com.atlassian.confluence.search.v2

Interface SearchIndexAccessor

All Known Implementing Classes:

DelegatingSearchIndexAccessor, EdgeSearchIndexAccessor, LuceneSearchIndexAccessor

@ExperimentalApi
public interface SearchIndexAccessor

Since:

7.16

Method Summary

Modifier and Type	Method and Description
void	execute(SearchIndexAction action) Executes a SearchIndexAction on the underlying search index.
int	numDocs() Get the total number of index documents for the underlying search index
void	optimize() Manually optimize the index for search by merging segments.
void	reset(Runnable replaceIndex) Execute a method after the index has been closed and before the index has been reopened
long	scan(SearchQuery searchQuery, Set<String> requestedFields, Consumer<Map<String,String[]>> consumer) Scans the index for documents that match the provided searchQuery.
long	scan(SearchQuery searchQuery, Set<String> requestedFields, Consumer<ScannedDocument> consumer, float defaultScore) Scans the index for documents that match the provided searchQuery.
SearchResults	search(ISearch search, Set<String> requestedFields) Perform a search with a given criteria, the returns searchResults only have the fields requested in the projection filled out, no other fields are valid in the searchResult.
void	snapshot(File destinationDirectory) Take a snapshot of the current index and store it into the destination directory
void	withBatchUpdate(BatchUpdateAction batchUpdateAction) Execute the update action in a batch

Method Detail

search

SearchResults search(ISearch search,
                     Set<String> requestedFields)
              throws InvalidSearchException

Perform a search with a given criteria, the returns searchResults only have the fields requested in the projection filled out, no other fields are valid in the searchResult. SearchResults will throw an FieldNotRequestedException if an attempt is made to access a field value that has not had it's key included in the projection Set.

Will throw an InvalidSearchException if some error occurred converting the search object into a search that could be performed in the back-end. Usually this is a sign that some plugin that was used to create the search is no longer available.

Parameters:

search - the search to perform

requestedFields - requested fields. Passing null will request all fields, and passing empty set will request none.

Returns:

a SearchResults instance

Throws:

InvalidSearchException

scan

long scan(SearchQuery searchQuery,
          Set<String> requestedFields,
          Consumer<Map<String,String[]>> consumer)

Scans the index for documents that match the provided searchQuery. The requestedFields set is used to specify which values need to be returned. The consumer will collect all the results found by the scan.

Note that it is possible for a document to match the query, but not have any of the requested fields. This will result in an empty String[] for those fields that don't exist on the document. If you don't want to return those documents, you should specify a consumer that filters these out.

Parameters:

searchQuery - SearchQuery to scan the index for.

requestedFields - Set of field names that are requested from the index. Any fields provided here that is not on the document will return an empty String[] When providing an empty set, the consumer will only get empty maps to consume.

consumer - Consumer that handles each found document.

Returns:

the number of documents that matched the searchQuery.

Throws:

SearchIndexAccessException - if there is an IO error.

scan

long scan(SearchQuery searchQuery,
          Set<String> requestedFields,
          Consumer<ScannedDocument> consumer,
          float defaultScore)

Scans the index for documents that match the provided searchQuery. The requestedFields set is used to specify which values need to be returned. The consumer will collect all the results found by the scan and convert them into ScannedDocument, which contains a score (if scoreIncluded is true) and a field-value map of the requested fields.

Note that it is possible for a document to match the query, but not have any of the requested fields as the other scan method.

This will result in an empty String[] for those fields that don't exist on the document. If you don't want to return those documents, you should specify a consumer that filters these out.

Parameters:

searchQuery - SearchQuery to scan the index for.

requestedFields - Set of field names that are requested from the index. Any fields provided here that is not on the document will return an empty String[] When providing an empty set, the consumer will only get empty maps to consume.

consumer - Consumer that handles each found document.

defaultScore - a default score to be included in the ScannedDocument if it is not possible to retrieve the score for the document.

Returns:

the number of documents that matched the searchQuery.

Throws:

SearchIndexAccessException - if there is an IO error.

execute

void execute(SearchIndexAction action)

Executes a SearchIndexAction on the underlying search index.

Parameters:

action - SearchIndexAction

Throws:

SearchIndexAccessException - if there is an io error.

numDocs

int numDocs()

Get the total number of index documents for the underlying search index

Returns:

the total number of index documents

withBatchUpdate

void withBatchUpdate(BatchUpdateAction batchUpdateAction)

Execute the update action in a batch

Parameters:

batchUpdateAction - the batch update action

snapshot

void snapshot(File destinationDirectory)
       throws SearchIndexAccessException

Take a snapshot of the current index and store it into the destination directory

Parameters:

destinationDirectory - the destination directory

Throws:

SearchIndexAccessException - if the snapshot could not be taken

Since:

7.16

reset

void reset(Runnable replaceIndex)

Execute a method after the index has been closed and before the index has been reopened

Parameters:

replaceIndex - the method that needs to be executed

optimize

void optimize()

Manually optimize the index for search by merging segments. Is generally horribly inefficient and rarely justified. Added for the usage-tracking-plugin index as it was using optimize before the Lucene Upgrade to prevent file handles from running out with a quickly ever-growing index which never has deletions. Please carefully consider whether to use this at all, it has the potential, especially on large instances to lock index usage for a long period of time. Per-segment search performance improvements and default merge policies should make this no longer needed. If you need it, the index was probably poorly designed.

Since:

7.17