README.md

Darts

This is a GO implementation of Double-ARray Trie System. It's a clone of the C++ version

Darts can be used as simple hash dictionary. You can also do very fast Common Prefix Search which is essential for morphological analysis, such as word split for CJK text indexing/searching.

Reference

What is Trie
An Implementation of Double-Array Trie

NEWS

• Support building Double-Array from DAWG有向无环图, reduce the on-disk dict half as Trie. Lookup performance increases 25%.

TO DO list

• Documentation/comments
• Benchmark

Switch from unicode to byte version

gofmt -tabs=false -tabwidth=4 -r='rune /*Key_type*/ -> byte /*Key_type*/' -w darts.go
gofmt -tabs=false -tabwidth=4 -r='rune /*Key_type*/ -> byte /*Key_type*/' -w dawg.go

Usage

Input Dictionary Format

Key\tFreq

Each key occupies one line. The file should be utf-8 encoded

Code example (unicode version)

package main

import (
    "darts"
    "fmt"
)

func main() {
    d, err:= darts.Import("darts.txt", "darts.lib", true)
    if err == nil {
        if d.ExactMatchSearch([]rune("考察队员", 0)) {
            fmt.Println("考察队员 is in dictionary")
        }
    }
}

Code example (byte version)

package main

import (
    "darts"
    "fmt"
)

func main() {
    d, err := darts.Import("darts.txt", "darts.lib", true)
    if err == nil {
        key := []byte("考察队员")
        r := d.CommonPrefixSearch(key, 0)
        for i := 0; i < len(r); i++ {
            fmt.Println(string(key[:r[i].PrefixLen]))
        }
    }
}

Performance

Using a 100K item dictionary, a simple search on eath key takes go map 46 ms, takes byte_key version of darts 14 ms, and for unicode_key version of darts 9.5 ms.

LICENSE

Apache License 2.0