zishuo/go-utils
1
0
Fork 0
mirror of https://github.com/unti-io/go-utils.git synced 2025-12-24 12:57:50 +08:00
Code Issues Packages Projects Releases Wiki Activity

v1.8.4

Browse Source
This commit is contained in:
陈兔子
2025-12-11 16:36:47 +08:00
parent a3191ec2d1
commit 7ac18f15f2
1 changed files with 32 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
utils/parse.go
View File

@@ -254,29 +254,8 @@ func (this *ParseClass) HtmlToTextRetainImg(html string, length int, isLine bool
block := regexp.MustCompile(`(?i)</?(p|div|h[1-6]|li|ul|ol|blockquote|address|article|section|header|footer|nav|pre|table|tr|td|th|hr)[^>]*>`)
content = block.ReplaceAllString(content, "\n")
imgRegex := regexp.MustCompile(`(?i)<img\b([^>]*)>`)
content = imgRegex.ReplaceAllStringFunc(content, func(imgTag string) string {
// 查找 class 属性的值（不使用反向引用）
classRegex := regexp.MustCompile(`(?i)\bclass\s*=\s*["']([^"']*)["']`)
if match := classRegex.FindStringSubmatch(imgTag); match != nil {
classes := match[1]
// 检查是否已有 emoji 类
if !regexp.MustCompile(`\bemoji\b`).MatchString(classes) {
newClasses := strings.TrimSpace(classes) + " emoji"
// 用统一的双引号替换整个 class 属性
return classRegex.ReplaceAllString(imgTag, `class="`+newClasses+`"`)
}
return imgTag
} else {
// 没有 class 属性，添加 class="emoji"
endRegex := regexp.MustCompile(`(\s*/?>)`)
if endRegex.MatchString(imgTag) {
return endRegex.ReplaceAllString(imgTag, ` class="emoji"$1`)
}
}
return imgTag
})
// 处理 img 标签，添加或追加 emoji 类
content = this.ProcessImgTags(content)
// 去掉除了 img 之外的所有标签（使用回调判断保留 img）
tag := regexp.MustCompile(`(?i)<[^>]+>`)
@@ -323,4 +302,34 @@ func (this *ParseClass) HtmlToTextRetainImg(html string, length int, isLine bool
// 超出长度则截取前 n 个字符并追加省略号
return string(runes[:length]) + "..."
}
// ProcessImgTags - 处理 img 标签，添加或追加 emoji 类
func (this *ParseClass) ProcessImgTags(content string) string {
imgRegex := regexp.MustCompile(`(?i)<img\b([^>]*)>`)
content = imgRegex.ReplaceAllStringFunc(content, func(imgTag string) string {
// 查找 class 属性的值（不使用反向引用）
classRegex := regexp.MustCompile(`(?i)\bclass\s*=\s*["']([^"']*)["']`)
if match := classRegex.FindStringSubmatch(imgTag); match != nil {
classes := match[1]
// 检查是否已有 emoji 类
if !regexp.MustCompile(`\bemoji\b`).MatchString(classes) {
newClasses := strings.TrimSpace(classes) + " emoji"
// 用统一的双引号替换整个 class 属性
return classRegex.ReplaceAllString(imgTag, `class="`+newClasses+`"`)
}
return imgTag
} else {
// 没有 class 属性，添加 class="emoji"
endRegex := regexp.MustCompile(`(\s*/?>)`)
if endRegex.MatchString(imgTag) {
return endRegex.ReplaceAllString(imgTag, ` class="emoji"$1`)
}
}
return imgTag
})
return content
}