初次提交

VoiceHub-main/app/utils/lyric/exclude.ts

@@ -0,0 +1,309 @@
+/**
+ * @fileoverview 跟元数据清理器有关的配置内容
+ */
+
+/**
+ * 默认的关键字列表，不包含冒号
+ *
+ * 建议更新这里的列表后运行 scripts\sort-keywords.ts 来排序并去重列表
+ */
+export const keywords = [
+  '版权',
+  '贝斯',
+  '贝斯Bass',
+  '本歌曲商用授权，前往小程序',
+  '编曲',
+  '编曲 Arranger',
+  '编曲Arranged By',
+  '编曲Arrangement',
+  '策划',
+  '策划 Planner',
+  '策划统筹',
+  '策划统筹Planner and coordinator',
+  '出品',
+  '出品、发行',
+  '出品/发行',
+  '出品/发行方',
+  '出品/发行公司',
+  '出品方Presented by',
+  '出品方Presenter',
+  '出品公司',
+  '出品公司Production Company',
+  '出品人',
+  '出品人 Publisher',
+  '出品人Chief Producer',
+  '出品人Presenter',
+  '出品人Producer',
+  '出品团队',
+  '出品团队Production Team',
+  '出品Present by',
+  '词',
+  '词 Lyricist',
+  '词曲提供',
+  '词曲提供Lyrics & Composition Provided by',
+  '词曲提供Lyrics and Composition Provided by',
+  '词曲提供Music and Lyrics Provided by',
+  '词曲提供Words and Music by',
+  '词曲协力',
+  '词Lyrics',
+  '大提琴',
+  '大提琴 Cello',
+  '灯光设计',
+  '第二小提琴',
+  '第二小提琴 Second Violin',
+  '第一小提琴',
+  '第一小提琴 First Violin',
+  '发行',
+  '发行方',
+  '发行方Publisher',
+  '发行公司',
+  '发行人',
+  '发行Distributed by',
+  '发行Distribution',
+  '发行Release',
+  '发行Released by',
+  '翻唱',
+  '封面',
+  '封面插画Cover illustration',
+  '封面设计',
+  '封面设计 Cover Design',
+  '钢琴/合成器演奏',
+  '钢琴演奏',
+  '歌名',
+  '歌手',
+  '工作室',
+  '鼓',
+  '鼓Drums',
+  '合成器演奏',
+  '合声',
+  '合声编写',
+  '和声',
+  '和声 Backing Vocals',
+  '和声/和声设计',
+  '和声编写',
+  '和声编写Backing Vocal Arrangement',
+  '和声编写Backing Vocals Design',
+  '和声演唱',
+  '和声Backing Vocal',
+  '和音',
+  '和音 Bvox',
+  '混缩',
+  '混音',
+  '混音 Mixed by',
+  '混音 Mixing Engineer',
+  '混音/母带工程师',
+  '混音工程师',
+  '混音工程师Mixing Engineer',
+  '混音师',
+  '混音室',
+  '混音室Mixing Studio',
+  '吉他',
+  '吉他 Guitar',
+  '吉他Guitars',
+  '监制',
+  '监制 Deputy Executive Producer',
+  '监制Chief Producer',
+  '监制Executive Director',
+  '监制Supervised production',
+  '键盘',
+  '乐队统筹',
+  '乐器演奏',
+  '乐团',
+  '联合策划',
+  '联合出品',
+  '联合出品Co-produced by',
+  '联合出品Co-production',
+  '联合出品Jointly Produced by',
+  '联合出品Published',
+  '联合推广',
+  '录/混音',
+  '录音',
+  '录音 Recording Engineer',
+  '录音师',
+  '录音师Recording Engineer',
+  '录音室',
+  '录音室 Recording Studio',
+  '录音制作',
+  '鸣谢',
+  '母带',
+  '母带 Mastered by',
+  '母带处理工程师',
+  '母带工程师',
+  '母带工程师Mastering Engineer',
+  '母带后期处理工程师',
+  '母带后期处理工程师Mastering Engineer',
+  '母带后期处理录音室',
+  '母带后期处理录音室Mastering Studio',
+  '母带后期制作人',
+  '母带后期制作人Mastering Producer',
+  '配唱制作人',
+  '配唱制作人 Vocals Producer',
+  '配唱制作人Vocal Producer',
+  '企划',
+  '企划宣传',
+  '企划宣传Propaganda',
+  '企划营销',
+  '企划制作A&R Planning',
+  '企划Planning',
+  '曲',
+  '曲 Composer',
+  '设计',
+  '设计Design',
+  '特别鸣谢',
+  '特别鸣谢/艺人支持',
+  '特别鸣谢Acknowledgement',
+  '特别企划',
+  '题记',
+  '统筹',
+  '统筹Coordinator',
+  '统筹Planning',
+  '推广',
+  '推广策划',
+  '推广策划Marketing Strategy',
+  '推广策划Promotion Planning',
+  '推广策划Promotion Strategy',
+  '推广统筹 Promotion Coordinator',
+  '推广宣传',
+  '推广营销 Marketing Promotion',
+  '维伴音乐',
+  '文案',
+  '文案Copywriting',
+  '弦乐',
+  '弦乐 Strings',
+  '弦乐编写',
+  '弦乐编写Strings Arrangement',
+  '弦乐监制',
+  '弦乐录音师',
+  '弦乐录音室',
+  '项目统筹',
+  '项目统筹 Project Coordinator',
+  '宣传',
+  '宣传/推广',
+  '宣推',
+  '演唱',
+  '艺术指导',
+  '艺术指导Art Director',
+  '音乐项目总监Project Executive',
+  '音乐制作',
+  '音乐制作发行',
+  '音乐制作Music Production',
+  '音频编辑',
+  '营销推广',
+  '营销推广/出品/发行',
+  '营销推广Marketing',
+  '营销推广Marketing Promotion',
+  '原唱',
+  '制作发行',
+  '制作公司',
+  '制作公司 Production',
+  '制作公司/OP',
+  '制作公司Produce Company',
+  '制作人',
+  '制作人 Producer',
+  '制作人Music Producer',
+  '制作人Record Producer',
+  '制作统筹',
+  '制作统筹Executive Producer',
+  '制作团队',
+  '制作助理',
+  '中提琴',
+  '中提琴 Viola',
+  '专辑',
+  '总策划',
+  '总策划 Chief Planner',
+  '总策划Chref Planner',
+  '总监制 Chief Executive Producer',
+  '总企划',
+  '作词',
+  '作曲',
+  'Acknowledgement',
+  'Additional Vocal by',
+  'Arranged By',
+  'Arrangement',
+  'Arranger',
+  'Art Director',
+  'Artist',
+  'Background Vocals by',
+  'Backing Vocal',
+  'Backing Vocal Arrangement',
+  'Backing Vocals',
+  'Backing Vocals Design',
+  'Bass',
+  'Cello',
+  'Chief Producer',
+  'Chref Planner',
+  'Co-produced by',
+  'Co-Producer',
+  'Co-production',
+  'Composed by',
+  'Composer',
+  'Copywriting',
+  'Cover Design',
+  'Design',
+  'Distributed by',
+  'Distribution',
+  'Drums',
+  'Executive Producer',
+  'First Violin',
+  'Guitar',
+  'Guitars',
+  'Jointly Produced by',
+  'Lyricist',
+  'Lyrics',
+  'Lyrics & Composition Provided by',
+  'Lyrics and Composition Provided by',
+  'Lyrics by',
+  'Marketing Promotion',
+  'Marketing Strategy',
+  'Mastered by',
+  'Mastering Engineer',
+  'Mixing Engineer',
+  'Music and Lyrics Provided by',
+  'Music Production',
+  'OP',
+  'OP、SP',
+  'OP/发行',
+  'OP/SP',
+  'Planner',
+  'Planner and coordinator',
+  'Presented by',
+  'Presenter',
+  'Produce Company',
+  'Produced by',
+  'Producer',
+  'Production Company',
+  'Production Team',
+  'Promotion Planning',
+  'Promotion Strategy',
+  'Propaganda',
+  'Published by',
+  'Publisher',
