dsh-live2dwaifu 1.0.6 → 1.1.0

package/README.md

@@ -145,6 +145,40 @@ dsh plugin --profile desktop add dsh-live2dwaifu
 
 ---
 
+## 喂她语料
+
+人设卡告诉她「是谁」，语料告诉她「怎么说话」。⚙ 抽屉里的 **语料** 按钮打开一个页面，有模型选择、大文本框、文件按钮和拖拽区。
+
+**四种喂法，怎么方便怎么来：**
+
+1. 直接粘贴——打字、贴聊天记录、贴剧本都行
+2. 点「选文件」，`.txt` / `.json` / `.jsonl` / `.csv`，可以多选
+3. 把文件拖到那个虚框里
+4. 一行一句地贴她的台词，什么都不用标
+
+**认得的写法**：
+
+```
+他：今天吃啥           ← 认得出谁说的
+你：锅里还有，热的。
+
+「嗯。我也是。」        ← 只认她是说话人
+
+- 先坐下。水在桌上。    ← 一行一句台词
+
+[{"user":"累吗","assistant":"还行。"}]   ← JSON 也行
+```
+
+说话人写 `他 / 我 / 用户` 算你，`你 / 她 / 助手` 算她；认不出来的一律当她说的——**只贴台词最省事**。
+
+页面右边会实时显示解析出来的条数，和前面十几条预览。贴错了能看见，改完再存。
+
+**存法有两种**：「加进去」是并到原来的里面，「替换成这些」是把之前清掉。第一次喂就用加。
+
+以后她说话会从语料里挑几条贴进去当样例——优先挑跟你这句话沾边的问答，另外随机抽几句她的台词找语感。所以**语料越多、越像她本人，她说得就越像**。
+
+存的位置：`~/.dsh/dsh-live2d/models/<id>/corpus.json`。
+
 ## 声音
 
