segment

README

Bluge Segment API

These interfaces define the relationship between Bluge and it's segment file formats.

The expectation is that these interfaces can be well-defined, and evolve slowly. By doing so, Bluge and the file formats themselves can evolve more quickly, and yet remain compatible.

Documentation

Index

• Variables
• type Automaton
• type CollectionStats
• type Data
  • func NewDataBytes(b []byte) *Data
  • func NewDataFile(f *os.File) (*Data, error)
  • func (d *Data) Len() int
  • func (d *Data) Read(start, end int) ([]byte, error)
  • func (d *Data) Reader() *DataReader
  • func (d *Data) Size() int
  • func (d *Data) Slice(start, end int) *Data
  • func (d *Data) WriteTo(w io.Writer) (int64, error)
• type DataReader
  • func (r *DataReader) Read(p []byte) (n int, err error)
• type Dictionary
• type DictionaryEntry
• type DictionaryIterator
• type DictionaryLookup
• type DocVisitState
• type Document
• type DocumentValueReader
• type DocumentValueVisitor
• type Field
• type FieldTerm
• type Location
• type Merger
• type Optimizable
• type OptimizableContext
• type OptimizablePostingsIterator
• type Posting
• type PostingsIterator
• type PostingsList
• type Segment
• type StoredFieldVisitor
• type Term
• type TermStats
• type VisitField
• type VisitLocation
• type VisitTerm

Variables

var ErrClosed = fmt.Errorf("index closed")

Types

type Automaton

type Automaton interface {

	// Start returns the start state
	Start() int

	// IsMatch returns true if and only if the state is a match
	IsMatch(int) bool

	// CanMatch returns true if and only if it is possible to reach a match
	// in zero or more steps
	CanMatch(int) bool

	// WillAlwaysMatch returns true if and only if the current state matches
	// and will always match no matter what steps are taken
	WillAlwaysMatch(int) bool

	// Accept returns the next state given the input to the specified state
	Accept(int, byte) int
}

type CollectionStats

type CollectionStats interface {

	// TotalDocumentCount returns the number of documents, regardless of whether or not
	// they have any terms for this field
	TotalDocumentCount() uint64

	// DocumentCount returns the number of documents with at least one term for this field
	DocumentCount() uint64

	// SumTotalTermFrequency returns to total number of tokens across all documents
	SumTotalTermFrequency() uint64

	Merge(CollectionStats)
}

type Data

type Data struct {
	// contains filtered or unexported fields
}

Data is an opaque representation of some data. This data could have been read onto the heap, it could be a live memory-mapped region, or it could be loaded on demand using traditional file I/O.

Micro-benchmarking supported using this concrete structure with simple conditional over an interface with multiple implementations.

func NewDataBytes

func NewDataBytes(b []byte) *Data

func NewDataFile

func NewDataFile(f *os.File) (*Data, error)

func (*Data) Len

func (d *Data) Len() int

func (*Data) Read

func (d *Data) Read(start, end int) ([]byte, error)

func (*Data) Reader

func (d *Data) Reader() *DataReader

func (*Data) Size

func (d *Data) Size() int

func (*Data) Slice

func (d *Data) Slice(start, end int) *Data

func (*Data) WriteTo

func (d *Data) WriteTo(w io.Writer) (int64, error)

type DataReader

type DataReader struct {
	// contains filtered or unexported fields
}

func (*DataReader) Read

func (r *DataReader) Read(p []byte) (n int, err error)

type Dictionary

type Dictionary interface {
	DictionaryLookup

	PostingsList(term []byte, except *roaring.Bitmap, prealloc PostingsList) (PostingsList, error)