+  'Record Producer',
+  'Recorded at',
+  'Recording Engineer',
+  'Recording Studio',
+  'Release',
+  'Released by',
+  'Repertoire Owner',
+  'Second Violin',
+  'Songs Title',
+  'SP',
+  'Strings',
+  'Strings Arrangement',
+  'Supervised production',
+  'Viola',
+  'Vocal Engineer',
+  'Vocal Producer',
+  'Vocals by',
+  'Vocals Produced by',
+  'Words and Music by',
+  'Written by'
+]
+
+/**
+ * 默认的正则表达式
+ */
+export const regexes = [
+  '(?:【.*?音乐人.*?】|\\(.*?音乐人.*?\\)|「.*?音乐人.*?」|（.*?音乐人.*?）|『.*?音乐人.*?』)',
+  '.*?未经.*?不得.*?'
+]

VoiceHub-main/app/utils/lyric/lyricFormat.ts

@@ -0,0 +1,230 @@
+import type { LyricLine } from '@applemusic-like-lyrics/lyric'
+import { stripLyricMetadata, type StripOptions } from './lyricStripper'
+import { keywords as defaultKeywords, regexes as defaultRegexes } from './exclude'
+import { cloneDeep } from 'lodash-es'
+
+/**
+ * 获取替换配置
+ */
+const getReplacementConfig = (settings: any) => {
+  const preset = settings.bracketReplacementPreset.value || 'dash'
+  const custom = settings.customBracketReplacement.value || '-'
+
+  let startStr = ' - '
+  const endStr = ' '
+  const isEnclosure = false
+
+  // 预设模式
+  if (preset === 'angleBrackets') {
+    return { startStr: '〔', endStr: '〕', isEnclosure: true }
+  } else if (preset === 'cornerBrackets') {
+    return { startStr: '「', endStr: '」', isEnclosure: true }
+  } else if (preset === 'custom') {
+    const trimmed = custom.trim()
+    // 启发式判断：如果是两个不同字符且不包含减号，视为成对符号 (如 "()")
+    const isPair = trimmed.length === 2 && trimmed[0] !== trimmed[1] && !trimmed.includes('-')
+    if (isPair) {
+      return { startStr: trimmed[0], endStr: trimmed[1], isEnclosure: true }
+    } else {
+      // 否则视为分隔符 (如 "-")，加上空格规范化
+      startStr = ' ' + trimmed + ' '
+      startStr = startStr.replace(/\s+/g, ' ')
+      return { startStr, endStr: ' ', isEnclosure: false }
+    }
+  } else if (preset === 'space') {
+    return { startStr: ' ', endStr: ' ', isEnclosure: false }
+  }
+
+  // 默认 dash 模式
+  return { startStr, endStr, isEnclosure }
+}
+
+/**
+ * 应用括号替换逻辑
+ */
+const applyBracketReplacement = (lines: LyricLine[], settings: any): LyricLine[] => {
+  if (!settings.replaceLyricBrackets.value) return lines
+
+  const { startStr, endStr, isEnclosure } = getReplacementConfig(settings)
+
+  // 正则：匹配全行括号 (如 "(Music)")
+  const fullBracketRegex = /^\s*[(（][^()（）]*[)）]\s*$/
+  // 正则：匹配左括号
+  const leftBracketRegex = /[(（]/g
+  // 正则：匹配右括号
+  const rightBracketRegex = /[)）]/g
+
+  /**
+   * 处理普通字符串 (用于翻译和罗马音)
+   */
+  const processString = (str: string): string => {
+    if (!str) return str
+
+    // 如果是全行括号且非闭合模式 (如分隔符模式)，则直接去除括号
+    if (!isEnclosure && fullBracketRegex.test(str)) {
+      return str
+        .replace(/^\s*[(（]/, '')
+        .replace(/[)）]\s*$/, '')
+        .trim()
+    }
+
+    let res = str.replace(leftBracketRegex, startStr)
+    if (isEnclosure) {
+      res = res.replace(rightBracketRegex, endStr)
+    } else {
+      // 分隔符模式：右括号如果是结尾则移除，否则替换为结束符(空格)
+      res = res.replace(/[)）](?=\s*$)/g, '').replace(rightBracketRegex, endStr)
+      // 清理重复分隔符 (如 " -  - " -> " - ")
+      if (startStr.includes('-')) {
+        res = res.replace(/(?:\s*-\s*){2,}/g, ' - ')
+      }
+    }
+    return res
+  }
+
+  return lines.map((line) => {
+    // 处理 words
+    if (line.words) {
+      const fullText = line.words.map((w) => w.word).join('')
+      const isFullBracket = fullBracketRegex.test(fullText)
+
+      // 整行是括号内容，且处于分隔符模式 (非闭合) -> 直接移除首尾括号，不加分隔符
+      if (isFullBracket && !isEnclosure) {
+        let removedStart = false
+        let removedEnd = false
+
+        // 移除第一个左括号
+        for (const word of line.words) {
+          if (!removedStart && leftBracketRegex.test(word.word)) {
+            word.word = word.word.replace(leftBracketRegex, '')
+            removedStart = true
+          }
+        }
+        // 移除最后一个右括号
+        for (let i = line.words.length - 1; i >= 0; i--) {
+          const word = line.words[i]
+          if (!removedEnd && rightBracketRegex.test(word.word)) {
+            const lastIndex = Math.max(word.word.lastIndexOf(')'), word.word.lastIndexOf('）'))
+            if (lastIndex !== -1) {
+              word.word = word.word.substring(0, lastIndex) + word.word.substring(lastIndex + 1)
+              removedEnd = true
+            }
+          }
+        }
+      } else {
+        // 普通替换逻辑
+        line.words.forEach((word, index) => {
+          // 替换左括号
+          word.word = word.word.replace(leftBracketRegex, startStr)
+
+          if (isEnclosure) {
+            // 闭合模式：直接替换右括号
+            word.word = word.word.replace(rightBracketRegex, endStr)
+          } else {
+            // 分隔符模式：
+            word.word = word.word.replace(rightBracketRegex, (_, offset, str) => {
+              // 如果右括号在单词末尾，且是整句的最后一个单词 -> 移除 (视为句尾闭合)
+              const isAtWordEnd = offset === str.length - 1
+              const isAtLineEnd = index === line.words.length - 1
