0ce86645b1
- Added a new webhook integration for Feishu/Lark groups, allowing for custom bot notifications. - Refactored deployment scripts to unify the deployment process, improving maintainability and reducing redundancy. - Updated the experience index to include recent entries and improved documentation for various roles. Made-with: Cursor
353 lines
10 KiB
Go
353 lines
10 KiB
Go
package handler
|
||
|
||
import (
|
||
"regexp"
|
||
"strings"
|
||
"unicode"
|
||
"unicode/utf8"
|
||
|
||
"soul-api/internal/database"
|
||
"soul-api/internal/model"
|
||
|
||
"gorm.io/gorm"
|
||
)
|
||
|
||
// linkTagPlainTextRe 正文里 #标签 的捕获：仅标签名，遇句读/引号/空格/#/@/</> 等即停止（勿用 [^\s]+ 以免吞入」。）
|
||
var linkTagPlainTextRe = regexp.MustCompile(`#([\p{L}\p{N}\p{Han}\p{Hiragana}\p{Katakana}\p{Hangul}\p{M}_-]+)`)
|
||
|
||
// sanitizeNameOrLabel 去除括号及内容，如 南风(管理) -> 南风
|
||
func sanitizeNameOrLabel(s string) string {
|
||
re := regexp.MustCompile(`\s*[（(][^)）]*(\)|）)?`)
|
||
return strings.TrimSpace(re.ReplaceAllString(s, ""))
|
||
}
|
||
|
||
// isLinkTagBodyRune 标签名允许的字符（用于首尾剥掉句读、引号等标点，避免 #碎片时间」。 被整段吞入）
|
||
func isLinkTagBodyRune(r rune) bool {
|
||
if r == '_' || r == '-' {
|
||
return true
|
||
}
|
||
if unicode.IsLetter(r) || unicode.IsNumber(r) {
|
||
return true
|
||
}
|
||
if unicode.Is(unicode.Han, r) {
|
||
return true
|
||
}
|
||
if unicode.Is(unicode.Hiragana, r) || unicode.Is(unicode.Katakana, r) || unicode.Is(unicode.Hangul, r) {
|
||
return true
|
||
}
|
||
if unicode.Is(unicode.Mn, r) || unicode.Is(unicode.Mc, r) || unicode.Is(unicode.Me, r) {
|
||
return true
|
||
}
|
||
return false
|
||
}
|
||
|
||
func trimLinkTagEdgePunct(s string) string {
|
||
for len(s) > 0 {
|
||
r, size := utf8.DecodeRuneInString(s)
|
||
if isLinkTagBodyRune(r) {
|
||
break
|
||
}
|
||
s = s[size:]
|
||
}
|
||
for len(s) > 0 {
|
||
r, size := utf8.DecodeLastRuneInString(s)
|
||
if isLinkTagBodyRune(r) {
|
||
break
|
||
}
|
||
s = s[:len(s)-size]
|
||
}
|
||
return s
|
||
}
|
||
|
||
// sanitizeLinkTagLabel 链接标签展示名：去括号说明、去首尾标点与杂符号（与 # 解析规则一致）
|
||
func sanitizeLinkTagLabel(s string) string {
|
||
s = strings.TrimSpace(s)
|
||
s = sanitizeNameOrLabel(s)
|
||
s = trimLinkTagEdgePunct(s)
|
||
return strings.TrimSpace(s)
|
||
}
|
||
|
||
// isValidNameOrLabel 仅允许包含至少一个汉字/字母/数字，纯符号（如 "、,、!）则跳过，不创建
|
||
func isValidNameOrLabel(s string) bool {
|
||
if s == "" {
|
||
return false
|
||
}
|
||
// 至少包含一个 Unicode 字母或数字
|
||
re := regexp.MustCompile(`[\p{L}\p{N}]`)
|
||
return re.MatchString(s)
|
||
}
|
||
|
||
// splitPersonAliases 解析 persons.aliases（逗号 / 中文逗号 / 顿号分隔）
|
||
func splitPersonAliases(s string) []string {
|
||
s = strings.TrimSpace(s)
|
||
if s == "" {
|
||
return nil
|
||
}
|
||
s = strings.ReplaceAll(s, "，", ",")
|
||
s = strings.ReplaceAll(s, "、", ",")
|
||
parts := strings.Split(s, ",")
|
||
out := make([]string, 0, len(parts))
|
||
for _, p := range parts {
|
||
t := strings.TrimSpace(p)
|
||
if t != "" {
|
||
out = append(out, t)
|
||
}
|
||
}
|
||
return out
|
||
}
|
||
|
||
// buildPersonAliasTokenMap 别名 -> person.token（同一别名多条人物时按 id 升序先命中者优先）
|
||
func buildPersonAliasTokenMap(db *gorm.DB) map[string]string {
|
||
var rows []model.Person
|
||
if err := db.Select("id", "token", "aliases").
