org.apache.tika.fork

Class ForkParser

java.lang.Object

org.apache.tika.parser.AbstractParser

org.apache.tika.fork.ForkParser

All Implemented Interfaces:

java.io.Closeable, java.io.Serializable, java.lang.AutoCloseable, Parser

public class ForkParser
extends AbstractParser
implements java.io.Closeable

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ForkParser()
ForkParser(java.lang.ClassLoader loader)
ForkParser(java.lang.ClassLoader loader, Parser parser)
ForkParser(java.nio.file.Path tikaBin, ParserFactoryFactory factoryFactory) If you have a directory with, say, tike-app.jar and you want the child process/server to build a parser and run it from that -- so that you can keep all of those dependencies out of your client code, use this initializer.
ForkParser(java.nio.file.Path tikaBin, ParserFactoryFactory parserFactoryFactory, java.lang.ClassLoader classLoader) EXPERT

Method Summary

Modifier and Type	Method and Description
void	close()
java.lang.String	getJavaCommand() Deprecated. since 1.8
java.util.List<java.lang.String>	getJavaCommandAsList() Returns the command used to start the forked server process.
int	getPoolSize() Returns the size of the process pool.
java.util.Set<MediaType>	getSupportedTypes(ParseContext context) Returns the set of media types supported by this parser when used with the given parse context.
void	parse(java.io.InputStream stream, org.xml.sax.ContentHandler handler, Metadata metadata, ParseContext context) This sends the objects to the server for parsing, and the server via the proxies acts on the handler as if it were updating it directly.
void	setJavaCommand(java.util.List<java.lang.String> java) Sets the command used to start the forked server process.
void	setJavaCommand(java.lang.String java) Deprecated. since 1.8
void	setMaxFilesProcessedPerServer(int maxFilesProcessedPerClient) If there is a slowly building memory leak in one of the parsers, it is useful to set a limit on the number of files processed by a server before it is shutdown and restarted.
void	setPoolSize(int poolSize) Sets the size of the process pool.
void	setServerParseTimeoutMillis(long serverParseTimeoutMillis) The maximum amount of time allowed for the server to try to parse a file.
void	setServerPulseMillis(long serverPulseMillis) The amount of time in milliseconds that the server should wait before checking to see if the parse has timed out or if the wait has timed out The default is 5 seconds.
void	setServerWaitTimeoutMillis(long serverWaitTimeoutMillis) The maximum amount of time allowed for the server to wait for a new request to parse a file.

Methods inherited from class org.apache.tika.parser.AbstractParser

parse

Methods inherited from class java.lang.Object

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

Constructor Detail

ForkParser

public ForkParser(java.nio.file.Path tikaBin,
                  ParserFactoryFactory factoryFactory)

If you have a directory with, say, tike-app.jar and you want the child process/server to build a parser and run it from that -- so that you can keep all of those dependencies out of your client code, use this initializer.

Parameters:

tikaBin - directory containing the tika-app.jar or similar -- full jar including tika-core and all desired parsers and dependencies

factoryFactory -

ForkParser

public ForkParser(java.nio.file.Path tikaBin,
                  ParserFactoryFactory parserFactoryFactory,
                  java.lang.ClassLoader classLoader)

EXPERT

Parameters:

tikaBin - directory containing the tika-app.jar or similar -- full jar including tika-core and all desired parsers and dependencies

parserFactoryFactory - -- the factory to use to generate the parser factory in the child process/server

classLoader - to use for all classes besides the parser in the child process/server

ForkParser

public ForkParser(java.lang.ClassLoader loader,
                  Parser parser)

Parameters:

loader - The ClassLoader to use

parser - the parser to delegate to. This one cannot be another ForkParser

ForkParser

public ForkParser(java.lang.ClassLoader loader)

ForkParser

public ForkParser()

Method Detail

getPoolSize

public int getPoolSize()

Returns the size of the process pool.

Returns:

process pool size

setPoolSize

public void setPoolSize(int poolSize)

Sets the size of the process pool.

Parameters:

poolSize - process pool size

getJavaCommand

@Deprecated
public java.lang.String getJavaCommand()

Deprecated. since 1.8

Returns the command used to start the forked server process.

Returns:

java command line

See Also:

getJavaCommandAsList()

getJavaCommandAsList

public java.util.List<java.lang.String> getJavaCommandAsList()

Returns the command used to start the forked server process.

