The Collapser class tracks collapse keys and the documents they match. More...

#include <collapser.h>

Collaboration diagram for Collapser:

Public Member Functions
	Collapser (Xapian::valueno slot_, Xapian::doccount collapse_max_, std::vector< Result > &results_, MSetCmp mcmp_)

	operator bool () const
	Return true if collapsing is active for this match. More...

collapse_result	check (Result &result, Xapian::Document::Internal &vsdoc)
	Check a new result. More...

void	process (collapse_result action, Xapian::doccount item)
	Handle a new Result. More...

void	result_has_moved (Xapian::doccount from, Xapian::doccount to)
	Process relocation of entry in results. More...

Xapian::doccount	get_collapse_count (const std::string &collapse_key, int percent_threshold, double min_weight) const

Xapian::doccount	get_docs_considered () const

Xapian::doccount	get_dups_ignored () const

Xapian::doccount	get_entries () const

Xapian::doccount	get_matches_lower_bound () const

void	finalise (double min_weight, int percent_threshold)

Public Attributes
Xapian::doccount	old_item = 0
	Replaced item when REPLACE is returned by collapse(). More...

Private Member Functions
bool	operator() (Xapian::doccount a, Xapian::doccount b) const
	Adapt mcmp to be usable with min_heap. More...

Private Attributes
std::unordered_map< std::string, CollapseData >	table
	Map from collapse key values to the items we're keeping for them. More...

Xapian::doccount	entry_count = 0
	How many items we're currently keeping in table. More...

Xapian::doccount	no_collapse_key = 0
	How many documents have we seen without a collapse key? More...

Xapian::doccount	dups_ignored = 0
	How many documents with duplicate collapse keys we have ignored. More...

Xapian::doccount	docs_considered = 0
	How many documents we've considered for collapsing. More...

Xapian::valueno	slot
	The value slot we're getting collapse keys from. More...

Xapian::doccount	collapse_max
	The maximum number of items to keep for each collapse key value. More...

std::vector< Result > &	results

MSetCmp	mcmp

CollapseData *	ptr = NULL
	Pointer to CollapseData when NEW or ADD is in progress. More...

Detailed Description

The Collapser class tracks collapse keys and the documents they match.

Definition at line 135 of file collapser.h.

Constructor & Destructor Documentation

◆ Collapser()

Collapser::Collapser	(	Xapian::valueno	slot_,
		Xapian::doccount	collapse_max_,
		std::vector< Result > &	results_,
		MSetCmp	mcmp_
	)

inline

Definition at line 184 of file collapser.h.

Member Function Documentation

◆ check()

collapse_result Collapser::check	(	Result &	result,
		Xapian::Document::Internal &	vsdoc
	)

Check a new result.

If this method determines the action to take is NEW or ADD then the proto-mset should be updated and then process() called to complete the update (if the result doesn't actually get added, then it's OK not to follow up with a call to process()).

Parameters

result	The new result.
vsdoc	Document for getting values.

Returns: How to handle result: EMPTY, NEW, ADD, REJECT or REPLACE.

Definition at line 126 of file collapser.cc.

References ADD, CollapseData::check_item(), EMPTY, Result::get_collapse_key(), Result::get_docid(), Xapian::Document::Internal::get_value(), NEW, REJECT, REPLACE, and Result::set_collapse_key().

Referenced by ProtoMSet::process().

◆ finalise()

void Collapser::finalise	(	double	min_weight,
		int	percent_threshold
	)

Definition at line 237 of file collapser.cc.

Referenced by ProtoMSet::finalise().

◆ get_collapse_count()

Xapian::doccount Collapser::get_collapse_count	(	const std::string &	collapse_key,
		int	percent_threshold,
		double	min_weight
	)		const

Definition at line 185 of file collapser.cc.

References Assert.

◆ get_docs_considered()

Xapian::doccount Collapser::get_docs_considered ( ) const

inline

Definition at line 243 of file collapser.h.

References docs_considered.

Referenced by ProtoMSet::finalise().

◆ get_dups_ignored()

Xapian::doccount Collapser::get_dups_ignored ( ) const

inline

Definition at line 245 of file collapser.h.

References dups_ignored.

Referenced by ProtoMSet::finalise().

◆ get_entries()

Xapian::doccount Collapser::get_entries ( ) const

inline

Definition at line 247 of file collapser.h.