	Iterator(a Automaton,
		startKeyInclusive, endKeyExclusive []byte) DictionaryIterator
}

type DictionaryEntry

type DictionaryEntry interface {
	Term() string
	Count() uint64
}

type DictionaryIterator

type DictionaryIterator interface {
	Next() (DictionaryEntry, error)
	Close() error
}

type DictionaryLookup

type DictionaryLookup interface {
	Contains(key []byte) (bool, error)
	Close() error
}

type DocVisitState

type DocVisitState interface {
}

type Document

type Document interface {
	Analyze()
	EachField(vf VisitField)
}

type DocumentValueReader

type DocumentValueReader interface {
	VisitDocumentValues(number uint64, visitor DocumentValueVisitor) error
}

type DocumentValueVisitor

type DocumentValueVisitor func(field string, term []byte)

DocumentValueVisitor is the callback function used by the DocumentValueReader's VisitDocumentValues method.

type Field

type Field interface {
	Name() string
	Length() int
	EachTerm(vt VisitTerm)
	Value() []byte

	Index() bool
	Store() bool
	IndexDocValues() bool
}

type FieldTerm

type FieldTerm interface {
	Term() []byte
	Frequency() int
	EachLocation(vl VisitLocation)
}

type Location

type Location interface {
	Field() string
	Start() int
	End() int
	Pos() int
	Size() int
}

type Merger

type Merger interface {
	WriteTo(w io.Writer, closeCh chan struct{}) (n int64, err error)
	DocumentNumbers() [][]uint64
}

type Optimizable

type Optimizable interface {
	Optimize(kind string, octx OptimizableContext) (OptimizableContext, error)
}

type OptimizableContext

type OptimizableContext interface {
	// Once all the optimzable resources have been provided the same
	// OptimizableContext instance, the optimization preparations are
	// finished or completed via the Finish() method.
	//
	// Depending on the optimization being performed, the Finish()
	// method might return a non-nil Optimized instance.  For example,
	// the Optimized instance might represent an optimized
	// PostingsIterator instance.
	Finish() (PostingsIterator, error)
}

type OptimizablePostingsIterator

type OptimizablePostingsIterator interface {
	ActualBitmap() *roaring.Bitmap
	DocNum1Hit() (uint64, bool)
	ReplaceActual(*roaring.Bitmap)
}

type Posting

type Posting interface {
	Number() uint64
	SetNumber(uint64)
	Frequency() int
	Norm() float64
	Locations() []Location
	Size() int
}

type PostingsIterator

type PostingsIterator interface {
	// The caller is responsible for copying whatever it needs from
	// the returned Posting instance before calling Next(), as some
	// implementations may return a shared instance to reduce memory
	// allocations.
	Next() (Posting, error)

	// Advance will return the posting with the specified doc number
	// or if there is no such posting, the next posting.
	// Callers MUST NOT attempt to pass a docNum that is less than or
	// equal to the currently visited posting doc Num.
	Advance(docNum uint64) (Posting, error)

	Size() int

	// is this postings iterator empty?
	Empty() bool

	Count() uint64

	Close() error
}

type PostingsList

type PostingsList interface {
	Iterator(includeFreq, includeNorm, includeLocations bool, prealloc PostingsIterator) (PostingsIterator, error)

	Size() int

	Count() uint64
}

type Segment

type Segment interface {
	Dictionary(field string) (Dictionary, error)

	VisitStoredFields(num uint64, visitor StoredFieldVisitor) error

	Count() uint64

	DocsMatchingTerms([]Term) (*roaring.Bitmap, error)

	Fields() []string

	CollectionStats(field string) (CollectionStats, error)

	Size() int

	DocumentValueReader(fields []string) (DocumentValueReader, error)

	WriteTo(w io.Writer, closeCh chan struct{}) (int64, error)

	Type() string
	Version() uint32
}

type StoredFieldVisitor

type StoredFieldVisitor func(field string, value []byte) bool

StoredFieldVisitor defines a callback to be visited for each stored field value. The return value determines if the visitor should keep going. Returning true continues visiting, false stops.

type Term

type Term interface {
	Field() string
	Term() []byte
}

type TermStats

type TermStats interface {

	// DocumentFrequency returns the number of documents using this term
	DocumentFrequency() uint64
}

type VisitField

type VisitField func(Field)

type VisitLocation

type VisitLocation func(Location)

type VisitTerm

type VisitTerm func(FieldTerm)