+
+              if (isAtWordEnd) {
+                return isAtLineEnd ? '' : endStr
+              }
+              return endStr
+            })
+          }
+        })
+
+        // 清理逻辑 (仅针对带减号的分隔符模式)
+        if (!isEnclosure && startStr.includes('-')) {
+          line.words.forEach((word, index) => {
+            // 单词内重复: " -  - " -> " - "
+            word.word = word.word.replace(/(?:\s*-\s*){2,}/g, ' - ')
+
+            // 单词间重复: 前一个词以 "- " 结尾，当前词以 " -" 开头 -> 移除前一个的尾部
+            if (index > 0) {
+              const prev = line.words[index - 1]
+              if (/-\s*$/.test(prev.word) && /^\s*-/.test(word.word)) {
+                prev.word = prev.word.replace(/-\s*$/, '')
+                // 确保当前词开头有完整分隔符
+                if (!/^\s*-\s+/.test(word.word)) {
+                  word.word = ' - ' + word.word.replace(/^\s*-\s*/, '')
+                }
+              }
+            }
+          })
+        }
+      }
+    }
+
+    // 处理翻译和罗马音
+    if (line.translatedLyric) line.translatedLyric = processString(line.translatedLyric)
+    if (line.romanLyric) line.romanLyric = processString(line.romanLyric)
+
+    return line
+  })
+}
+
+/**
+ * 格式化歌词
+ * 主要处理括号替换、关键词过滤等
+ * @param lines 歌词行
+ * @param settings 歌词设置对象 (从 useLyricSettings 获取)
+ * @param metadata 可选的歌曲元数据，用于更精确的歌词清洗
+ */
+export const formatLyric = (
+  lines: LyricLine[],
+  settings: any,
+  metadata?: { title?: string; artists?: string[] }
+): LyricLine[] => {
+  const { enableExcludeLyrics, excludeLyricsUserKeywords, excludeLyricsUserRegexes } = settings
+
+  if (!lines || lines.length === 0) return []
+
+  // 使用 cloneDeep 避免修改原始引用 (如果需要)
+  // 但 formatLyric 通常被视为纯函数处理流程，这里我们操作的是 parsedLyrics 的结果
+  // 为了安全起见，克隆一份
+  let processedLines = cloneDeep(lines)
+
+  // 1. 智能元数据清洗 (Smart Metadata Stripping)
+  if (enableExcludeLyrics && enableExcludeLyrics.value) {
+    const userKeywords = excludeLyricsUserKeywords.value || []
+    const userRegexes = excludeLyricsUserRegexes.value || []
+
+    const mergedKeywords = [...new Set([...defaultKeywords, ...userKeywords])]
+    const mergedRegexes = [...new Set([...defaultRegexes, ...userRegexes])]
+
+    const options: StripOptions = {
+      keywords: mergedKeywords,
+      regexPatterns: mergedRegexes,
+      matchMetadata: metadata
+    }
+
+    processedLines = stripLyricMetadata(processedLines, options)
+  }
+
+  // 2. 基础过滤 (保留)
+  processedLines = processedLines.filter((line) => {
+    const text = line.words.map((w) => w.word).join('')
+    const transText = line.translatedLyric || ''
+
+    if (!text.trim()) {
+      if (!transText.trim() || transText.includes('著作权') || transText.includes('QQ音乐')) {
+        return false
+      }
+    }
+
+    if (text.includes('著作权') || text.includes('QQ音乐')) return false
+
+    if (transText.includes('著作权') || transText.includes('QQ音乐') || transText.trim() === '//') {
+      line.translatedLyric = ''
+    }
+
+    return true
+  })
+
+  // 3. 括号替换
+  processedLines = applyBracketReplacement(processedLines, settings)
+
+  return processedLines
+}

VoiceHub-main/app/utils/lyric/lyricParser.ts

@@ -0,0 +1,506 @@
+import { cloneDeep } from 'lodash-es'
+import type { LyricLine } from '@applemusic-like-lyrics/lyric'
+import { parseTTML as parseTTMLLib } from '@applemusic-like-lyrics/lyric'
+import { extractLyricContent } from './qrc-parser'
+import { parseLrc } from './parseLrc'
+
+/**
+ * LRC 格式类型
+ */
+export enum LrcFormat {
+  /** 普通逐行 LRC */
+  Line = 'line',
+  /** 逐字 LRC：[00:28.850]曲[00:32.455]：[00:36.060]钱 */
+  WordByWord = 'word-by-word',
+  /** 增强型 LRC (ESLyric)：[01:37.305]<01:37.624>怕<01:37.943>你 */
+  Enhanced = 'enhanced'
+}
+
+/** LyricWord 类型 */
+type LyricWord = { word: string; startTime: number; endTime: number; romanWord: string }
+
+// 预编译正则表达式
+const META_TAG_REGEX = /^\[[a-z]+:/i
+const TIME_TAG_REGEX = /\[(\d{2}):(\d{2})\.(\d{1,})\]/g
+const ENHANCED_TIME_TAG_REGEX = /<(\d{2}):(\d{2})\.(\d{1,})>/
+// 移除全局带状态的正则，改为在函数内使用 matchAll 或重新构建
+const LINE_TIME_REGEX = /^\[(\d{2}):(\d{2})\.(\d{1,})\]/
+
+// QRC 解析相关正则 - 提前编译
+const QRC_LINE_PATTERN = /^\[(\d+),(\d+)\](.*)$/
+const QRC_WORD_PATTERN = /([^(]*)\((\d+),(\d+)\)/g
+
+const DEFAULT_WORD_DURATION = 1000
+const ALIGN_TOLERANCE_MS = 300
+
+/**
+ * 解析时间戳为毫秒
+ * 使用字符串补齐处理，避免浮点数计算误差
+ */
+const parseTimeToMs = (min: string, sec: string, ms: string): number => {
+  const minutes = parseInt(min, 10)
+  const seconds = parseInt(sec, 10)
+  // 补齐到 3 位 (例如 "5" -> "500", "05" -> "050", "1234" -> "123")
+  const msNormalized = ms.padEnd(3, '0').slice(0, 3)
+  const milliseconds = parseInt(msNormalized, 10)
+  return minutes * 60 * 1000 + seconds * 1000 + milliseconds
+}
+
+/**
+ * 创建 LyricWord 对象
+ */
+const createWord = (word: string, startTime: number, endTime: number = startTime): LyricWord => ({
+  word,
+  startTime,
+  endTime,
+  romanWord: ''
+})
+
+/**
+ * 创建 LyricLine 对象
+ */
+const createLine = (words: LyricWord[], startTime: number, endTime: number = 0): LyricLine => ({
