multisearch

README

Amortized O(n) multiple search for Go

Amortized O(n) search with multiple search terms and stopwords

Documentation

Index

• func WordFinder(r rune) bool
• type Engine
  • func NewEngine(stemmer Stemmer, tokenizer Tokenizer) Engine
• type Match
• type Stemmer
• type Token
• type Tokenizer

Functions

func WordFinder

func WordFinder(r rune) bool

WordFinder is an implementation of a Tokenizer designed to break up text on word boundaries.

Types

type Engine

type Engine interface {
	// Ignore registers a new ignored word.
	Ignore(input string) error

	// Match registers and returns a new Match.
	Match(needle string) (Match, error)

	// Process breaks up the input into a slice of Tokens based on Matches
	// and ignores.
	Process(input string) []Token
}

Engine represents a multisearch engine capable of registering new matches, ignores (stopwords) and - most importantly - processing new inputs.

func NewEngine

func NewEngine(stemmer Stemmer, tokenizer Tokenizer) Engine

NewEngine provides a public constructor for the Engine.

type Match

type Match interface {
	// Size provides the total length of the Match.
	Size() int

	// String returns a full string representation of the Match.
	String() string
}

Match represents a matched term.

type Stemmer

type Stemmer interface {
	// StemString stems a string to produce another string. The original
	// is not changed in the process.
	StemString(word string) string
}

Stemmer does the job of stemming words.

type Token

type Token interface {
	// Ignored tells whether this given piece of text has been ignored.
	Ignored() bool

	// Matches returns an array of Matches that were matched against
	// this particular piece of text.
	Matches() []Match

	// String returns the content of the token.
	String() string
}

Token represents a piece of processed text. From user's perspective this is behaves like a single linked list.

type Tokenizer

type Tokenizer func(rune) bool

Tokenizer is a function capable of taking a rune as an input and deciding whether the rune should be captured (true) or ignored (false).