Implement function to prune string keeping HTML closing tags

backend/app/notify/prune_html.go

@@ -0,0 +1,120 @@
+package notify
+
+import (
+	"fmt"
+	"strings"
+
+	"golang.org/x/net/html"
+)
+
+type stringArr struct {
+	data []string
+	len  int
+}
+
+// Push adds element to the end
+func (s *stringArr) Push(v string) {
+	s.data = append(s.data, v)
+	s.len += len(v)
+}
+
+// Pop removes element from end and returns it
+func (s *stringArr) Pop() string {
+	l := len(s.data)
+	newData, v := s.data[:l-1], s.data[l-1]
+	s.data = newData
+	s.len -= len(v)
+	return v
+}
+
+// Unshift adds element to the start
+func (s *stringArr) Unshift(v string) {
+	s.data = append([]string{v}, s.data...)
+	s.len += len(v)
+}
+
+// Shift removes element from start and returns it
+func (s *stringArr) Shift() string {
+	v, newData := s.data[0], s.data[1:]
+	s.data = newData
+	s.len -= len(v)
+	return v
+}
+
+// String returns all strings concatenated
+func (s stringArr) String() string {
+	return strings.Join(s.data, "")
+}
+
+// Len returns total length of all strings concatenated
+func (s stringArr) Len() int {
+	return s.len
+}
+
+// pruneHTML prunes string keeping HTML closing tags
+func pruneHTML(htmlText string, maxLength int) string {
+	result := stringArr{}
+	endTokens := stringArr{}
+
+	suffix := "..."
+	suffixLen := len(suffix)
+
+	tokenizer := html.NewTokenizer(strings.NewReader(htmlText))
+	for {
+		if tokenizer.Next() == html.ErrorToken {
+			return result.String()
+		}
+		token := tokenizer.Token()
+
+		switch token.Type {
+		case html.CommentToken, html.DoctypeToken:
+			// skip tokens without content
+			continue
+
+		case html.StartTagToken:
+			// <token></token>
+			// len(token) * 2 + len("<></>")
+			totalLenToAppend := len(token.Data)*2 + 5
+
+			lengthAfterChange := result.Len() + totalLenToAppend + endTokens.Len() + suffixLen
+
+			if lengthAfterChange > maxLength {
+				return result.String() + suffix + endTokens.String()
+			}
+
+			endTokens.Unshift(fmt.Sprintf("</%s>", token.Data))
+
+		case html.EndTagToken:
+			endTokens.Shift()
+
+		case html.TextToken, html.SelfClosingTagToken:
+			lengthAfterChange := result.Len() + len(token.String()) + endTokens.Len() + suffixLen
+
+			if lengthAfterChange > maxLength {
+				text := pruneStringToWord(token.String(), maxLength-result.Len()-endTokens.Len()-suffixLen)
+				return result.String() + text + suffix + endTokens.String()
+			}
+		}
+
+		result.Push((token.String()))
+	}
+}
+
+// pruneStringToWord prunes string to specified length respecting words
+func pruneStringToWord(text string, maxLength int) string {
+	if maxLength <= 0 {
+		return ""
+	}
+
+	result := ""
+
+	arr := strings.Split(text, " ")
+	for _, s := range arr {
+		if len(result)+len(s) > maxLength {
+			return strings.TrimRight(result, " ")
+		}
+		result += s + " "
+	}
+
+	return text
+}