+  words,
+  startTime,
+  endTime,
+  translatedLyric: '',
+  romanLyric: '',
+  isBG: false,
+  isDuet: false
+})
+
+/**
+ * 检测 LRC 格式类型
+ */
+export const detectLrcFormat = (content: string): LrcFormat => {
+  const lines = content.split(/\r?\n/)
+  for (const rawLine of lines) {
+    const line = rawLine.trim()
+    if (!line || META_TAG_REGEX.test(line)) continue
+    // 检查增强型LRC
+    if (ENHANCED_TIME_TAG_REGEX.test(line)) {
+      return LrcFormat.Enhanced
+    }
+    // 检查逐字LRC
+    const matches = line.match(TIME_TAG_REGEX)
+    if (matches && matches.length > 1) {
+      return LrcFormat.WordByWord
+    }
+  }
+  return LrcFormat.Line
+}
+
+/**
+ * 解析逐字 LRC 格式
+ * 优化：在解析过程中直接计算 endTime，避免二次遍历
+ */
+export const parseWordByWordLrc = (content: string): LyricLine[] => {
+  const result: LyricLine[] = []
+  let prevLine: LyricLine | null = null
+  const WORD_BY_WORD_PATTERN = /\[(\d{2}):(\d{2})\.(\d{1,})\]([^[\\]]*)/g
+
+  for (const rawLine of content.split(/\r?\n/)) {
+    const line = rawLine.trim()
+    if (!line || META_TAG_REGEX.test(line)) continue
+
+    const words: LyricWord[] = []
+    let lineStartTime = Infinity
+
+    let prevWord: LyricWord | null = null
+
+    const matches = line.matchAll(WORD_BY_WORD_PATTERN)
+
+    for (const match of matches) {
+      const startTime = parseTimeToMs(match[1], match[2], match[3])
+      const wordText = match[4]
+
+      if (!wordText && words.length === 0) continue
+
+      lineStartTime = Math.min(lineStartTime, startTime)
+
+      // 设置上一个字的结束时间
+      if (prevWord) {
+        prevWord.endTime = startTime
+      }
+
+      if (wordText) {
+        const newWord = createWord(wordText, startTime)
+        words.push(newWord)
+        prevWord = newWord
+      }
+    }
+
+    // 处理行内最后一个字
+    if (prevWord) {
+      prevWord.endTime = prevWord.startTime + DEFAULT_WORD_DURATION
+    }
+
+    if (words.length > 0) {
+      const lineObj = createLine(words, lineStartTime === Infinity ? 0 : lineStartTime)
+      // 设置行结束时间为最后一个字的结束时间
+      lineObj.endTime = words[words.length - 1].endTime
+
+      // 修正上一行的结束时间 (Single Pass)
+      if (prevLine) {
+        const prevLastWord = prevLine.words[prevLine.words.length - 1]
+        // 只有当当前行开始时间晚于上一行最后一个字的开始时间时，才进行截断
+        if (lineObj.startTime > prevLastWord.startTime) {
+          prevLastWord.endTime = Math.min(prevLastWord.endTime, lineObj.startTime)
+          prevLine.endTime = prevLastWord.endTime
+        }
+      }
+
+      result.push(lineObj)
+      prevLine = lineObj
+    }
+  }
+
+  return result
+}
+
+/**
+ * 解析增强型 LRC 格式 (ESLyric)
+ */
+export const parseEnhancedLrc = (content: string): LyricLine[] => {
+  const result: LyricLine[] = []
+  let prevLine: LyricLine | null = null
+  const ENHANCED_WORD_PATTERN = /<(\d{2}):(\d{2})\.(\d{1,})>([^<]*)/g
+
+  for (const rawLine of content.split(/\r?\n/)) {
+    const line = rawLine.trim()
+    if (!line || META_TAG_REGEX.test(line)) continue
+
+    const lineTimeMatch = LINE_TIME_REGEX.exec(line)
+    if (!lineTimeMatch) continue
+
+    const lineStartTime = parseTimeToMs(lineTimeMatch[1], lineTimeMatch[2], lineTimeMatch[3])
+    const contentAfterTime = line.slice(lineTimeMatch[0].length)
+
+    const words: LyricWord[] = []
+
+    // 检查是否有增强型标记
+    if (ENHANCED_TIME_TAG_REGEX.test(contentAfterTime)) {
+      let prevWord: LyricWord | null = null
+
+      const matches = contentAfterTime.matchAll(ENHANCED_WORD_PATTERN)
+
+      for (const match of matches) {
+        const startTime = parseTimeToMs(match[1], match[2], match[3])
+        const wordText = match[4]
+
+        if (prevWord) {
+          prevWord.endTime = startTime
+        }
+
+        if (wordText) {
+          const newWord = createWord(wordText, startTime)
+          words.push(newWord)
+          prevWord = newWord
+        }
+      }
+
+      if (prevWord) {
+        prevWord.endTime = prevWord.startTime + DEFAULT_WORD_DURATION // 默认兜底
+      }
+    } else {
+      // 无增强型标记，作为整行处理
+      const text = contentAfterTime.trim()
+      if (text) {
+        words.push(createWord(text, lineStartTime, lineStartTime + DEFAULT_WORD_DURATION)) // 默认持续1s
+      }
+    }
+
+    if (words.length > 0) {
+      const lineObj = createLine(words, lineStartTime)
+      lineObj.endTime = words[words.length - 1].endTime
+
+      // 修正上一行的结束时间 (Single Pass)
+      if (prevLine) {
+        const prevLastWord = prevLine.words[prevLine.words.length - 1]
+        if (lineObj.startTime > prevLastWord.startTime) {
+          prevLastWord.endTime = Math.min(prevLastWord.endTime, lineObj.startTime)
+          prevLine.endTime = prevLastWord.endTime
+        }
+      }
+
+      result.push(lineObj)
+      prevLine = lineObj
+    }
+  }
+
+  return result
+}
+
+/**
+ * 智能解析 LRC 歌词
+ */
+export const parseSmartLrc = (content: string): { format: LrcFormat; lines: LyricLine[] } => {
+  const format = detectLrcFormat(content)
+
+  let lines: LyricLine[]
