IndexSearcher Class Reference

#include <index_searcher.hpp>

Public Member Functions
	IndexSearcher ()
virtual	~IndexSearcher ()
MatchedPatternFragment	simpleSearch (boost::shared_ptr< HashGenerator > hashGenerator, boost::shared_ptr< std::vector< sauchar_t > > T, boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > > markers, boost::shared_ptr< std::vector< saidx_t > > SA, const std::string &pattern, bool byWhitespace=false) throw (ConcordiaException)
MatchedPatternFragment	lexiconSearch (boost::shared_ptr< HashGenerator > hashGenerator, boost::shared_ptr< std::vector< sauchar_t > > T, boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > > markers, boost::shared_ptr< std::vector< saidx_t > > SA, const std::string &pattern, bool byWhitespace=false) throw (ConcordiaException)
std::vector< AnubisSearchResult >	anubisSearch (boost::shared_ptr< ConcordiaConfig > config, boost::shared_ptr< HashGenerator > hashGenerator, boost::shared_ptr< std::vector< sauchar_t > > T, boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > > markers, boost::shared_ptr< std::vector< saidx_t > > SA, const std::string &pattern) throw (ConcordiaException)
boost::shared_ptr< ConcordiaSearchResult >	concordiaSearch (boost::shared_ptr< HashGenerator > hashGenerator, boost::shared_ptr< std::vector< sauchar_t > > T, boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > > markers, boost::shared_ptr< std::vector< saidx_t > > SA, const std::string &pattern, bool byWhitespace=false) throw (ConcordiaException)

Detailed Description

Class for searching the index with a sentence. In all searches the sentence is first hashed and then used as a query.

IndexSearcher performs the simpleSearch on its own, but uses a ConcordiaSearcher object to carry out concordiaSearch.

Constructor & Destructor Documentation

IndexSearcher::IndexSearcher ( )

explicit

Constructor.

IndexSearcher::~IndexSearcher ( )

virtual

Destructor.

Member Function Documentation

std::vector< AnubisSearchResult > IndexSearcher::anubisSearch	(	boost::shared_ptr< ConcordiaConfig >	config,
		boost::shared_ptr< HashGenerator >	hashGenerator,
		boost::shared_ptr< std::vector< sauchar_t > >	T,
		boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > >	markers,
		boost::shared_ptr< std::vector< saidx_t > >	SA,
		const std::string &	pattern
	)
throw	(	ConcordiaException
	)

Deprecated:

Finds the examples from the index, whose resemblance to the pattern is maximal. This method may perform very slow, try using concordiaSearch instead.

Parameters

config	concordia config object (to read the anubis threshold parameter)
hashGenerator	hash generator to be used to convert input sentence to a hash
T	hashed index to search in
markers	markers array for the needs of searching
SA	suffix array for the needs of searching
pattern	string pattern to be searched in the index.
byWhitespace	whether to tokenize the pattern by white space

Returns

vector of results

Exceptions

ConcordiaException

boost::shared_ptr< ConcordiaSearchResult > IndexSearcher::concordiaSearch	(	boost::shared_ptr< HashGenerator >	hashGenerator,
		boost::shared_ptr< std::vector< sauchar_t > >	T,
		boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > >	markers,
		boost::shared_ptr< std::vector< saidx_t > >	SA,
		const std::string &	pattern,
		bool	byWhitespace = false
	)
throw	(	ConcordiaException
	)

Performs concordia lookup on the RAM-based index. This is a unique library functionality, designed to facilitate Computer-Aided Translation. For more info see Concordia searching.

Parameters

hashGenerator	hash generator to be used to convert input sentence to a hash
T	hashed index to search in
markers	markers array for the needs of searching
SA	suffix array for the needs of searching
pattern	pattern to be searched in the index.
byWhitespace	whether to tokenize the pattern by white space

Returns

result of the search

Exceptions

ConcordiaException

Here is the call graph for this function:

MatchedPatternFragment IndexSearcher::lexiconSearch	(	boost::shared_ptr< HashGenerator >	hashGenerator,
		boost::shared_ptr< std::vector< sauchar_t > >	T,
		boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > >	markers,
		boost::shared_ptr< std::vector< saidx_t > >	SA,
		const std::string &	pattern,
		bool	byWhitespace = false
	)
throw	(	ConcordiaException
	)

Performs a search useful for lexicons in the following scenario: Concordia gets fed by a lexicon (glossary) instead of a TM. The lexicon search performs as simple search - it requires the match to cover the whole pattern, but additionally the lexicon search requires that the match is the whole example source.

Parameters

hashGenerator	hash generator to be used to convert input sentence to a hash
T	hashed index to search in
markers	markers array for the needs of searching
SA	suffix array for the needs of searching
pattern	string pattern to be searched in the index.

Returns

matched pattern fragment, containing occurences of the pattern in the index

Exceptions

ConcordiaException

Here is the call graph for this function:

MatchedPatternFragment IndexSearcher::simpleSearch	(	boost::shared_ptr< HashGenerator >	hashGenerator,
		boost::shared_ptr< std::vector< sauchar_t > >	T,
		boost::shared_ptr< std::vector< SUFFIX_MARKER_TYPE > >	markers,
		boost::shared_ptr< std::vector< saidx_t > >	SA,
		const std::string &	pattern,
		bool	byWhitespace = false
	)
throw	(	ConcordiaException
	)

Performs a simple substring lookup in RAM-based index. For more info see Simple substring lookup.

Parameters

hashGenerator	hash generator to be used to convert input sentence to a hash
T	hashed index to search in
markers	markers array for the needs of searching
SA	suffix array for the needs of searching
pattern	string pattern to be searched in the index.

Returns

matched pattern fragment, containing occurences of the pattern in the index

Exceptions

ConcordiaException

Here is the call graph for this function:

---

The documentation for this class was generated from the following files:

• /home/rafalj/projects/concordia/concordia/concordia/index_searcher.hpp
• /home/rafalj/projects/concordia/concordia/concordia/index_searcher.cpp