References entry_count.

◆ get_matches_lower_bound()

Xapian::doccount Collapser::get_matches_lower_bound ( ) const

Definition at line 212 of file collapser.cc.

Referenced by ProtoMSet::finalise().

◆ operator bool()

Collapser::operator bool ( ) const

inline

Return true if collapsing is active for this match.

Definition at line 194 of file collapser.h.

References collapse_max.

◆ operator()()

bool Collapser::operator()	(	Xapian::doccount	a,
		Xapian::doccount	b
	)		const

inlineprivate

Adapt mcmp to be usable with min_heap.

Definition at line 176 of file collapser.h.

References mcmp, and results.

◆ process()

void Collapser::process	(	collapse_result	action,
		Xapian::doccount	item
	)

Handle a new Result.

Parameters

action	The collapse_result returned by check().
item	The new item (index into results).

Definition at line 164 of file collapser.cc.

References ADD, Assert, and NEW.

Referenced by ProtoMSet::process().

◆ result_has_moved()

void Collapser::result_has_moved	(	Xapian::doccount	from,
		Xapian::doccount	to
	)

inline

Process relocation of entry in results.

Parameters

from	The old item (index into results).
to	The new item (index into results).

Definition at line 223 of file collapser.h.

References Assert, rare, CollapseData::result_has_moved(), results, and table.

Referenced by ProtoMSet::handle_min_weight_pending().

Member Data Documentation

◆ collapse_max

Xapian::doccount Collapser::collapse_max

private

The maximum number of items to keep for each collapse key value.

Definition at line 166 of file collapser.h.

Referenced by operator bool().

◆ docs_considered

Xapian::doccount Collapser::docs_considered = 0

private

How many documents we've considered for collapsing.

We use this statistic to improve matches_estimated (by considering the rate of collapsing).

Definition at line 160 of file collapser.h.

Referenced by get_docs_considered().

◆ dups_ignored

Xapian::doccount Collapser::dups_ignored = 0

private

How many documents with duplicate collapse keys we have ignored.

We use this statistic to improve matches_estimated (by considering the rate of collapsing) and matches_upper_bound.

Definition at line 153 of file collapser.h.

Referenced by get_dups_ignored().

◆ entry_count

Xapian::doccount Collapser::entry_count = 0

private

How many items we're currently keeping in table.

Definition at line 140 of file collapser.h.

Referenced by get_entries().

◆ mcmp

MSetCmp Collapser::mcmp

private

Definition at line 170 of file collapser.h.

Referenced by operator()().

◆ no_collapse_key

Xapian::doccount Collapser::no_collapse_key = 0

private

How many documents have we seen without a collapse key?

We use this statistic to improve matches_lower_bound.

Definition at line 146 of file collapser.h.

◆ old_item

Xapian::doccount Collapser::old_item = 0

Replaced item when REPLACE is returned by collapse().

Definition at line 182 of file collapser.h.

Referenced by ProtoMSet::process().

◆ ptr

CollapseData* Collapser::ptr = NULL

private

Pointer to CollapseData when NEW or ADD is in progress.

Definition at line 173 of file collapser.h.

◆ results

std::vector<Result>& Collapser::results

private

Definition at line 168 of file collapser.h.

Referenced by operator()(), and result_has_moved().

◆ slot

Xapian::valueno Collapser::slot

private

The value slot we're getting collapse keys from.

Definition at line 163 of file collapser.h.

◆ table

std::unordered_map<std::string, CollapseData> Collapser::table

private

Map from collapse key values to the items we're keeping for them.

Definition at line 137 of file collapser.h.

Referenced by result_has_moved().

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

matcher/collapser.h
matcher/collapser.cc

Public Member Functions

Public Attributes

Private Member Functions

Private Attributes

Detailed Description

Constructor & Destructor Documentation

◆ Collapser()

Member Function Documentation

◆ check()

◆ finalise()

◆ get_collapse_count()

◆ get_docs_considered()

◆ get_dups_ignored()

◆ get_entries()

◆ get_matches_lower_bound()

◆ operator bool()

◆ operator()()

◆ process()

◆ result_has_moved()

Member Data Documentation

◆ collapse_max

◆ docs_considered

◆ dups_ignored

◆ entry_count

◆ mcmp

◆ no_collapse_key

◆ old_item

◆ ptr

◆ results

◆ slot

◆ table