+  switch (format) {
+    case LrcFormat.WordByWord:
+      lines = parseWordByWordLrc(content)
+      break
+    case LrcFormat.Enhanced:
+      lines = parseEnhancedLrc(content)
+      break
+    default:
+      lines = parseLrc(content) || []
+  }
+
+  console.log(`[LyricParser] 检测到歌词格式: ${format}, 共 ${lines.length} 行`)
+  return { format, lines }
+}
+
+/**
+ * 判断解析结果是否为逐字格式
+ */
+export const isWordLevelFormat = (format: LrcFormat): boolean =>
+  format === LrcFormat.WordByWord || format === LrcFormat.Enhanced
+
+/**
+ * 歌词内容对齐
+ * 使用双指针算法实现 O(N) 复杂度
+ * @param lyrics 歌词数据 (Readonly)
+ * @param otherLyrics 其他歌词数据
+ * @param key 对齐类型
+ * @returns 对齐后的歌词数据 (新副本)
+ */
+export const alignLyrics = (
+  lyrics: Readonly<LyricLine[]>,
+  otherLyrics: Readonly<LyricLine[]>,
+  key: 'translatedLyric' | 'romanLyric'
+): LyricLine[] => {
+  if (!lyrics.length || !otherLyrics.length) return cloneDeep(lyrics) as LyricLine[]
+
+  const result = cloneDeep(lyrics) as LyricLine[]
+
+  let i = 0
+  let j = 0
+
+  while (i < result.length && j < otherLyrics.length) {
+    const line = result[i]
+    const other = otherLyrics[j]
+    const diff = line.startTime - other.startTime
+
+    if (Math.abs(diff) <= ALIGN_TOLERANCE_MS) {
+      // 匹配成功
+      line[key] = other.words.map((word) => word.word).join('')
+      i++
+      j++
+    } else if (diff < 0) {
+      // 当前歌词时间较早，移动当前指针
+      i++
+    } else {
+      // 目标歌词时间较早，移动目标指针
+      j++
+    }
+  }
+  return result
+}
+
+/**
+ * 解析 QRC 内容为行数据
+ */
+const parseQRCContent = (
+  rawContent: string
+): Array<{
+  startTime: number
+  endTime: number
+  words: Array<{ word: string; startTime: number; endTime: number }>
+}> => {
+  // 使用策略模式提取 LyricContent (自动适配 Browser/Node 环境)
+  const content = extractLyricContent(rawContent) || rawContent
+
+  const result: Array<{
+    startTime: number
+    endTime: number
+    words: Array<{ word: string; startTime: number; endTime: number }>
+  }> = []
+
+  for (const rawLine of content.split('\n')) {
+    const line = rawLine.trim()
+    if (!line) continue
+
+    // 跳过元数据标签 [ti:xxx] [ar:xxx] 等
+    if (META_TAG_REGEX.test(line)) continue
+
+    const lineMatch = QRC_LINE_PATTERN.exec(line)
+    if (!lineMatch) continue
+
+    const lineStart = parseInt(lineMatch[1], 10)
+    const lineDuration = parseInt(lineMatch[2], 10)
+    const lineContent = lineMatch[3]
+
+    // 解析逐字
+    const words: Array<{ word: string; startTime: number; endTime: number }> = []
+
+    const matches = lineContent.matchAll(QRC_WORD_PATTERN)
+
+    for (const match of matches) {
+      const wordText = match[1]
+      const wordStart = parseInt(match[2], 10)
+      const wordDuration = parseInt(match[3], 10)
+
+      if (wordText) {
+        words.push({
+          word: wordText,
+          startTime: wordStart,
+          endTime: wordStart + wordDuration
+        })
+      }
+    }
+
+    if (words.length > 0) {
+      result.push({
+        startTime: lineStart,
+        endTime: lineStart + lineDuration,
+        words
+      })
+    }
+  }
+  return result
+}
+
+/**
+ * 解析 QQ 音乐 QRC 格式歌词
+ * @param qrcContent QRC 原始内容
+ * @param trans 翻译歌词
+ * @param roma 罗马音歌词（QRC 格式）
+ * @returns LyricLine 数组
+ */
+export const parseQRCLyric = (qrcContent: string, trans?: string, roma?: string): LyricLine[] => {
+  // 解析主歌词
+  const qrcLines = parseQRCContent(qrcContent)
+  let result: LyricLine[] = qrcLines.map((qrcLine) => {
+    return {
+      words: qrcLine.words.map((word) => ({
+        ...word,
+        romanWord: ''
+      })),
+      startTime: qrcLine.startTime,
+      endTime: qrcLine.endTime,
+      translatedLyric: '',
+      romanLyric: '',
+      isBG: false,
+      isDuet: false
+    }
+  })
+
+  // 处理翻译
+  if (trans) {
+    let transLines = parseLrc(trans)
+    if (transLines?.length) {
+      // 过滤包含 "//" 或 "作品的著作权" 的翻译行
+      transLines = transLines.filter((line) => {
+        const text = line.words.map((w) => w.word).join('')
+        return !text.includes('//') && !text.includes('作品的著作权')
+      })
+      result = alignLyrics(result, transLines, 'translatedLyric')
+    }
+  }
+
+  // 处理音译
+  if (roma) {
+    const qrcRomaLines = parseQRCContent(roma)
+    if (qrcRomaLines?.length) {
+      const romaLines: LyricLine[] = qrcRomaLines.map((line) => {
+        return {
+          words: [
+            {
+              startTime: line.startTime,
+              endTime: line.endTime,
+              word: line.words.map((w) => w.word).join(''),
+              romanWord: ''
+            }
+          ],
+          startTime: line.startTime,
+          endTime: line.endTime,
+          translatedLyric: '',
+          romanLyric: '',
+          isBG: false,
+          isDuet: false
+        }
+      })
+      result = alignLyrics(result, romaLines, 'romanLyric')
+    }
+  }
+
+  return result
+}
+
+/**
+ * 清洗 TTML 中不需要的翻译
+ * @param ttmlContent 原始 TTML 内容
+ * @returns 清洗后的 TTML 内容
+ */
+export const cleanTTMLTranslations = (ttmlContent: string): string => {