Returned list is unmodifiable.

Returns:

java command line args

setJavaCommand

public void setJavaCommand(java.util.List<java.lang.String> java)

Sets the command used to start the forked server process. The arguments "-jar" and "/path/to/bootstrap.jar" or "-cp" and "/path/to/tika_bin" are appended to the given command when starting the process. The default setting is {"java", "-Xmx32m"}.

Creates a defensive copy.

Parameters:

java - java command line

setJavaCommand

@Deprecated
public void setJavaCommand(java.lang.String java)

Deprecated. since 1.8

Sets the command used to start the forked server process. The given command line is split on whitespace and the arguments "-jar" and "/path/to/bootstrap.jar" are appended to it when starting the process. The default setting is "java -Xmx32m".

Parameters:

java - java command line

See Also:

setJavaCommand(List)

getSupportedTypes

public java.util.Set<MediaType> getSupportedTypes(ParseContext context)

Description copied from interface: Parser

Returns the set of media types supported by this parser when used with the given parse context.

Specified by:

getSupportedTypes in interface Parser

Parameters:

context - parse context

Returns:

immutable set of media types

parse

public void parse(java.io.InputStream stream,
                  org.xml.sax.ContentHandler handler,
                  Metadata metadata,
                  ParseContext context)
           throws java.io.IOException,
                  org.xml.sax.SAXException,
                  TikaException

This sends the objects to the server for parsing, and the server via the proxies acts on the handler as if it were updating it directly.

If using a RecursiveParserWrapper, there are two options:

1. Send in a class that extends RecursiveParserWrapperHandler, and the server will proxy back the data as best it can[0].
2. Send in a class that extends AbstractRecursiveParserWrapperHandler and the server will act on the class but not proxy back the data. This can be used, for example, if all you want to do is write to disc, extend AbstractRecursiveParserWrapperHandler to write to disc when AbstractRecursiveParserWrapperHandler.endDocument(ContentHandler, Metadata) is called, and the server will take care of the writing via the handler.

NOTE:[0] "the server will proxy back the data as best it can". If the handler implements Serializable and is actually serializable, the server will send it and the Metadata back upon RecursiveParserWrapperHandler.endEmbeddedDocument(ContentHandler, Metadata) or RecursiveParserWrapperHandler.endEmbeddedDocument(ContentHandler, Metadata). If the handler does not implement Serializable or if there is a NotSerializableException thrown during serialization, the server will call ContentHandler#toString() on the ContentHandler and set that value with the AbstractRecursiveParserWrapperHandler.TIKA_CONTENT key and then serialize and proxy that data back.

Specified by:

parse in interface Parser

Parameters:

stream - the document stream (input)

handler - handler for the XHTML SAX events (output)

metadata - document metadata (input and output)

context - parse context

Throws:

java.io.IOException

org.xml.sax.SAXException

TikaException

close

public void close()

Specified by:

close in interface java.io.Closeable

Specified by:

close in interface java.lang.AutoCloseable

setServerPulseMillis

public void setServerPulseMillis(long serverPulseMillis)

The amount of time in milliseconds that the server should wait before checking to see if the parse has timed out or if the wait has timed out The default is 5 seconds.

Parameters:

serverPulseMillis - milliseconds to sleep before checking if there has been any activity

setServerParseTimeoutMillis

public void setServerParseTimeoutMillis(long serverParseTimeoutMillis)

The maximum amount of time allowed for the server to try to parse a file. If more than this time elapses, the server shuts down, and the ForkParser throws an exception.

Parameters:

serverParseTimeoutMillis -

setServerWaitTimeoutMillis

public void setServerWaitTimeoutMillis(long serverWaitTimeoutMillis)

The maximum amount of time allowed for the server to wait for a new request to parse a file. The server will shutdown after this amount of time, and a new server will have to be started by a new client.

Parameters:

serverWaitTimeoutMillis -

setMaxFilesProcessedPerServer

public void setMaxFilesProcessedPerServer(int maxFilesProcessedPerClient)

If there is a slowly building memory leak in one of the parsers, it is useful to set a limit on the number of files processed by a server before it is shutdown and restarted. Default value is -1.

Parameters:

maxFilesProcessedPerClient - maximum number of files that a server can handle before the parser shuts down a client and creates a new process. If set to -1, the server is never restarted because of the number of files handled.