
final Path changed = (Path) event.context();
System.out.println(changed);
if (changed.endsWith("myFile.txt")) {
System.out.println("My file has changed");
}
}
// reset the key
boolean valid = wk.reset();
if (!valid) {
System.out.println("Key has been unregisterede");
}
}
}
Here we check whether the changed file is "myFile.txt", if it is then do whatever.

Other answers are right that you must watch a directory and filter for your particular file. However, you probably want a thread running in the background. The accepted answer can block indefinitely on watchService.take(); and doesn't close the WatchService. A solution suitable for a separate thread might look like:
public class FileWatcher extends Thread {
private final File file;
private AtomicBoolean stop = new AtomicBoolean(false);
public FileWatcher(File file) {
this.file = file;
}
public boolean isStopped() { return stop.get(); }
public void stopThread() { stop.set(true); }
public void doOnChange() {
// Do whatever action you want here
}
#Override
public void run() {
try (WatchService watcher = FileSystems.getDefault().newWatchService()) {
Path path = file.toPath().getParent();
path.register(watcher, StandardWatchEventKinds.ENTRY_MODIFY);
while (!isStopped()) {
WatchKey key;
try { key = watcher.poll(25, TimeUnit.MILLISECONDS); }
catch (InterruptedException e) { return; }
if (key == null) { Thread.yield(); continue; }
for (WatchEvent<?> event : key.pollEvents()) {
WatchEvent.Kind<?> kind = event.kind();
#SuppressWarnings("unchecked")
WatchEvent<Path> ev = (WatchEvent<Path>) event;
Path filename = ev.context();
if (kind == StandardWatchEventKinds.OVERFLOW) {
Thread.yield();
continue;
} else if (kind == java.nio.file.StandardWatchEventKinds.ENTRY_MODIFY
&& filename.toString().equals(file.getName())) {
doOnChange();
}
boolean valid = key.reset();
if (!valid) { break; }
}
Thread.yield();
}
} catch (Throwable e) {
// Log or rethrow the error
}
}
}
I tried working from the accepted answer and this article. You should be able to use this thread with new FileWatcher(new File("/home/me/myfile")).start() and stop it by calling stopThread() on the thread.

No it isn't possible to register a file, the watch service doesn't work this way. But registering a directory actually watches changes on the directory children (the files and sub-directories), not the changes on the directory itself.
If you want to watch a file, then you register the containing directory with the watch service. Path.register() documentation says:
WatchKey java.nio.file.Path.register(WatchService watcher, Kind[] events, Modifier...
modifiers) throws IOException
Registers the file located by this path with a watch service.
In this release, this path locates a directory that exists. The directory is registered with the watch service so that entries in the directory can be watched
Then you need to process events on entries, and detect those related to the file you are interested in, by checking the context value of the event. The context value represents the name of the entry (actually the path of the entry relatively to the path of its parent, which is exactly the child name). You have an example here.

Apache offers a FileWatchDog class with a doOnChange method.
private class SomeWatchFile extends FileWatchdog {
protected SomeWatchFile(String filename) {
super(filename);
}
#Override
protected void doOnChange() {
fileChanged= true;
}
}
And where ever you want you can start this thread:
SomeWatchFile someWatchFile = new SomeWatchFile (path);
someWatchFile.start();
The FileWatchDog class polls a file's lastModified() timestamp. The native WatchService from Java NIO is more efficient, since notifications are immediate.

You cannot watch an individual file directly but you can filter out what you don't need.
Here is my FileWatcher class implementation:
import java.io.File;
import java.nio.file.*;
import java.nio.file.WatchEvent.Kind;
import static java.nio.file.StandardWatchEventKinds.*;
public abstract class FileWatcher
{
private Path folderPath;
private String watchFile;
public FileWatcher(String watchFile)
{
Path filePath = Paths.get(watchFile);
boolean isRegularFile = Files.isRegularFile(filePath);
if (!isRegularFile)
{
// Do not allow this to be a folder since we want to watch files
throw new IllegalArgumentException(watchFile + " is not a regular file");
}
// This is always a folder
folderPath = filePath.getParent();
// Keep this relative to the watched folder
this.watchFile = watchFile.replace(folderPath.toString() + File.separator, "");
}
public void watchFile() throws Exception
{
// We obtain the file system of the Path
FileSystem fileSystem = folderPath.getFileSystem();
// We create the new WatchService using the try-with-resources block
try (WatchService service = fileSystem.newWatchService())
{
// We watch for modification events
folderPath.register(service, ENTRY_MODIFY);
// Start the infinite polling loop
while (true)
{
// Wait for the next event
WatchKey watchKey = service.take();
for (WatchEvent<?> watchEvent : watchKey.pollEvents())
{
// Get the type of the event
Kind<?> kind = watchEvent.kind();
if (kind == ENTRY_MODIFY)
{
Path watchEventPath = (Path) watchEvent.context();
// Call this if the right file is involved
if (watchEventPath.toString().equals(watchFile))
{
onModified();
}
}
}
if (!watchKey.reset())
{
// Exit if no longer valid
break;
}
}
}
}
public abstract void onModified();
}
To use this, you just have to extend and implement the onModified() method like so:
import java.io.File;
public class MyFileWatcher extends FileWatcher
{
public MyFileWatcher(String watchFile)
{
super(watchFile);
}
#Override
public void onModified()
{
System.out.println("Modified!");
}
}
Finally, start watching the file:
String watchFile = System.getProperty("user.home") + File.separator + "Desktop" + File.separator + "Test.txt";
FileWatcher fileWatcher = new MyFileWatcher(watchFile);
fileWatcher.watchFile();

