-
public final class CrawlFilterTODO : configurable
-
-
Nested Class Summary
Nested Classes Modifier and Type Class Description public classCrawlFilter.Companion
-
Field Summary
Fields Modifier and Type Field Description private final PageCategorypageTypeprivate StringurlFilterprivate TextFiltertextFilterprivate BlockFilterblockFilterprivate StringstartKeyprivate StringendKeyprivate StringreversedStartKeyprivate StringreversedEndKeyprivate final ImmutableConfigconf
-
Constructor Summary
Constructors Constructor Description CrawlFilter(ImmutableConfig conf)
-
Method Summary
-
-
Method Detail
-
getPageType
final PageCategory getPageType()
-
getUrlFilter
final String getUrlFilter()
-
setUrlFilter
final Unit setUrlFilter(String urlFilter)
-
getTextFilter
final TextFilter getTextFilter()
-
setTextFilter
final Unit setTextFilter(TextFilter textFilter)
-
getBlockFilter
final BlockFilter getBlockFilter()
-
setBlockFilter
final Unit setBlockFilter(BlockFilter blockFilter)
-
getStartKey
final String getStartKey()
-
setStartKey
final Unit setStartKey(String startKey)
-
getReversedStartKey
final String getReversedStartKey()
-
setReversedStartKey
final Unit setReversedStartKey(String reversedStartKey)
-
getReversedEndKey
final String getReversedEndKey()
-
setReversedEndKey
final Unit setReversedEndKey(String reversedEndKey)
-
getConf
final ImmutableConfig getConf()
-
filter
final WebPage filter(WebPage page)
-
testUrlSatisfied
final Boolean testUrlSatisfied(String url)
-
testKeyRangeSatisfied
final Boolean testKeyRangeSatisfied(String reversedUrl)
-
testTextSatisfied
final Boolean testTextSatisfied(String text)
-
isDisallowed
final Boolean isDisallowed(Node node)
-
isDetailUrl
final Boolean isDetailUrl(String url)
-
isSearchUrl
final Boolean isSearchUrl(String url)
-
isMediaUrl
final Boolean isMediaUrl(String url)
-
isIndexUrl
final Boolean isIndexUrl(String url)
-
-
-
-