+  const lang_counter = (ttml_text: string) => {
+    // 使用正则匹配所有 xml:lang="xx-XX" 格式的字符串
+    const langRegex = /(?<=<(span|translation)[^<>]+)xml:lang="([^"]+)"/g
+    const matches = ttml_text.matchAll(langRegex)
+
+    // 提取匹配结果并去重
+    const langSet = new Set<string>()
+    for (const match of matches) {
+      if (match[2]) langSet.add(match[2])
+    }
+
+    return Array.from(langSet)
+  }
+
+  const lang_filter = (langs: string[]): string | null => {
+    if (langs.length <= 1) return null
+
+    const lang_matcher = (target: string) => {
+      return langs.find((lang) => {
+        try {
+          // @ts-expect-error Intl.Locale type definition is missing maximize method
+          return new Intl.Locale(lang).maximize().script === target
+        } catch {
+          return false
+        }
+      })
+    }
+
+    const hans_matched = lang_matcher('Hans')
+    if (hans_matched) return hans_matched
+
+    const hant_matched = lang_matcher('Hant')
+    if (hant_matched) return hant_matched
+
+    const major = langs.find((key) => key.startsWith('zh'))
+    if (major) return major
+
+    return langs[0]
+  }
+
+  const ttml_cleaner = (ttml_text: string, major_lang: string | null): string => {
+    // 如果没有指定主语言，直接返回原文本（或者根据需求返回空）
+    if (major_lang === null) return ttml_text
+
+    /**
+     * 替换逻辑回调函数
+     * @param match 完整匹配到的标签字符串
+     * @param lang 正则中第一个捕获组匹配到的语言代码
+     */
+    const replacer = (match: string, lang: string) => (lang === major_lang ? match : '')
+    const translationRegex = /<translation[^>]+xml:lang="([^"]+)"[^>]*>[\s\S]*?<\/translation>/g
+    const spanRegex = /<span[^>]+xml:lang="([^" ]+)"[^>]*>[\s\S]*?<\/span>/g
+    return ttml_text.replace(translationRegex, replacer).replace(spanRegex, replacer)
+  }
+
+  const context_lang = lang_counter(ttmlContent)
+  const major = lang_filter(context_lang)
+  const cleaned_ttml = ttml_cleaner(ttmlContent, major)
+
+  return cleaned_ttml.replace(/\n\s*/g, '')
+}
+
+/**
+ * 简单的 TTML 解析器 (Wrapper for library)
+ */
+export const parseTTML = (content: string) => {
+  return parseTTMLLib(content)
+}

VoiceHub-main/app/utils/lyric/lyricStripper.ts

@@ -0,0 +1,411 @@
+/**
+ * @fileoverview 元数据行清理器
+ *
+ * 用于清理歌词中开头和结尾的元数据行，例如：
+ *
+ * (歌曲名) - (歌手名)
+ * 词：...
+ * 曲：...
+ * 编曲：...
+ * 和声编唱：...
+ * 人声编辑：...
+ * 混音：...
+ * 母带：...
+ * 监制：...
+ * 出品：...
+ * 真正的歌词行 1
+ * 真正的歌词行 2
+ *
+ * @see https://github.com/apoint123/Unilyric/blob/afd351c54eca7137cf8ee4ea5652d9ee55c20e32/lyrics_helper_rs/src/converter/processors/metadata_stripper.rs
+ */
+
+import type { LyricLine } from '@applemusic-like-lyrics/lyric'
+
+const STRICT_MATCH_SEPARATORS = [
+  ':',
+  '：',
+  ',',
+  '，',
+  '.',
+  '。',
+  '!',
+  '！',
+  '-',
+  '_',
+  '(',
+  '（',
+  '[',
+  '【',
+  '{',
+  '『',
+  '「'
+]
+
+/**
+ * 扫描限制配置
+ */
+export interface ScanLimitConfig {
+  /**
+   * 扫描比例 (0.0 - 1.0)
+   */
+  ratio: number
+  /**
+   * 最小扫描行数
+   */
+  minLines: number
+  /**
+   * 最大扫描行数
+   */
+  maxLines: number
+}
+
+/**
+ * 歌词清理配置
+ */
+export interface StripOptions {
+  /**
+   * 要匹配的关键词列表
+   */
+  keywords: string[]
+  /**
+   * 要匹配的正则字符串列表
+   */
+  regexPatterns: string[]
+  /**
+   * 这里的正则表达式会被视为弱匹配
+   */
+  softMatchRegexes?: string[]
+  /**
+   * 歌曲元数据，用于检查第一行是否为 "歌曲 - 歌手" 格式
+   */
+  matchMetadata?: {
+    title?: string
+    artists?: string[]
+  }
+}
+
+const DEFAULT_HEADER_LIMIT: ScanLimitConfig = {
+  ratio: 0.2,
+  minLines: 20,
+  maxLines: 70
+}
+
+const DEFAULT_FOOTER_LIMIT: ScanLimitConfig = {
+  ratio: 0.2,
+  minLines: 20,
+  maxLines: 50
+}
+
+function calculateScanLimit(config: ScanLimitConfig, totalLines: number): number {
+  const proportional = Math.ceil(totalLines * config.ratio)
+
+  return Math.min(Math.min(Math.max(proportional, config.minLines), config.maxLines), totalLines)
+}
+
+function getLineText(line: LyricLine): string {
+  if (!line || !line.words) return ''
+  return line.words
+    .map((w) => w.word)
+    .join('')
+    .trim()
+}
+
+/**
+ * 清理文本以便检查
+ *
+ * 移除行首尾的括号，例如 (作曲: xxx) -> 作曲: xxx
+ */
+function cleanTextForCheck(text: string): string {
+  let processed = text.trim()
+  const brackets = [
+    ['(', ')'],
+    ['（', '）'],
+    ['【', '】'],
+    ['[', ']'],
+    ['{', '}'],
+    ['『', '』'],
+    ['「', '」']
+  ]
+
+  let changed = true
+  let loopCount = 0
+
+  while (changed && loopCount < 5) {
+    changed = false
+    loopCount++
+
+    for (const [open, close] of brackets) {
+      if (processed.startsWith(open)) {
+        // 类似于这样的格式：(作曲：周杰伦)
+        if (processed.endsWith(close)) {
+          processed = processed.slice(open.length, processed.length - close.length).trim()
+          changed = true
+          break
+        }
+
+        // 类似于这样的格式：(Live) 作曲：周杰伦
+        const closeIdx = processed.indexOf(close)
+        if (closeIdx > -1) {
+          const contentAfter = processed.slice(closeIdx + close.length).trim()
+
+          if (contentAfter.length > 0) {
