meta::index::ranker Class Reference

A ranker scores a query against all the documents in an inverted index, returning a list of documents sorted by relevance. More...

#include <ranker.h>

Inheritance diagram for meta::index::ranker:

Public Types
using	filter_function_type = std::function< bool(doc_id did)>

Public Member Functions
template<class ForwardIterator , class Function = bool (*)(doc_id)>
std::vector< search_result >	score (inverted_index &idx, ForwardIterator begin, ForwardIterator end, uint64_t num_results=10, Function &&filter=passthrough)
std::vector< search_result >	score (inverted_index &idx, const corpus::document &query, uint64_t num_results=10, const filter_function_type &filter=[](doc_id) { return true;})
virtual	~ranker ()=default
	Default destructor.
virtual void	save (std::ostream &out) const =0
	Saves the ranker to a stream. More...
virtual std::vector< search_result >	rank (ranker_context &ctx, uint64_t num_results, const filter_function_type &filter)=0
	Scores a query using a document-at-a-time strategy. More...

Static Public Member Functions
static bool	passthrough (doc_id)

Detailed Description

A ranker scores a query against all the documents in an inverted index, returning a list of documents sorted by relevance.

Member Function Documentation

score() [1/2]

template<class ForwardIterator , class Function = bool (*)(doc_id)>

std::vector<search_result> meta::index::ranker::score ( inverted_index &  idx, ForwardIterator  begin, ForwardIterator  end, uint64_t  num_results = 10, Function &&  filter = passthrough  )

inline

Parameters

idx	The index this ranker is operating on
begin	A forward iterator to the beginning of the term weights (pairs of std::string and a weight)
end	A forward iterator to the end of the above range
num_results	The number of results to return in the vector
filter	A filtering function to apply to each doc_id; returns true if the document should be included in results

score() [2/2]

std::vector< search_result > meta::index::ranker::score	(	inverted_index &	idx,
		const corpus::document &	query,
		uint64_t	num_results = 10,
		const filter_function_type &	filter = [](doc_id) { return true; }
	)

Parameters

idx	The index this ranker is operating on
query	The current query
num_results	The number of results to return in the vector
filter	A filtering function to apply to each doc_id; returns true if the document should be included in results

save()

virtual void meta::index::ranker::save ( std::ostream &  out ) const

pure virtual

Saves the ranker to a stream.

This should save the ranker's id, followed by any parameters needed for reconstruction.

Implemented in meta::index::rocchio, meta::index::kl_divergence_prf, meta::index::okapi_bm25, meta::index::pivoted_length, meta::index::jelinek_mercer, meta::index::absolute_discount, and meta::index::dirichlet_prior.

rank()

virtual std::vector<search_result> meta::index::ranker::rank ( ranker_context &  ctx, uint64_t  num_results, const filter_function_type &  filter  )

pure virtual

Scores a query using a document-at-a-time strategy.

You should not override this unless you desire a completely different ranking strategy than document-at-a-time, which might be the case if you are implementing a new pseudo-relevance feedback method.

Parameters

ctx	The ranker_context holding the postings lists
num_results	The number of search results to return
filter	The filter function to be used

Implemented in meta::index::ranking_function, meta::index::rocchio, and meta::index::kl_divergence_prf.

---

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

• /home/chase/projects/meta/include/meta/index/ranker/ranker.h
• /home/chase/projects/meta/src/index/ranker/ranker.cpp