Package org.apache.tika.pipes.reporter.fs

Class FileSystemJsonlReporter

java.lang.Object

org.apache.tika.plugins.AbstractTikaExtension

org.apache.tika.pipes.reporters.PipesReporterBase

org.apache.tika.pipes.reporter.fs.FileSystemJsonlReporter

All Implemented Interfaces:

Closeable, AutoCloseable, PipesReporter, TikaExtension

public class FileSystemJsonlReporter
extends PipesReporterBase

Append-only per-document audit log: one JSON object per line for every result accepted by the includes/excludes filter. The line's id is the FetchEmitTuple.getId() verbatim, so consumers join on it.

Each line is written and flushed to the OS synchronously in report(org.apache.tika.pipes.api.FetchEmitTuple, org.apache.tika.pipes.api.PipesResult, long), so it survives the driver process dying (not a host crash; there is no fsync). A write failure (disk full, etc.) throws from that report(org.apache.tika.pipes.api.FetchEmitTuple, org.apache.tika.pipes.api.PipesResult, long) and every later one rather than dropping lines silently.

Field Summary

Fields inherited from class org.apache.tika.plugins.AbstractTikaExtension

pluginConfig

Constructor Summary

Constructors

Constructor	Description
FileSystemJsonlReporter(ExtensionConfig pluginConfig, FileSystemJsonlReporterConfig config)

Method Summary

Modifier and Type	Method	Description
static FileSystemJsonlReporter	build(ExtensionConfig pluginConfig)
void	close()
void	error(String msg)	This is called if the process has crashed.
void	error(Throwable t)	This is called if the process has crashed.
void	report(FetchEmitTuple t, PipesResult result, long elapsed)
void	report(TotalCountResult totalCountResult)	Make sure to override PipesReporter.supportsTotalCount() to return true
boolean	supportsTotalCount()	Override this if your reporter supports total count.

Methods inherited from class org.apache.tika.pipes.reporters.PipesReporterBase

accept

Methods inherited from class org.apache.tika.plugins.AbstractTikaExtension

getExtensionConfig

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface org.apache.tika.plugins.TikaExtension

getExtensionConfig

Constructor Details

FileSystemJsonlReporter

public FileSystemJsonlReporter(ExtensionConfig pluginConfig,
 FileSystemJsonlReporterConfig config)
                        throws TikaConfigException,
IOException

Throws:

TikaConfigException

IOException

Method Details

build

public static FileSystemJsonlReporter build(ExtensionConfig pluginConfig)
                                     throws TikaConfigException,
IOException

Throws:

TikaConfigException

IOException

report

public void report(FetchEmitTuple t,
 PipesResult result,
 long elapsed)

report

public void report(TotalCountResult totalCountResult)

Description copied from interface: PipesReporter

Make sure to override PipesReporter.supportsTotalCount() to return true

Parameters:

totalCountResult -

supportsTotalCount

public boolean supportsTotalCount()

Description copied from interface: PipesReporter

Override this if your reporter supports total count.

Returns:

false as the baseline implementation

error

public void error(Throwable t)

Description copied from interface: PipesReporter

This is called if the process has crashed. Implementers should not rely on close() to be called after this.

Parameters:

t -

error

public void error(String msg)

Description copied from interface: PipesReporter

This is called if the process has crashed. Implementers should not rely on close() to be called after this.

Parameters:

msg -

close

public void close()
           throws IOException

Throws:

IOException