org.apache.tika.batch.fs.builders

Class FSCrawlerBuilder

java.lang.Object

org.apache.tika.batch.fs.builders.FSCrawlerBuilder

All Implemented Interfaces:

ICrawlerBuilder, ObjectFromDOMAndQueueBuilder<FileResourceCrawler>

public class FSCrawlerBuilder
extends java.lang.Object
implements ICrawlerBuilder

Builds either an FSDirectoryCrawler or an FSListCrawler.

Field Summary

Fields

Modifier and Type	Field and Description
private static java.lang.String	CRAWL_ORDER
private static java.lang.String	EXCLUDE_FILE_PAT_ATTR
private static java.lang.String	INCLUDE_FILE_PAT_ATTR
private static java.lang.String	INPUT_DIR_ATTR
private static java.lang.String	INPUT_START_DIR_ATTR
private static java.lang.String	MAX_CONSEC_WAIT_MILLIS
private static java.lang.String	MAX_FILE_SIZE_BYTES_ATTR
private static java.lang.String	MAX_FILES_TO_ADD_ATTR
private static java.lang.String	MAX_FILES_TO_CONSIDER_ATTR
private static java.lang.String	MIN_FILE_SIZE_BYTES_ATTR

Constructor Summary

Constructors

Constructor and Description
FSCrawlerBuilder()

Method Summary

Modifier and Type	Method and Description
FileResourceCrawler	build(org.w3c.dom.Node node, java.util.Map<java.lang.String,java.lang.String> runtimeAttributes, java.util.concurrent.ArrayBlockingQueue<FileResource> queue)
private DocumentSelector	buildSelector(java.util.Map<java.lang.String,java.lang.String> attributes)
private FSDirectoryCrawler.CRAWL_ORDER	getCrawlOrder(java.lang.String s)

Methods inherited from class java.lang.Object

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

Field Detail

MAX_CONSEC_WAIT_MILLIS

private static final java.lang.String MAX_CONSEC_WAIT_MILLIS

See Also:

Constant Field Values

MAX_FILES_TO_ADD_ATTR

private static final java.lang.String MAX_FILES_TO_ADD_ATTR

See Also:

Constant Field Values

MAX_FILES_TO_CONSIDER_ATTR

private static final java.lang.String MAX_FILES_TO_CONSIDER_ATTR

See Also:

Constant Field Values

CRAWL_ORDER

private static final java.lang.String CRAWL_ORDER

See Also:

Constant Field Values

INPUT_DIR_ATTR

private static final java.lang.String INPUT_DIR_ATTR

See Also:

Constant Field Values

INPUT_START_DIR_ATTR

private static final java.lang.String INPUT_START_DIR_ATTR

See Also:

Constant Field Values

MAX_FILE_SIZE_BYTES_ATTR

private static final java.lang.String MAX_FILE_SIZE_BYTES_ATTR

See Also:

Constant Field Values

MIN_FILE_SIZE_BYTES_ATTR

private static final java.lang.String MIN_FILE_SIZE_BYTES_ATTR

See Also:

Constant Field Values

INCLUDE_FILE_PAT_ATTR

private static final java.lang.String INCLUDE_FILE_PAT_ATTR

See Also:

Constant Field Values

EXCLUDE_FILE_PAT_ATTR

private static final java.lang.String EXCLUDE_FILE_PAT_ATTR

See Also:

Constant Field Values

Constructor Detail

FSCrawlerBuilder

public FSCrawlerBuilder()

Method Detail

build

public FileResourceCrawler build(org.w3c.dom.Node node,
                                 java.util.Map<java.lang.String,java.lang.String> runtimeAttributes,
                                 java.util.concurrent.ArrayBlockingQueue<FileResource> queue)

Specified by:

build in interface ICrawlerBuilder

Specified by:

build in interface ObjectFromDOMAndQueueBuilder<FileResourceCrawler>

getCrawlOrder

private FSDirectoryCrawler.CRAWL_ORDER getCrawlOrder(java.lang.String s)

buildSelector

private DocumentSelector buildSelector(java.util.Map<java.lang.String,java.lang.String> attributes)