|
||
Where("aliases IS NOT NULL AND aliases != ?", "").
|
||
Order("id ASC").
|
||
Find(&rows).Error; err != nil || len(rows) == 0 {
|
||
return map[string]string{}
|
||
}
|
||
out := make(map[string]string)
|
||
for _, p := range rows {
|
||
tok := strings.TrimSpace(p.Token)
|
||
if tok == "" {
|
||
continue
|
||
}
|
||
for _, a := range splitPersonAliases(p.Aliases) {
|
||
key := sanitizeNameOrLabel(a)
|
||
if key == "" || !isValidNameOrLabel(key) {
|
||
continue
|
||
}
|
||
if _, exists := out[key]; !exists {
|
||
out[key] = tok
|
||
}
|
||
}
|
||
}
|
||
return out
|
||
}
|
||
|
||
// ensurePersonByName 按名称或 aliases 确保 Person 存在，不存在则创建（含存客宝计划），返回 token
|
||
func ensurePersonByName(db *gorm.DB, name string, aliasToken map[string]string) (token string, err error) {
|
||
name = strings.TrimSpace(name)
|
||
if name == "" {
|
||
return "", nil
|
||
}
|
||
clean := sanitizeNameOrLabel(name)
|
||
if clean == "" || !isValidNameOrLabel(clean) {
|
||
return "", nil
|
||
}
|
||
var p model.Person
|
||
if db.Where("name = ?", clean).First(&p).Error == nil {
|
||
return p.Token, nil
|
||
}
|
||
// 按净化后的名字再查一次
|
||
if db.Where("name = ?", name).First(&p).Error == nil {
|
||
return p.Token, nil
|
||
}
|
||
if aliasToken != nil {
|
||
if tok := aliasToken[clean]; tok != "" {
|
||
return tok, nil
|
||
}
|
||
if tok := aliasToken[name]; tok != "" {
|
||
return tok, nil
|
||
}
|
||
}
|
||
// 无匹配人物：先建占位 users（无 open_id），再建 Person 并绑定 user_id，便于超级个体/会员体系与 @ 一致
|
||
u, err := createUserPlaceholderForArticleMention(db, clean)
|
||
if err != nil {
|
||
return "", err
|
||
}
|
||
created, err := createPersonMinimal(db, clean, u.ID)
|
||
if err != nil {
|
||
_ = db.Unscoped().Where("id = ?", u.ID).Delete(&model.User{})
|
||
return "", err
|
||
}
|
||
return created.Token, nil
|
||
}
|
||
|
||
// getPersonNameByToken 按 token 查 Person 返回 name，用于修复已损坏的 mention（显示 token 而非名字）
|
||
func getPersonNameByToken(db *gorm.DB, token string) string {
|
||
if token == "" {
|
||
return ""
|
||
}
|
||
var p model.Person
|
||
if db.Select("name").Where("token = ?", token).First(&p).Error != nil {
|
||
return ""
|
||
}
|
||
return p.Name
|
||
}
|
||
|
||
// ensureLinkTagByLabel 按 label 确保 LinkTag 存在，不存在则创建
|
||
func ensureLinkTagByLabel(db *gorm.DB, label string) (tagID string, err error) {
|
||
label = strings.TrimSpace(label)
|
||
if label == "" {
|
||
return "", nil
|
||
}
|
||
clean := sanitizeLinkTagLabel(label)
|
||
if clean == "" || !isValidNameOrLabel(clean) {
|
||
return "", nil
|
||
}
|
||
var t model.LinkTag
|
||
if db.Where("label = ?", clean).First(&t).Error == nil {
|
||
return t.TagID, nil
|
||
}
|
||
if db.Where("label = ?", label).First(&t).Error == nil {
|
||