+            processed = contentAfter
+            changed = true
+            break
+          }
+        }
+      }
+    }
+  }
+  return processed
+}
+
+// 强匹配行：匹配关键词加冒号，或者匹配正则表达式的行
+// 弱匹配行：带有冒号，但不匹配关键词或正则表达式的行。如果夹在强匹配行之间，多半是元数据行但是没有对应的规则。但也有可能是演唱者标识，“男：...”这样的，为了避免误删，如果后面全是弱匹配行，就不删它们
+// 真正的歌词行：既不匹配规则，又没有冒号的行，作为防火墙来阻止对之后行的移除。避免元数据在歌词中间，把中间的歌词也移除了
+
+function isStrictMatch(text: string, keywords: string[], regexes: RegExp[]): boolean {
+  const cleaned = cleanTextForCheck(text)
+
+  // 转小写并移除空格进行匹配
+  const normalizedText = cleaned.toLowerCase().replace(/\s+/g, '')
+
+  for (const kw of keywords) {
+    const normalizedKw = kw.toLowerCase().replace(/\s+/g, '')
+
+    if (normalizedText.startsWith(normalizedKw)) {
+      const remainder = normalizedText.slice(normalizedKw.length)
+
+      // 如果剩余部分为空，说明完全匹配
+      if (remainder.length === 0) {
+        return true
+      }
+
+      // 检查分隔符
+      // 允许的分隔符包括：冒号、逗号、句号、感叹号、连字符、括号等
+      // 注意：normalizedText 已经移除了空格
+      if (STRICT_MATCH_SEPARATORS.includes(remainder.charAt(0))) {
+        return true
+      }
+    }
+  }
+
+  for (const reg of regexes) {
+    if (reg.test(text)) {
+      return true
+    }
+  }
+
+  return false
+}
+
+function looksLikeMetadata(text: string, softRegexes: RegExp[]): boolean {
+  const cleaned = cleanTextForCheck(text)
+  if (
+    cleaned.includes(':') ||
+    cleaned.includes('：') ||
+    // 第一行的 歌曲名 - 歌手名 这样的格式
+    cleaned.includes('-')
+  ) {
+    return true
+  }
+
+  for (const reg of softRegexes) {
+    if (reg.test(text)) {
+      return true
+    }
+  }
+
+  return false
+}
+
+/**
+ * 扫描头部，寻找正文开始的位置
+ */
+function findHeaderCutoff(
+  lines: readonly LyricLine[],
+  startIndex: number,
+  keywords: string[],
+  regexes: RegExp[],
+  softRegexes: RegExp[],
+  limit: number
+): number {
+  let lastValidMetadataIndex = startIndex - 1
+
+  console.groupCollapsed(`[LyricStripper] ⬇️ 开始头部扫描 (Start: ${startIndex}, Limit: ${limit})`)
+
+  for (let i = startIndex; i < limit; i++) {
+    if (i >= lines.length) break
+
+    const text = getLineText(lines[i])
+
+    if (!text) {
+      continue
+    }
+
+    const strict = isStrictMatch(text, keywords, regexes)
+    const weak = looksLikeMetadata(text, softRegexes)
+
+    let status = '❌ NONE'
+    if (strict) status = '✅ STRICT'
+    else if (weak) status = '⚠️ WEAK'
+
+    console.log(`Line [${i}]: "${text}" | Result: ${status}`)
+
+    if (!strict && !weak) {
+      break
+    }
+
+    if (strict) {
+      lastValidMetadataIndex = i
+    }
+  }
+  console.groupEnd()
+
+  return lastValidMetadataIndex + 1
+}
+
+/**
+ * 扫描尾部，寻找正文结束的位置
+ */
+function findFooterCutoff(
+  lines: readonly LyricLine[],
+  startIndex: number,
+  keywords: string[],
+  regexes: RegExp[],
+  softRegexes: RegExp[],
+  limit: number
+): number {
+  if (startIndex >= lines.length) return startIndex
+
+  const scanEnd = Math.max(startIndex, lines.length - limit)
+  let firstValidFooterIndex = lines.length
+
+  console.groupCollapsed(`[LyricStripper] ⬆️ 开始尾部扫描 (Limit: ${limit})`)
+
+  for (let i = lines.length - 1; i >= scanEnd; i--) {
+    const text = getLineText(lines[i])
+
+    if (!text) {
+      continue
+    }
+
+    const strict = isStrictMatch(text, keywords, regexes)
+    const weak = looksLikeMetadata(text, softRegexes)
+
+    let status = '❌ NONE'
+    if (strict) status = '✅ STRICT'
+    else if (weak) status = '⚠️ WEAK'
+
+    console.log(`Line [${i}]: "${text}" | Result: ${status}`)
+
+    if (!strict && !weak) {
+      break
+    }
+
+    if (strict) {
+      firstValidFooterIndex = i
+    }
+  }
+  console.groupEnd()
+
+  return firstValidFooterIndex
+}
+
+/**
+ * 剥离歌词中的元数据行
+ * @param lines 原始歌词行数组
+ * @param options 包含关键词和正则的配置
+ * @returns 清理后的新歌词行数组
+ */
+export function stripLyricMetadata(
+  lines: readonly LyricLine[],
+  options: StripOptions
+): LyricLine[] {
+  if (!lines || lines.length === 0) return []
+
+  let scanStartIndex = 0
+
+  if (options.matchMetadata) {
+    const { title, artists } = options.matchMetadata
+    const firstLineText = getLineText(lines[0])
+
+    if (title && artists && artists.length > 0 && firstLineText) {
+      const lowerText = firstLineText.toLowerCase()
+      const lowerTitle = title.toLowerCase()
+
+      if (lowerText.includes(lowerTitle)) {
+        const hasAnyArtist = artists.some((artist) => lowerText.includes(artist.toLowerCase()))
+
+        if (hasAnyArtist) {
+          console.log(
+            `[LyricStripper] 在第一行匹配到歌曲元数据: "${firstLineText}" (Title: ${title}, Artists: ${artists.join(', ')})`
+          )
+          scanStartIndex = 1
+        }
+      }
+    }
+  }
+
+  if (
+    scanStartIndex === 0 &&
+    (!options.keywords || options.keywords.length === 0) &&
+    (!options.regexPatterns || options.regexPatterns.length === 0) &&
+    (!options.softMatchRegexes || options.softMatchRegexes.length === 0)
+  ) {
+    return [...lines]
+  }
+
+  const regexes: RegExp[] = []
+  if (options.regexPatterns) {
