110 lines
3.4 KiB
Go
110 lines
3.4 KiB
Go
package main
|
|
|
|
import (
|
|
"bytes"
|
|
"regexp"
|
|
"strings"
|
|
"unicode"
|
|
|
|
"git.pcux.cn/xuans/Smail/dto/parsemail"
|
|
"golang.org/x/net/html"
|
|
)
|
|
|
|
const maxWeChatVerificationScanBytes = 64 << 10
|
|
|
|
const weChatVerificationKeyword = `(?:验证码|校验码|动态码|认证码|一次性(?:密码|口令)|verification[\s-]*code|security[\s-]*code|authentication[\s-]*code|one[\s-]*time[\s-]*(?:password|code)|passcode|\botp\b)`
|
|
|
|
var weChatVerificationPatterns = []*regexp.Regexp{
|
|
regexp.MustCompile(`(?i)` + weChatVerificationKeyword + `(?:[[:space:][:punct:]：，。！【】（）]|是|为|is|equals|如下|your|please|use){0,20}\b([A-Z0-9]{4,8})\b`),
|
|
regexp.MustCompile(`(?i)\b([A-Z0-9]{4,8})\b(?:[[:space:][:punct:]：，。！【】（）]|是|为|is|your|您的){0,20}` + weChatVerificationKeyword),
|
|
}
|
|
|
|
func extractWeChatVerificationCode(email *parsemail.Email) (string, bool) {
|
|
if email == nil {
|
|
return "", false
|
|
}
|
|
// 分别检查标题和各正文，避免跨字段拼接出不存在的验证码。
|
|
if code, ok := findWeChatVerificationCode(boundedWeChatVerificationText(email.Subject)); ok {
|
|
return code, true
|
|
}
|
|
text := email.Text[:min(len(email.Text), maxWeChatVerificationScanBytes)]
|
|
plain := string(text)
|
|
if len(email.Text) > len(text) {
|
|
plain += "_"
|
|
}
|
|
if code, ok := findWeChatVerificationCode(plain); ok {
|
|
return code, true
|
|
}
|
|
return findWeChatVerificationCode(weChatHTMLNotificationText(email.HTML))
|
|
}
|
|
|
|
func findWeChatVerificationCode(content string) (string, bool) {
|
|
content = strings.Map(func(char rune) rune {
|
|
if unicode.IsSpace(char) {
|
|
return ' '
|
|
}
|
|
return char
|
|
}, content)
|
|
for _, pattern := range weChatVerificationPatterns {
|
|
for _, match := range pattern.FindAllStringSubmatch(content, -1) {
|
|
code := match[1]
|
|
// 排除普通英文单词，只识别至少含一个数字的四至八位字母数字组合。
|
|
if strings.ContainsAny(code, "0123456789") {
|
|
return code, true
|
|
}
|
|
}
|
|
}
|
|
return "", false
|
|
}
|
|
|
|
func boundedWeChatVerificationText(content string) string {
|
|
if len(content) > maxWeChatVerificationScanBytes {
|
|
// 用标识符边界阻止截断的长串被误识别为完整验证码，不进入通知正文。
|
|
return truncateWeChatText(content, maxWeChatVerificationScanBytes) + "_"
|
|
}
|
|
return content
|
|
}
|
|
|
|
func weChatHTMLNotificationText(content []byte) string {
|
|
// 限制原始输入，避免大邮件在通知提取阶段产生无界扫描和额外分配。
|
|
truncated := len(content) > maxWeChatVerificationScanBytes
|
|
content = content[:min(len(content), maxWeChatVerificationScanBytes)]
|
|
tokenizer := html.NewTokenizer(bytes.NewReader(content))
|
|
var text strings.Builder
|
|
hiddenDepth := 0
|
|
for {
|
|
switch tokenizer.Next() {
|
|
case html.ErrorToken:
|
|
if truncated {
|
|
text.WriteByte('_')
|
|
}
|
|
return text.String()
|
|
case html.StartTagToken, html.SelfClosingTagToken:
|
|
name, _ := tokenizer.TagName()
|
|
if weChatHiddenHTMLTag(string(name)) {
|
|
hiddenDepth++
|
|
}
|
|
case html.EndTagToken:
|
|
name, _ := tokenizer.TagName()
|
|
if hiddenDepth > 0 && weChatHiddenHTMLTag(string(name)) {
|
|
hiddenDepth--
|
|
}
|
|
case html.TextToken:
|
|
if hiddenDepth == 0 {
|
|
// 分词器负责解码实体；不读取属性、注释、脚本及样式中的内容。
|
|
text.WriteByte(' ')
|
|
text.Write(tokenizer.Text())
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func weChatHiddenHTMLTag(name string) bool {
|
|
switch name {
|
|
case "head", "script", "style", "template":
|
|
return true
|
|
default:
|
|
return false
|
|
}
|
|
}
|