Not sure about others, but I groan at the amount of code needed to watch a single file for changes using the basic WatchService API. It has to be simpler!
Here are a couple of alternatives using third party libraries:
Using Apache Commons Configuration
Using spring-loaded package from the Spring Framework (didn't find an example implementation for this off-hand, but it looks straight-forward to use)

I have created a wrapper around Java 1.7's WatchService that allows registering a directory and any number of glob patterns. This class will take care of the filtering and only emit events you are interested in.
try {
DirectoryWatchService watchService = new SimpleDirectoryWatchService(); // May throw
watchService.register( // May throw
new DirectoryWatchService.OnFileChangeListener() {
#Override
public void onFileCreate(String filePath) {
// File created
}
#Override
public void onFileModify(String filePath) {
// File modified
}
#Override
public void onFileDelete(String filePath) {
// File deleted
}
},
<directory>, // Directory to watch
<file-glob-pattern-1>, // E.g. "*.log"
<file-glob-pattern-2>, // E.g. "input-?.txt"
<file-glob-pattern-3>, // E.g. "config.ini"
... // As many patterns as you like
);
watchService.start(); // The actual watcher runs on a new thread
} catch (IOException e) {
LOGGER.error("Unable to register file change listener for " + fileName);
}
Complete code is in this repo.

I extended the solution by BullyWiiPlaza a bit, for integration with javafx.concurrent, e.g. javafx.concurrent.Taskand javafx.concurrent.Service.
Also I added possibility to track multiple files.
Task:
import javafx.concurrent.Task;
import lombok.extern.slf4j.Slf4j;
import java.io.File;
import java.nio.file.*;
import java.util.*;
import static java.nio.file.StandardWatchEventKinds.ENTRY_MODIFY;
#Slf4j
public abstract class FileWatcherTask extends Task<Void> {
static class Entry {
private final Path folderPath;
private final String watchFile;
Entry(Path folderPath, String watchFile) {
this.folderPath = folderPath;
this.watchFile = watchFile;
}
#Override
public boolean equals(Object o) {
if (this == o) return true;
if (o == null || getClass() != o.getClass()) return false;
Entry entry = (Entry) o;
return Objects.equals(folderPath, entry.folderPath) && Objects.equals(watchFile, entry.watchFile);
}
#Override
public int hashCode() {
return Objects.hash(folderPath, watchFile);
}
}
private final List<Entry> entryList;
private final Map<WatchKey, Entry> watchKeyEntryMap;
public FileWatcherTask(Iterable<String> watchFiles) {
this.entryList = new ArrayList<>();
this.watchKeyEntryMap = new LinkedHashMap<>();
for (String watchFile : watchFiles) {
Path filePath = Paths.get(watchFile);
boolean isRegularFile = Files.isRegularFile(filePath);
if (!isRegularFile) {
// Do not allow this to be a folder since we want to watch files
throw new IllegalArgumentException(watchFile + " is not a regular file");
}
// This is always a folder
Path folderPath = filePath.getParent();
// Keep this relative to the watched folder
watchFile = watchFile.replace(folderPath.toString() + File.separator, "");
Entry entry = new Entry(folderPath, watchFile);
entryList.add(entry);
log.debug("Watcher initialized for {} entries. ({})", entryList.size(), entryList.stream().map(e -> e.watchFile + "-" + e.folderPath).findFirst().orElse("<>"));
}
}
public FileWatcherTask(String... watchFiles) {
this(Arrays.asList(watchFiles));
}
public void watchFile() throws Exception {
// We obtain the file system of the Path
// FileSystem fileSystem = folderPath.getFileSystem();
// TODO: use the actual file system instead of default
FileSystem fileSystem = FileSystems.getDefault();
// We create the new WatchService using the try-with-resources block
try (WatchService service = fileSystem.newWatchService()) {
log.debug("Watching filesystem {}", fileSystem);
for (Entry e : entryList) {
// We watch for modification events
WatchKey key = e.folderPath.register(service, ENTRY_MODIFY);
watchKeyEntryMap.put(key, e);
}
// Start the infinite polling loop
while (true) {
// Wait for the next event
WatchKey watchKey = service.take();
for (Entry e : entryList) {
// Call this if the right file is involved
var hans = watchKeyEntryMap.get(watchKey);
if (hans != null) {
for (WatchEvent<?> watchEvent : watchKey.pollEvents()) {
// Get the type of the event
WatchEvent.Kind<?> kind = watchEvent.kind();
if (kind == ENTRY_MODIFY) {
Path watchEventPath = (Path) watchEvent.context();
onModified(e.watchFile);
}
if (!watchKey.reset()) {
// Exit if no longer valid
log.debug("Watch key {} was reset", watchKey);
break;
}
}
}
}
}
}
}
#Override
protected Void call() throws Exception {
watchFile();
return null;
}
public abstract void onModified(String watchFile);
}
Service:
public abstract class FileWatcherService extends Service<Void> {
private final Iterable<String> files;
public FileWatcherService(Iterable<String> files) {
this.files = files;
}
#Override
protected Task<Void> createTask() {
return new FileWatcherTask(files) {
#Override
public void onModified(String watchFile) {
FileWatcherService.this.onModified(watchFile);
}
};
}
abstract void onModified(String watchFile);
}