Package org.apache.tika.parser.inference

Class InferenceUnit

java.lang.Object

org.apache.tika.parser.inference.InferenceUnit

public final class InferenceUnit
extends Object

One thing offered to the dispatcher: the bytes of an image, a page render, a media clip, with the document they came from and, when known, its parent. Captured at the offer so a task can run after the document has closed; the bytes live in a file the dispatcher owns until the flush, and the id paths let the flush find the metadata the recursive wrapper kept once the parser's own objects are no longer read. A task writes its result on the destination, which the dispatcher decides: the parent for an inline part or a page render, the document itself otherwise, and the top-level document for everything when the parse produces one metadata object.

Constructor Summary

Constructors

Constructor	Description
InferenceUnit(MediaType type, Metadata target, Metadata parent, String text)	A InputKind.TEXT unit: the target's extracted text, held in memory.
InferenceUnit(InputKind kind, MediaType type, Metadata target, Metadata parent, Path path)
InferenceUnit(InputKind kind, MediaType type, Metadata target, Metadata parent, Path path, int page)	A unit that is one page of its target: page is 1-based.

Method Summary

Modifier and Type	Method	Description
byte[]	getBytes()	Readable during a task run only; the file is deleted when the flush ends.
Metadata	getDestination()	Where a task writes this unit's results.
InputKind	getKind()
int	getPage()	The 1-based page this unit renders, for InputKind.PAGES; -1 otherwise.
Metadata	getParent()	The document the target is embedded in; null at top level.
String	getParentIdPath()
Path	getPath()
long	getSize()
Metadata	getTarget()	The document the bytes belong to.
String	getTargetIdPath()
String	getText()	The text of a InputKind.TEXT unit; null for the rest.
MediaType	getType()
boolean	isLifted()	Whether the results go on another document than the target, which a locator then names.
static boolean	lifts(Metadata target)	Whether results for this document belong on the one it is part of, not on it.

Methods inherited from class java.lang.Object

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

Constructor Details

InferenceUnit

public InferenceUnit(InputKind kind,
 MediaType type,
 Metadata target,
 Metadata parent,
 Path path)
              throws IOException

Throws:

IOException

InferenceUnit

public InferenceUnit(InputKind kind,
 MediaType type,
 Metadata target,
 Metadata parent,
 Path path,
 int page)
              throws IOException

A unit that is one page of its target: page is 1-based.

Throws:

IOException

InferenceUnit

public InferenceUnit(MediaType type,
 Metadata target,
 Metadata parent,
 String text)

A InputKind.TEXT unit: the target's extracted text, held in memory.

Method Details

lifts

public static boolean lifts(Metadata target)

Whether results for this document belong on the one it is part of, not on it.

getDestination

public Metadata getDestination()

Where a task writes this unit's results.

isLifted

public boolean isLifted()

Whether the results go on another document than the target, which a locator then names.

getText

public String getText()

The text of a InputKind.TEXT unit; null for the rest.

getPage

public int getPage()

The 1-based page this unit renders, for InputKind.PAGES; -1 otherwise.

getKind

public InputKind getKind()

getType

public MediaType getType()

getTarget

public Metadata getTarget()

The document the bytes belong to.

getParent

public Metadata getParent()

The document the target is embedded in; null at top level.

getTargetIdPath

public String getTargetIdPath()

getParentIdPath

public String getParentIdPath()

getPath

public Path getPath()

getSize

public long getSize()

getBytes

public byte[] getBytes()
                throws IOException

Readable during a task run only; the file is deleted when the flush ends.

Throws:

IOException