+    options.regexPatterns.forEach((p) => {
+      try {
+        if (p.trim()) {
+          regexes.push(new RegExp(p, 'i')) // 忽略大小写
+        }
+      } catch (e) {
+        console.warn(`[LyricStripper] 无效的正则表达式: ${p}`, e)
+      }
+    })
+  }
+
+  const softRegexes: RegExp[] = []
+  if (options.softMatchRegexes) {
+    options.softMatchRegexes.forEach((p) => {
+      try {
+        if (p.trim()) softRegexes.push(new RegExp(p, 'i'))
+      } catch (e) {
+        console.warn(`[LyricStripper] 无效的正则表达式: ${p}`, e)
+      }
+    })
+  }
+
+  const keywords = options.keywords || []
+  const totalLines = lines.length
+
+  const headerConfig = DEFAULT_HEADER_LIMIT
+  const headerLimit = calculateScanLimit(headerConfig, totalLines)
+
+  const startIdx = findHeaderCutoff(
+    lines,
+    scanStartIndex,
+    keywords,
+    regexes,
+    softRegexes,
+    headerLimit
+  )
+
+  const footerConfig = DEFAULT_FOOTER_LIMIT
+  const footerLimit = calculateScanLimit(footerConfig, totalLines)
+
+  const endIdx = findFooterCutoff(lines, startIdx, keywords, regexes, softRegexes, footerLimit)
+
+  if (startIdx === 0 && endIdx === lines.length) {
+    return [...lines]
+  }
+
+  const newLength = endIdx - startIdx
+  console.log(`[LyricStripper] 清理完成，总行数从 ${totalLines} 变为 ${newLength}`)
+
+  return lines.slice(startIdx, endIdx)
+}

VoiceHub-main/app/utils/lyric/parseLrc.ts

@@ -0,0 +1,69 @@
+import type { LyricLine } from '@applemusic-like-lyrics/lyric'
+
+const TIME_TAG_REGEX = /\[(\d{2}):(\d{2})\.(\d{2,3})\]/g
+const META_TAG_REGEX = /^\[[a-z]+:/i
+
+export const parseLrc = (lrcContent: string): LyricLine[] => {
+  if (!lrcContent) return []
+
+  const lines = lrcContent.split(/\r?\n/)
+  const result: LyricLine[] = []
+
+  for (const line of lines) {
+    const trimmedLine = line.trim()
+    if (!trimmedLine) continue
+
+    // 跳过元数据
+    if (META_TAG_REGEX.test(trimmedLine)) continue
+
+    const matches = [...trimmedLine.matchAll(TIME_TAG_REGEX)]
+    if (matches.length === 0) continue
+
+    // 提取歌词文本（移除所有时间标签）
+    const text = trimmedLine.replace(TIME_TAG_REGEX, '').trim()
+    // 允许空行（作为间奏），但通常我们忽略完全空的行，除非为了显示间隔
+    if (!text) continue
+
+    // 处理一行多个时间标签的情况（重复歌词）
+    for (const match of matches) {
+      const min = parseInt(match[1], 10)
+      const sec = parseInt(match[2], 10)
+      // 补齐毫秒
+      const msStr = match[3].padEnd(3, '0').slice(0, 3)
+      const ms = parseInt(msStr, 10)
+      const startTime = min * 60 * 1000 + sec * 1000 + ms
+
+      result.push({
+        startTime,
+        endTime: 0, // 稍后计算
+        words: [{ word: text, startTime, endTime: 0, romanWord: '' }],
+        translatedLyric: '',
+        romanLyric: '',
+        isBG: false,
+        isDuet: false
+      })
+    }
+  }
+
+  // 按时间排序
+  result.sort((a, b) => a.startTime - b.startTime)
+
+  // 计算结束时间
+  for (let i = 0; i < result.length; i++) {
+    const line = result[i]
+    const nextLine = result[i + 1]
+
+    if (nextLine) {
+      line.endTime = nextLine.startTime
+    } else {
+      line.endTime = line.startTime + 5000 // 最后一行默认 5秒
+    }
+
+    // 更新单词结束时间
+    if (line.words.length > 0) {
+      line.words[0].endTime = line.endTime
+    }
+  }
+
+  return result
+}

VoiceHub-main/app/utils/lyric/qrc-parser.ts

@@ -0,0 +1,64 @@
+/**
+ * QRC 解析
+ */
+
+// 定义函数签名类型
+type QRCParserFn = (xmlStr: string) => string
+
+// 实现 DOM 策略
+const domParser: QRCParserFn = (xmlStr) => {
+  if (!xmlStr || !xmlStr.trim()) return ''
+  try {
+    const parser = new DOMParser()
+    const doc = parser.parseFromString(xmlStr, 'text/xml')
+    // 检查解析错误（如未转义的 & 等）
+    const parseError = doc.querySelector('parsererror')
+    if (parseError) {
+      // DOM 解析失败，回退到正则解析
+      return regexParser(xmlStr)
+    }
+
+    return doc.documentElement.getAttribute('LyricContent') || ''
+  } catch (e) {
+    console.error('[QRC Parser] Fatal Error:', e)
+    // 回退到正则解析
+    return regexParser(xmlStr)
+  }
+}
+
+const decodeXml = (str: string) => {
+  return str
+    .replace(/"/g, '"')
+    .replace(/'/g, "'")
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&amp;/g, '&')
+}
+
+// 实现正则策略
+const regexParser: QRCParserFn = (xmlStr) => {
+  if (!xmlStr) return ''
+
+  // 尝试贪婪匹配 (处理 LyricContent="... "..." ..." 这种非标准 XML 引号未转义的情况)
+  // 匹配从 LyricContent=" 开始，直到标签结束前的最后一个引号
+  const greedyMatch = /LyricContent\s*=\s*"([\s\S]*)"\s*\/?>/.exec(xmlStr)
+  if (greedyMatch) {
+    const content = greedyMatch[1]
+    // 启发式检查：如果提取的内容中包含类似 ` Attribute="` 的结构，说明贪婪匹配吃掉了其他属性
+    // 这种情况下回退到非贪婪匹配
+    if (!/\s+\w+\s*=\s*"/.test(content)) {
+      return decodeXml(content)
+    }
+  }
+
+  // 标准非贪婪匹配 (兼容标准 XML 属性)
+  const match = /LyricContent\s*=\s*"([^"]*)"/.exec(xmlStr)
+
+  // 如果没有匹配到 XML 结构，假设输入本身就是内容 (保持原有逻辑的兼容性)
+  const result = match?.[1] || xmlStr
+  return decodeXml(result)
+}
+
+// 避免每次调用时的运行时检查
+export const extractLyricContent: QRCParserFn =
+  typeof DOMParser !== 'undefined' ? domParser : regexParser