oracle.ifs.core.extraction.util

Class NlpSolverSearchHelper

java.lang.Object

oracle.ifs.core.extraction.util.NlpSolverSearchHelper

public class NlpSolverSearchHelper
extends Object

An NlpSolverSearchHelper performs utilities for assisting in searching via Oracle text during an NlpSolver.solve() process.

Author:

tschnell

(Unpublished)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	NlpSolverSearchHelper.SearchType Enumeration for the various search types.

Field Summary

Fields

Modifier and Type	Field and Description
static String	RESERVED_TEXT_CHARACTERS The reserved characters.
static Set	s_ReservedWords Oracle Text reserved words, will be escaped in 'simple' inxire iO token strings.

Method Summary

Modifier and Type	Method and Description
static String	getContainsClauseTerm(String searchTerm, NlpSolverSearchHelper.SearchType type) Return the text string to use in full text searches, conforming to the database's syntax.
static String	getContainsClauseTermOracle(String searchTerm, NlpSolverSearchHelper.SearchType type) Return the Oracle text string to use in the CONTAINS clause, based on the specified term and the type of search requested.
static String	getContainsClauseTermPostgres(String searchTerm, NlpSolverSearchHelper.SearchType type) Return the Postgres text string to use in to_tsquery() function, based on the specified term and the type of search requested.
static String	normalizeSearchPhrase(String searchPhrase) Normalize a full text search phrase for entity phrases.
static String	normalizeSearchTermOracle(String ctxString) Normalize a search string for context search, i.e. replace double quotes with curly parentheses '{}', escape all special characters as given by reservedCharacters with '\', and finally replace '*' by '%' and '?'
protected static String	normalizeToken(String token) Normalize given token.
protected static String[]	tokenizeSearchTermOracle(String searchString) Split context search string into tokens.
protected static String[]	tokenizeSearchTermPostgres(String searchString) Split search term into tokens.

Methods inherited from class java.lang.Object

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

Field Detail

RESERVED_TEXT_CHARACTERS

public static final String RESERVED_TEXT_CHARACTERS

The reserved characters.

See Also:

Constant Field Values

s_ReservedWords

public static final Set s_ReservedWords

Oracle Text reserved words, will be escaped in 'simple' inxire iO token strings.

Method Detail

normalizeSearchPhrase

public static String normalizeSearchPhrase(String searchPhrase)

Normalize a full text search phrase for entity phrases.

Parameters:

searchPhrase - the search phrase

Returns:

the normalized search phrase

getContainsClauseTerm

public static String getContainsClauseTerm(String searchTerm,
                                           NlpSolverSearchHelper.SearchType type)

Return the text string to use in full text searches, conforming to the database's syntax.

For Oracle, the term will be used in a CONTAINS clause. For Postgres, the term will be used in a to_tsquery() function.

Parameters:

searchTerm - the BASIC search term

searchType - the search type

Returns:

the EXPANDED search term

(Unpublished)

getContainsClauseTermOracle

public static String getContainsClauseTermOracle(String searchTerm,
                                                 NlpSolverSearchHelper.SearchType type)

Return the Oracle text string to use in the CONTAINS clause, based on the specified term and the type of search requested.

Parameters:

searchTerm - the BASIC search term

searchType - the search type

Returns:

the EXPANDED search term

(Unpublished)

getContainsClauseTermPostgres

public static String getContainsClauseTermPostgres(String searchTerm,
                                                   NlpSolverSearchHelper.SearchType type)

Return the Postgres text string to use in to_tsquery() function, based on the specified term and the type of search requested.

Parameters:

searchTerm - the BASIC search term

searchType - the search type

Returns:

the EXPANDED search term

(Unpublished)

tokenizeSearchTermOracle

protected static String[] tokenizeSearchTermOracle(String searchString)

Split context search string into tokens. Tokens are space separated and may be escaped by double quotes or grouped by the '(' and ')' characters. A quoted phrase or a group of tokens will be returned as single token, including the quote of group characters.

Any token must contain at least one non-reserved characters (?), or it will be removed from list.

Tokens containing any Oracle Text reserved word, like AND, will be escaped.

Examples:

• car garden : car | garden
• car* near(garden, flower) : car* | near(garden, flower)
• me and you : me | "and" | you
• :

Parameters:

searchString - context search string containing list of tokens

Returns:

the list of tokens as array, may be of size 0

tokenizeSearchTermPostgres

protected static String[] tokenizeSearchTermPostgres(String searchString)

Split search term into tokens.

Parameters:

searchString - context search string containing list of tokens

Returns:

the list of tokens as array, may be of size 0

normalizeSearchTermOracle

public static String normalizeSearchTermOracle(String ctxString)

Normalize a search string for context search, i.e. replace double quotes with curly parentheses '{}', escape all special characters as given by reservedCharacters with '\', and finally replace '*' by '%' and '?' by '_'.

If no search string is found, null is returned.

For complex searches all Oracle Text commands, like "and", "or" and "near" may be used or the special characters '"', '*', '?', . All other characters with special meaning for Oracle Text (InterMedia) are escaped.

Warning: the new version does not remove leading and trailing wildcards automaticly. Call com.inxire.ioc.entity.AdvSearchParameter#ensureTwoCharacters(String) to avoid search strings with less than two characters.

Parameters:

ctxString - search string, with '*' and '?' as wildcards

Returns:

the normalized string

normalizeToken

protected static String normalizeToken(String token)

Normalize given token. A token must a least contain one text character, or it will be removed.

Any Oracle text reserved word will be escaped by double quotes.

Parameters:

token - some token as used for inxire iO content search

Returns:

the normalized token, either unchanged, escaped or cleared (NULL)