From 0124ebadce1394f3736de8ed331fc9c51d127dea Mon Sep 17 00:00:00 2001
From: Wang Bin <wangbin.zibo@gmail.com>
Date: Wed, 29 Apr 2015 18:51:38 +0800
Subject: [PATCH] put dictionary to a seperated module

---
 dictionary/dictionary.go | 121 +++++++++++++++++++++++++++++++++++++++
 dictionary/token.go      |  19 ++++++
 2 files changed, 140 insertions(+)
 create mode 100644 dictionary/dictionary.go
 create mode 100644 dictionary/token.go

diff --git a/dictionary/dictionary.go b/dictionary/dictionary.go
new file mode 100644
index 0000000..7fa7625
--- /dev/null
+++ b/dictionary/dictionary.go
@@ -0,0 +1,121 @@
+package dictionary
+
+import (
+	"bufio"
+	"math"
+	"os"
+	"path/filepath"
+	"strconv"
+	"strings"
+	"sync"
+)
+
+type Dictionary struct {
+	total, logTotal float64
+	freqMap         map[string]float64
+	sync.RWMutex
+}
+
+func (d *Dictionary) addToken(token Token) {
+	d.freqMap[token.text] = token.frequency
+	d.total += token.frequency
+	runes := []rune(token.text)
+	n := len(runes)
+	for i := 0; i < n; i++ {
+		frag := string(runes[:i+1])
+		if _, ok := d.freqMap[frag]; !ok {
+			d.freqMap[frag] = 0.0
+		}
+	}
+}
+
+func (d *Dictionary) updateLogTotal() {
+	d.logTotal = math.Log(d.total)
+}
+
+func (d *Dictionary) AddToken(token Token) {
+	d.Lock()
+	d.addToken(token)
+	d.updateLogTotal()
+	d.Unlock()
+}
+
+func (d Dictionary) Total() float64 {
+	return d.total
+}
+
+func (d Dictionary) LogTotal() float64 {
+	return d.logTotal
+}
+
+func (d Dictionary) Frequency(key string) (float64, bool) {
+	d.RLock()
+	freq, ok := d.freqMap[key]
+	d.RUnlock()
+	return freq, ok
+}
+
+func (d *Dictionary) LoadDictionary(fileName string) error {
+	return d.loadDictionary(fileName, false)
+}
+
+func (d *Dictionary) LoadUserDictionary(fileName string) error {
+	return d.loadDictionary(fileName, true)
+}
+
+func (d *Dictionary) loadDictionary(fileName string, isUserDictionary bool) error {
+	filePath, err := dictPath(fileName)
+	if err != nil {
+		return err
+	}
+	dictFile, err := os.Open(filePath)
+	if err != nil {
+		return err
+	}
+	defer dictFile.Close()
+
+	scanner := bufio.NewScanner(dictFile)
+	var token Token
+	var line string
+	var fields []string
+
+	d.Lock()
+	defer d.Unlock()
+
+	if !isUserDictionary && len(d.freqMap) > 0 {
+		d.freqMap = make(map[string]float64)
+	}
+	for scanner.Scan() {
+		line = scanner.Text()
+		fields = strings.Split(line, " ")
+		token.text = strings.Replace(fields[0], "\ufeff", "", 1)
+		if length := len(fields); length > 1 {
+			token.frequency, err = strconv.ParseFloat(fields[1], 64)
+			if err != nil {
+				return err
+			}
+			if length > 2 {
+				token.pos = fields[2]
+			}
+		}
+		d.addToken(token)
+	}
+	d.updateLogTotal()
+	if err = scanner.Err(); err != nil {
+		return err
+	}
+	return nil
+}
+
+func dictPath(dictFileName string) (string, error) {
+	if filepath.IsAbs(dictFileName) {
+		return dictFileName, nil
+	}
+	var dictFilePath string
+	cwd, err := os.Getwd()
+	if err != nil {
+		return dictFilePath, err
+	}
+	dictFilePath = filepath.Clean(filepath.Join(cwd, dictFileName))
+	return dictFilePath, nil
+}
diff --git a/dictionary/token.go b/dictionary/token.go
new file mode 100644
index 0000000..f4124a1
--- /dev/null
+++ b/dictionary/token.go
@@ -0,0 +1,19 @@
+package dictionary
+
+type Token struct {
+	text      string
+	frequency float64
+	pos       string
+}
+
+func (t Token) Text() string {
+	return t.text
+}
+
+func (t Token) Frequency() float64 {
+	return t.frequency
+}
+
+func (t Token) Pos() string {
+	return t.pos
+}