return t.TagID, nil
|
||
}
|
||
t = model.LinkTag{TagID: clean, Label: clean, Type: "url"}
|
||
if err := db.Create(&t).Error; err != nil {
|
||
return "", err
|
||
}
|
||
return t.TagID, nil
|
||
}
|
||
|
||
// ParseAutoLinkContent 解析 content 中的 @人物 和 #标签，确保存在并转为带 data-id 的 span，后端统一处理
|
||
func ParseAutoLinkContent(content string) (string, error) {
|
||
if content == "" || (!strings.Contains(content, "@") && !strings.Contains(content, "#")) {
|
||
return content, nil
|
||
}
|
||
db := database.DB()
|
||
aliasToken := buildPersonAliasTokenMap(db)
|
||
|
||
// 1. 提取所有 @name 和 #label（排除 <> 避免匹配到 HTML 标签内）
|
||
mentionRe := regexp.MustCompile(`@([^\s@#<>]+)`)
|
||
tagRe := linkTagPlainTextRe
|
||
names := make(map[string]string) // cleanName -> token
|
||
labels := make(map[string]string) // cleanLabel -> tagId
|
||
|
||
for _, m := range mentionRe.FindAllStringSubmatch(content, -1) {
|
||
if len(m) < 2 {
|
||
continue
|
||
}
|
||
raw := m[1]
|
||
clean := sanitizeNameOrLabel(raw)
|
||
if clean == "" || !isValidNameOrLabel(clean) || names[clean] != "" {
|
||
continue
|
||
}
|
||
token, err := ensurePersonByName(db, clean, aliasToken)
|
||
if err == nil && token != "" {
|
||
names[clean] = token
|
||
names[raw] = token // 原始名也映射
|
||
}
|
||
}
|
||
for _, m := range tagRe.FindAllStringSubmatch(content, -1) {
|
||
if len(m) < 2 {
|
||
continue
|
||
}
|
||
raw := m[1]
|
||
clean := sanitizeLinkTagLabel(raw)
|
||
if clean == "" || !isValidNameOrLabel(clean) || labels[clean] != "" {
|
||
continue
|
||
}
|
||
tagID, err := ensureLinkTagByLabel(db, clean)
|
||
if err == nil && tagID != "" {
|
||
labels[clean] = tagID
|
||
labels[raw] = tagID
|
||
}
|
||
}
|
||
|
||
// 2. 占位符替换：先把已有 mention/linkTag span 保护起来
|
||
mentionSpanRe := regexp.MustCompile(`<span[^>]*data-type="mention"[^>]*>.*?</span>`)
|
||
linkTagSpanRe := regexp.MustCompile(`<span[^>]*data-type="linkTag"[^>]*>.*?</span>`)
|
||
placeholders := []string{}
|
||
|
||
content = mentionSpanRe.ReplaceAllStringFunc(content, func(m string) string {
|
||
// 回填 data-id、data-label：TipTap 用 data-label 显示名字，缺则显示 token
|
||
idRe := regexp.MustCompile(`data-id="([^"]*)"`)
|
||
labelRe := regexp.MustCompile(`data-label="([^"]*)"`)
|
||
innerRe := regexp.MustCompile(`>([^<]+)<`)
|
||
nickname := ""
|
||
if labelRe.MatchString(m) {
|
||
sub := labelRe.FindStringSubmatch(m)
|
||
if len(sub) >= 2 && strings.TrimSpace(sub[1]) != "" {
|
||
nickname = sanitizeNameOrLabel(sub[1])
|
||
}
|
||
}
|
||
if nickname == "" && innerRe.MatchString(m) {
|
||
sub := innerRe.FindStringSubmatch(m)
|
||
if len(sub) >= 2 {
|
||
nickname = sanitizeNameOrLabel(strings.TrimPrefix(sub[1], "@"))
|
||
}
|
||
}
|
||
