Refactor: Improve space insertion logic for Pinyin conversion

phrase/paragraph.go

@@ -17,26 +17,27 @@ import (
 
 var (
 	spacesReg = regexp.MustCompile(`[\s]+`)
-	allowReg  = regexp.MustCompile(`[a-zA-Z0-9\.,\?\!;\(\)\[\]\&\=\-_@\s]`)
 
 	// Option set pinyin style args option
 	Option = gpy.Args{
 		Style:     gpy.Normal,
 		Heteronym: true,
 	}
 
-	hanSymbols = map[string]string{
-		"？": "?",
-		"！": "!",
-		"：": ":",
-		"。": ".",
-		"，": ",",
-		"；": ";",
-		"（": "(",
-		"）": ")",
-		"【": "[",
-		"】": "]",
-		"、": ",",
+	hanSymbols = map[rune]rune{
+		'？': '?',
+		'！': '!',
+		'：': ':',
+		'。': '.',
+		'，': ',',
+		'；': ';',
+		'（': '(',
+		'）': ')',
+		'【': '[',
+		'】': ']',
+		'、': ',',
+		'“': '"',
+		'”': '"',
 	}
 )
 
@@ -66,52 +67,35 @@ func Join(a []string) (s string) {
 // including letters, numbers, symbols
 func Paragraph(p string, segs ...gse.Segmenter) (s string) {
 	p = pinyinPhrase(p, segs...)
-
+	var b strings.Builder
+	var last rune
 	for _, r := range p {
 		if unicode.Is(unicode.Han, r) {
 			// Han chars
+			if last != 0 && !isPunctOrSymbol(last) {
+				b.WriteRune(' ')
+			}
 			result := gpy.HanPinyin(string(r), Option)
 			if len(result) == 0 {
 				continue
 			}
-
 			if len(result[0]) == 0 {
 				continue
 			}
-
-			s += " " + string(result[0][0]) + " "
+			b.WriteString(result[0][0])
+		} else if symbol, ok := hanSymbols[r]; ok {
+			// Han symbols
+			b.WriteRune(symbol)
 		} else {
-			// Not han chars
-			char := string(r)
-
-			if allowReg.MatchString(char) {
-				s += char
-			} else {
-				if hanSymbols[char] != "" {
-					s += hanSymbols[char]
-				}
-			}
+			// Ohter
+			b.WriteRune(r)
 		}
+		last = r
 	}
+	s = b.String()
 
 	// trim the two continuous spaces
 	s = spacesReg.ReplaceAllString(s, " ")
-	m := map[string]string{
-		" ,": ",",
-		" .": ".",
-		" ?": "?",
-		" !": "!",
-		" ;": ";",
-		" :": ":",
-		" (": "(",
-		") ": ")",
-		"[ ": "[",
-		" ]": "]",
-	}
-
-	for k, v := range m {
-		s = strings.Replace(s, k, v, -1)
-	}
 
 	s = strings.TrimSpace(s)
 	return

phrase/paragraph_test.go

@@ -31,6 +31,15 @@ func TestParagraph(t *testing.T) {
 		"香港维多利亚港":         "xiang gang wei duo li ya gang",
 		"上海外滩, 陆家嘴上海中心大厦": "shang hai wai tan, lu jia zui shang hai zhong xin da sha",
 		"北京八达岭长城":         "bei jing ba da ling chang cheng",
+
+		"你好Golang":         "ni hao Golang",
+		"2006年":            "2006 nian",
+		"价格$100":           "jia ge$100",
+		"邮箱：[email protected]": "you xiang:[email protected]",
+		"你好！":              "ni hao!",
+		"“你好”":             "\"ni hao\"",
+		"中文#标签":            "zhong wen#biao qian",
+		"阅读《Astérix》":      "yue du《Astérix》",
 	}
 
 	seg, err := gse.New("zh, ../examples/dict.txt")

phrase/phrase_cut.go

@@ -8,6 +8,7 @@ package phrase
 
 import (
 	"strings"
+	"unicode"
 
 	"github.com/go-ego/gpy"
 	"github.com/go-ego/gse"
@@ -62,10 +63,21 @@ func Match(word string) string {
 	return match
 }
 
+func isPunctOrSymbol(r rune) bool {
+	return unicode.IsOneOf([]*unicode.RangeTable{unicode.Punct, unicode.Symbol}, r)
+}
+
 func matchs(s, word string) string {
-	match := Match(word)
-	if match != "" {
-		s = strings.Replace(s, word, " "+match+" ", 1)
+	if match := Match(word); match != "" {
+		if before, after, found := strings.Cut(s, word); found {
+			if before := []rune(before); len(before) > 0 && !isPunctOrSymbol(before[len(before)-1]) {
+				match = " " + match
+			}
+			if after := []rune(after); len(after) > 0 && !isPunctOrSymbol(after[0]) {
+				match += " "
+			}
+			s = strings.Replace(s, word, match, 1)
+		}
 	}
 
 	return s

phrase/phrase_test.go

@@ -18,12 +18,12 @@ func Test_pinyinPhrase(t *testing.T) {
 	expects := map[string]string{
 		"西雅图太空针, The Space Nedle":            "西雅图 tai kong 针, The Space Nedle",
 		"旧金山湾金门大桥":                           "旧金山湾金门 da qiao",
-		"纽约帝国大厦, 纽约时代广场, 世贸中心":               "纽约帝国 da sha , 纽约时代 guang chang , 世贸中心",
+		"纽约帝国大厦, 纽约时代广场, 世贸中心":               "纽约帝国 da sha, 纽约时代 guang chang, 世贸中心",
 		"多伦多加拿大国家电视塔, the CN Tower, 尼亚加拉大瀑布": "多伦多加拿 da guo jia dian 视塔, the CN Tower, 尼亚加拉大 pu bu",
 		"伦敦泰晤士河, 大笨钟":                        "lun dun 泰晤士河, 大笨钟",
 
 		"洛杉矶好莱坞":          "洛杉矶 hao lai wu",
-		"东京都, 东京晴空塔":      "东 jing du , 东京 qing kong 塔",
+		"东京都, 东京晴空塔":      "东 jing du, 东京 qing kong 塔",
 		"巴黎埃菲尔铁塔":         "巴黎 ai fei er tie ta",
 		"香港维多利亚港":         "xiang gang 维多利亚港",
 		"上海外滩, 陆家嘴上海中心大厦": "shang hai 外滩, 陆家嘴 shang hai 中心 da sha",