com.yahoo.prelude

Class IndexFacts

java.lang.Object

com.yahoo.prelude.IndexFacts

public class IndexFacts
extends Object

A central repository for information about indices. Standard usage is

 IndexFacts.Session session = indexFacts.newSession(query); // once when starting to process a query
 session.getIndex(indexName).[get index info]
 

Author:

Steinar Knutsen

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
private static class	IndexFacts.DocumentTypeListOffset
class	IndexFacts.Session Create an instance of this to look up index facts with a given query.

Field Summary

Fields

Modifier and Type	Field and Description
private Map<String,List<String>>	clusterByDocument
private Map<String,List<String>>	clusters A map of document types contained in each cluster indexed by cluster name
private boolean	frozen
private boolean	hasNGramIndices Whether this has (any) NGram indexes.
private Map<String,SearchDefinition>	searchDefinitions A Map of all known search definitions indexed by name
static String	unionName Deprecated. do not use
private SearchDefinition	unionSearchDefinition A search definition which contains the union of all settings.

Constructor Summary

Constructors

Constructor and Description
IndexFacts()
IndexFacts(IndexModel indexModel)

Method Summary

Modifier and Type	Method and Description
private static void	addEntry(Map<String,List<String>> result, String key, String value)
void	addIndex(String sdName, Index index) Deprecated. set indexes at creation time instead
void	addIndex(String sdName, String indexName) Deprecated. set indexes at creation time instead
private IndexFacts.DocumentTypeListOffset	chooseSearchDefinition(List<String> documentTypes, int index) Chooses the correct search definition, default if in doubt.
List<String>	clustersHavingSearchDefinition(String searchDefinitionName)
private Collection<String>	emptyCollectionIfNull(Collection<String> collection)
private void	ensureNotFrozen()
void	freeze() Freeze this to prevent further changes.
private String	getCanonicNameFromDocumentTypes(String indexName, List<String> documentTypes)
String	getDefaultPosition(String sdName)
private Index	getIndexByCanonicNameFromDocumentTypes(String canonicName, List<String> documentTypes)
private Index	getIndexFromDocumentTypes(String indexName, List<String> documentTypes)
boolean	hasNGramIndices() Whether this contains any index which has isNGram()==true.
private static Map<String,List<String>>	invert(Map<String,List<String>> clusters)
boolean	isFrozen()
private boolean	isIndexFromDocumentTypes(String indexName, List<String> documentTypes)
private boolean	isInitialized()
IndexFacts.Session	newSession(Collection<String> sources, Collection<String> restrict)
IndexFacts.Session	newSession(Collection<String> sources, Collection<String> restrict, Set<String> candidateDocumentTypes)
IndexFacts.Session	newSession(Query query)
private Set<String>	resolveDocumentTypes(Collection<String> sources, Collection<String> restrict, Set<String> candidateDocumentTypes) Given a search list which is a mixture of document types and cluster names, and a restrict list which is a list of document types, return a set of all valid document types for this combination.
private Set<String>	resolveDocumentTypes(Query query) Calls resolveDocumentTypes(query.getModel().getSources(), query.getModel().getRestrict())
void	setClusters(Map<String,List<String>> clusters) Public only for testing.
private void	setMasterClusters(Map<String,List<String>> clusters)
void	setSearchDefinitions(Map<String,SearchDefinition> searchDefinitions, SearchDefinition unionSearchDefinition) Deprecated. set indexes at creation time instead

Methods inherited from class java.lang.Object

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

Field Detail

clusterByDocument

private Map<String,List<String>> clusterByDocument

searchDefinitions

private Map<String,SearchDefinition> searchDefinitions

A Map of all known search definitions indexed by name

clusters

private Map<String,List<String>> clusters

A map of document types contained in each cluster indexed by cluster name

unionName

@Deprecated
public static final String unionName

Deprecated. do not use

The name of the default search definition, which is the union of all known document types.

