JASS::index_postings Class Reference

Non-thread-safe object that accumulates a single postings list during indexing. More...

#include <index_postings.h>

Collaboration diagram for JASS::index_postings:

Classes
class	posting
	The representation of a single postings as a tuple of docid, term frequency and position. More...

Public Member Functions
	index_postings (allocator &memory_pool)
	Constructor. More...
virtual void	push_back (JASS::compress_integer::integer document_id, index_postings_impact::impact_type amount=1)
	Add to the end of the postings list for this term a term frequency of the given amount. More...
virtual void	push_back (const std::vector< JASS::posting > &data)
	Add a set of postings to the end of the postings list (warning) More...
compress_integer::integer	linearize (uint8_t *temporary, size_t temporary_size, compress_integer::integer *ids, index_postings_impact::impact_type *frequencies, size_t id_and_frequencies_length) const
	Turn the internal format used to accumulate postings into a docid and term-frequencies array. More...
void	impact_order (size_t documents_in_collection, index_postings_impact &postings_list) const
	Return the postings list impact ordered postings list with impact headers. More...
void	impact_order (size_t documents_in_collection, index_postings_impact &postings_list, compress_integer::integer document_frequency, compress_integer::integer *document_ids, index_postings_impact::impact_type *term_frequencies) const
	Return the postings list impact ordered postings list with impact headers. More...
void	text_render (std::ostream &stream) const
	Dump a human-readable version of the postings list down the stream. Format is: <DocID, TF, Pos, Pos, Pos>... More...

Static Public Member Functions
static void	unittest (void)
	Unit test this class.

Private Attributes
compress_integer::integer	highest_document
	The higest document number seen in this postings list (counting from 1)
dynamic_array< uint8_t >	document_ids
	Array holding the docids (variable byte encoded)
dynamic_array< index_postings_impact::impact_type >	term_frequencies
	Array holding the term frequencies (as integers)

Static Private Attributes
static constexpr size_t	initial_size = 16
	Initially allocate space for 4 elements.
static constexpr double	growth_factor = 1.5
	Grow dynamic arrays by a factor of 1.5.

Detailed Description

Non-thread-safe object that accumulates a single postings list during indexing.

Constructor & Destructor Documentation

index_postings()

JASS::index_postings::index_postings ( allocator &  memory_pool )

inline

Constructor.

Parameters

memory_pool

[in] All allocation is from this allocator.

Member Function Documentation

impact_order() [1/2]

void JASS::index_postings::impact_order ( size_t  documents_in_collection, index_postings_impact &  postings_list  ) const

inline

Return the postings list impact ordered postings list with impact headers.

Parameters

documents_in_collection	[in] The number of documents in the collection,
postings_list	[in / out] The constructed impact ordered postings list.

impact_order() [2/2]

void JASS::index_postings::impact_order ( size_t  documents_in_collection, index_postings_impact &  postings_list, compress_integer::integer  document_frequency, compress_integer::integer *  document_ids, index_postings_impact::impact_type *  term_frequencies  ) const

inline

Return the postings list impact ordered postings list with impact headers.

Parameters

documents_in_collection	[in] The number of documents in the collection,
postings_list	[out] The constructed impact ordered postings list.
document_frequency	[in] the document frequency of this term (the length of id_list and tf_list).
document_ids	[in] The list of document ids.
term_frequencies	[in] The list of term frequencies.

linearize()

compress_integer::integer JASS::index_postings::linearize ( uint8_t *  temporary, size_t  temporary_size, compress_integer::integer *  ids, index_postings_impact::impact_type *  frequencies, size_t  id_and_frequencies_length  ) const

inline

Turn the internal format used to accumulate postings into a docid and term-frequencies array.

Parameters

temporary	[in] Buffer used as scratch space.
temporary_size	[in] Length, in bytes, of temporary.
ids	[out] Buffer to store the document ids.
frequencies	[out] Buffer to store the term frequencies.
id_and_frequencies_length	[in] The length of the id and frequencies buffers.

Returns

Returns the document frequency of this term, or 0 on failure.

push_back() [1/2]

virtual void JASS::index_postings::push_back ( JASS::compress_integer::integer  document_id, index_postings_impact::impact_type  amount = 1  )

inlinevirtual

Add to the end of the postings list for this term a term frequency of the given amount.

Parameters

document_id	[in] The document whose term count is to be incremented.
amount	[in] The amount to add to the score. default = 1.

The amount is generally not useful unless pre-computed term frequencies (or impact scores) are known in advance. This might happen if a forward index is being inverted (i.e. term:count values are known).

push_back() [2/2]

virtual void JASS::index_postings::push_back ( const std::vector< JASS::posting > &  data )

inlinevirtual

Add a set of postings to the end of the postings list (warning)

This method assumes the caller is either adding postings to an index one at a time or is adding postings to an index block at a time. WARNING: it with JASS_assert() if there are already postings in the index from the document-at-a-time adding approach.

Parameters

data	[in] The postings list to use. Document ids are assumed to be D1-encoded.

text_render()

void JASS::index_postings::text_render ( std::ostream &  stream ) const

inline

Dump a human-readable version of the postings list down the stream. Format is: <DocID, TF, Pos, Pos, Pos>...

Parameters

stream

[in] The stream to write to.

---

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

• source/index_postings.h