// 若 inner 是 token（如已损坏显示 @pZefXfWlon...），用 token 查 Person 取真实名字
|
||
if idRe.MatchString(m) {
|
||
sub := idRe.FindStringSubmatch(m)
|
||
if len(sub) >= 2 && strings.TrimSpace(sub[1]) != "" {
|
||
tokenVal := sub[1]
|
||
innerSub := innerRe.FindStringSubmatch(m)
|
||
innerText := ""
|
||
if len(innerSub) >= 2 {
|
||
innerText = strings.TrimPrefix(innerSub[1], "@")
|
||
}
|
||
// 若 inner 看起来像 token（长串字母数字）且与 data-id 一致，用 DB 查真实名字
|
||
if innerText == tokenVal && len(tokenVal) >= 20 {
|
||
if realName := getPersonNameByToken(db, tokenVal); realName != "" {
|
||
nickname = realName
|
||
}
|
||
}
|
||
}
|
||
}
|
||
if nickname != "" {
|
||
if token, ok := names[nickname]; ok && token != "" {
|
||
if idRe.MatchString(m) {
|
||
m = idRe.ReplaceAllString(m, `data-id="`+token+`"`)
|
||
} else {
|
||
m = strings.Replace(m, `data-type="mention"`, `data-type="mention" data-id="`+token+`"`, 1)
|
||
}
|
||
}
|
||
// 确保有 data-label，否则 TipTap 会显示 token 而非名字
|
||
needLabel := !labelRe.MatchString(m)
|
||
if !needLabel {
|
||
if sub := labelRe.FindStringSubmatch(m); len(sub) >= 2 && strings.TrimSpace(sub[1]) == "" {
|
||
needLabel = true
|
||
}
|
||
}
|
||
if needLabel {
|
||
m = strings.Replace(m, `data-type="mention"`, `data-type="mention" data-label="`+nickname+`"`, 1)
|
||
}
|
||
}
|
||
placeholders = append(placeholders, m)
|
||
return "\x00PLACEHOLDER\x00"
|
||
})
|
||
|
||
content = linkTagSpanRe.ReplaceAllStringFunc(content, func(m string) string {
|
||
placeholders = append(placeholders, m)
|
||
return "\x00PLACEHOLDER\x00"
|
||
})
|
||
|
||
// 3. 替换纯文本 @name 和 #label
|
||
content = mentionRe.ReplaceAllStringFunc(content, func(m string) string {
|
||
raw := strings.TrimPrefix(m, "@")
|
||
clean := sanitizeNameOrLabel(raw)
|
||
token := names[clean]
|
||
if token == "" {
|
||
token = names[raw]
|
||
}
|
||
if token == "" {
|
||
return m
|
||
}
|
||
return `<span data-type="mention" data-id="` + token + `" data-label="` + raw + `" class="mention-tag">@` + raw + `</span>`
|
||
})
|
||
|
||
content = tagRe.ReplaceAllStringFunc(content, func(m string) string {
|
||
raw := strings.TrimPrefix(m, "#")
|
||
clean := sanitizeLinkTagLabel(raw)
|
||
if clean == "" || !isValidNameOrLabel(clean) {
|
||
return m
|
||
}
|
||
tagID := labels[clean]
|
||
if tagID == "" {
|
||
tagID = labels[raw]
|
||
}
|
||
if tagID == "" {
|
||
return m
|
||
}
|
||
return `<span data-type="linkTag" data-url="" data-tag-type="url" data-tag-id="` + tagID + `" data-page-path="" data-app-id="" class="link-tag-node">#` + clean + `</span>`
|
||
})
|
||
|
||
// 4. 恢复占位符
|
||
for _, p := range placeholders {
|
||
content = strings.Replace(content, "\x00PLACEHOLDER\x00", p, 1)
|
||
}
|
||
|
||
return content, nil
|
||
}
|