See Also:

Constant Field Values

unionSearchDefinition

private SearchDefinition unionSearchDefinition

A search definition which contains the union of all settings.

frozen

private boolean frozen

hasNGramIndices

private boolean hasNGramIndices

Whether this has (any) NGram indexes. Calculated at freeze time.

Constructor Detail

IndexFacts

public IndexFacts()

IndexFacts

public IndexFacts(IndexModel indexModel)

Method Detail

setMasterClusters

private void setMasterClusters(Map<String,List<String>> clusters)

invert

private static Map<String,List<String>> invert(Map<String,List<String>> clusters)

addEntry

private static void addEntry(Map<String,List<String>> result,
                             String key,
                             String value)

clustersHavingSearchDefinition

public List<String> clustersHavingSearchDefinition(String searchDefinitionName)

setClusters

public void setClusters(Map<String,List<String>> clusters)

Public only for testing.

setSearchDefinitions

@Deprecated
public void setSearchDefinitions(Map<String,SearchDefinition> searchDefinitions,
                                             SearchDefinition unionSearchDefinition)

Deprecated. set indexes at creation time instead

isInitialized

private boolean isInitialized()

isIndexFromDocumentTypes

private boolean isIndexFromDocumentTypes(String indexName,
                                         List<String> documentTypes)

getCanonicNameFromDocumentTypes

private String getCanonicNameFromDocumentTypes(String indexName,
                                               List<String> documentTypes)

getIndexFromDocumentTypes

private Index getIndexFromDocumentTypes(String indexName,
                                        List<String> documentTypes)

getIndexByCanonicNameFromDocumentTypes

private Index getIndexByCanonicNameFromDocumentTypes(String canonicName,
                                                     List<String> documentTypes)

resolveDocumentTypes

private Set<String> resolveDocumentTypes(Query query)

Calls resolveDocumentTypes(query.getModel().getSources(), query.getModel().getRestrict())

resolveDocumentTypes

private Set<String> resolveDocumentTypes(Collection<String> sources,
                                         Collection<String> restrict,
                                         Set<String> candidateDocumentTypes)

Given a search list which is a mixture of document types and cluster names, and a restrict list which is a list of document types, return a set of all valid document types for this combination. Most use-cases for fetching index settings will involve calling this method with the the incoming query's Model.getSources() and Model.getRestrict() as input parameters before calling any other method of this class.

Parameters:

sources - the search list for a query

restrict - the restrict list for a query

Returns:

a (possibly empty) set of valid document types

emptyCollectionIfNull

private Collection<String> emptyCollectionIfNull(Collection<String> collection)

chooseSearchDefinition

private IndexFacts.DocumentTypeListOffset chooseSearchDefinition(List<String> documentTypes,
                                                                 int index)

Chooses the correct search definition, default if in doubt.

Returns:

the search definition to use

freeze

public void freeze()

Freeze this to prevent further changes.

hasNGramIndices

public boolean hasNGramIndices()

Whether this contains any index which has isNGram()==true. This is free to ask on a frozen instance.

isFrozen

public boolean isFrozen()

Returns:

whether it is permissible to update this object

ensureNotFrozen

private void ensureNotFrozen()

addIndex

@Deprecated
public void addIndex(String sdName,
                                 String indexName)

Deprecated. set indexes at creation time instead

Add a string to be accepted as an index name when parsing a query. For testing only.

Parameters:

sdName - name of search definition containing index, if null, modify default set

indexName - name of index, actual or otherwise

addIndex

@Deprecated
public void addIndex(String sdName,
                                 Index index)

Deprecated. set indexes at creation time instead

Adds an index to the specified index, and the default index settings, overriding any current settings for this index

getDefaultPosition

public String getDefaultPosition(String sdName)

newSession

public IndexFacts.Session newSession(Query query)

newSession

public IndexFacts.Session newSession(Collection<String> sources,
                                     Collection<String> restrict)

newSession

public IndexFacts.Session newSession(Collection<String> sources,
                                     Collection<String> restrict,
                                     Set<String> candidateDocumentTypes)