 她说话时一个字一声很短的电音，标点停一拍，句尾沉一下。不是朗读，是“哔哔”那种。

package/assets/corpus.html

@@ -0,0 +1,290 @@
+<!DOCTYPE html>
+<html lang="zh">
+<head>
+<meta charset="utf-8" />
+<title>语料 — dsh-live2dwaifu</title>
+<style>
+  :root { --bg:#12141a; --card:#15181e; --line:#2c313a; --txt:#dfe4ec; --dim:#7d858f; --acc:#6aa9ff; }
+  * { box-sizing: border-box; }
+  body { margin:0; min-height:100vh; background:radial-gradient(1100px 560px at 50% -10%, #1b1f27, var(--bg));
+    color:var(--txt); font:13px/1.6 "Microsoft YaHei", system-ui, sans-serif; padding:22px 24px 28px; }
+  h1 { font-size:15px; margin:0 0 4px; font-weight:600; }
+  p.sub { margin:0 0 14px; color:var(--dim); font-size:12px; }
+  .row { display:flex; gap:12px; align-items:center; margin-bottom:12px; flex-wrap:wrap; }
+  select, button { font:inherit; }
+  select { padding:8px 10px; border-radius:9px; border:1px solid #3c424c; background:#181c22; color:var(--txt); font-size:13px; }
+  .btns { display:flex; gap:8px; flex-wrap:wrap; }
+  button.act { padding:8px 14px; border-radius:9px; border:1px solid #3c424c; background:#232830;
+    color:#dfe4ec; cursor:pointer; font-size:12px; }
+  button.act.primary { background:#27405f; border-color:#3e6b9e; color:#eaf2ff; }
+  button.act:hover { filter:brightness(1.14); }
+  .cols { display:flex; gap:14px; align-items:flex-start; flex-wrap:wrap; }
+  .col { flex:1 1 380px; min-width:320px; }
+  .box { border:1px solid var(--line); border-radius:12px; background:var(--card); padding:12px; }
+  .box h2 { font-size:11px; letter-spacing:.6px; color:var(--dim); margin:0 0 8px; font-weight:600; }
+  textarea { width:100%; height:220px; resize:vertical; padding:10px 12px; border-radius:10px;
+    border:1px solid #3c424c; background:#181c22; color:var(--txt);
+    font:12.5px/1.75 "Microsoft YaHei", system-ui, sans-serif; outline:none; }
+  textarea:focus { border-color:#3f4a5c; }
+  #drop { margin-top:10px; border:1px dashed #3c424c; border-radius:10px; padding:14px; text-align:center;
+    color:var(--dim); font-size:12px; cursor:pointer; transition:all .15s; }
+  #drop.hot { border-color:var(--acc); color:#cfe2ff; background:rgba(106,169,255,.06); }
+  .stat { font-size:12px; color:var(--dim); margin-bottom:6px; }
+  .stat b { color:#e2e8f1; font-weight:600; }
+  #preview { max-height:260px; overflow-y:auto; font-size:12px; line-height:1.75; }
+  #preview div { padding:3px 0; border-bottom:1px solid rgba(255,255,255,.04); }
+  #preview .who { color:#6f7783; }
+  .tips { color:#6f7783; font-size:11.5px; line-height:1.8; margin-top:10px; }
+  code { background:rgba(255,255,255,.06); border-radius:4px; padding:1px 5px; font-size:11px; }
+  .state { color:var(--dim); font-size:11.5px; margin-left:auto; }
+</style>
+</head>
+<body>
+
+<h1>语料</h1>
+<p class="sub">喂她一些她说过的话，她就照着这个调子说。不用排版整齐，能读懂就行。</p>
+
+<div class="row">
+  <select id="models"></select>
+  <div class="btns">
+    <button class="act" id="pick">选文件</button>
+    <button class="act primary" id="append">加进去</button>
+    <button class="act" id="replace">替换成这些</button>
+  </div>
+  <span class="state" id="state"></span>
+</div>
+
+<div class="cols">
+  <div class="col box">
+    <h2>把话贴进来</h2>
+    <textarea id="raw" spellcheck="false" placeholder="随便贴，几种写法都认：
+
+他：今天吃啥
+你：锅里还有，热的。
+
+或者只贴她的台词，一行一句：
+锅里还有，热的。
+先坐下。水在桌上。
+
+也支持她说过的话带引号：「嗯。我也是。」"></textarea>
+    <div id="drop">把 .txt / .json / .jsonl 拖到这里，或者点一下选文件（可以多选）</div>
+    <input id="files" type="file" multiple accept=".txt,.json,.jsonl,.csv" style="display:none" />
+    <div class="tips">
+      认得的写法：<code>他：… / 你：…</code>、<code>「…」</code>、<code>- …</code>、一行一句、
+      JSON 数组 <code>[{"user":"…","assistant":"…"}]</code>。<br />
+      说话人写对才分得清谁说的：他/我/用户 算你，你/她/助手 算她；认不出的一律当她说的。
+    </div>
+  </div>
+
+  <div class="col box">
+    <h2>解析出来长这样</h2>
+    <div class="stat" id="stat">还没有内容</div>
+    <div id="preview"></div>
+    <div class="tips" id="saved"></div>
+  </div>
+</div>
+
+<script>
+(function () {
+  var $ = function (id) { return document.getElementById(id); };
+  var modelId = '';
+  var parsed = { lines: [], pairs: [] };
+  var stored = { lines: [], pairs: [] };
+  var files = [];
+
+  var USERS = /^(他|我|用户|user|User|USER|问|Q)$/;
+  var HERS = /^(你|她|它|助手|bot|Bot|BOT|答|A|B)$/;
+  var SPEAK = /^\s*([^\s：:]{1,6})\s*[：:]\s*(.+)$/;
+
+  function unquote(s) {
+    var m = String(s).trim().match(/^[「“"'']([\s\S]+)[」”"'']$/);
+    return m ? m[1].trim() : String(s).trim();
+  }
+
+  function parse(raw) {
+    var lines = [], pairs = [];
+    var pending = null;
+    var rows = String(raw || '').split(/\r?\n/);
+    for (var i = 0; i < rows.length; i += 1) {
+      var line = rows[i].trim();
+      if (!line) continue;
+
+      if (line.charAt(0) === '{' || line.charAt(0) === '[') {
+        try {
+          var o = JSON.parse(line);
+          var arr = Array.isArray(o) ? o : [o];
+          var used = false;
+          for (var k = 0; k < arr.length; k += 1) {
+            var it = arr[k] || {};
+            var a = it.assistant || it.a || it.answer || it.reply || it.response || it.text || '';
+            var u = it.user || it.q || it.question || it.prompt || it.input || '';
+            if (a) {
+              if (u) pairs.push({ user: String(u), assistant: String(a) });
+              else lines.push(String(a));
+              used = true;
+            }
+          }
+          if (used) continue;
+        } catch (e) { /* 不是 JSON，按普通行处理 */ }
+      }
+
+      line = line.replace(/^[-*•·]\s*/, '');
+      var m = line.match(SPEAK);
+      if (m) {
+        var who = m[1].trim();
+        var said = unquote(m[2]);
+        if (!said) continue;
+        if (USERS.test(who)) { pending = said; continue; }
+        if (HERS.test(who) || true) {
+          pairs.push({ user: pending || '', assistant: said });
+          pending = null;
+          continue;
+        }
+      }
+      var t = unquote(line);
+      if (t) lines.push(t);
+    }
+    return { lines: lines, pairs: pairs };
+  }
+
+  function render() {
+    var total = parsed.lines.length + parsed.pairs.length;
+    $('stat').innerHTML = total
+      ? '读到 <b>' + parsed.pairs.length + '</b> 组问答、<b>' + parsed.lines.length + '</b> 句台词'
+      : '还没有内容';
+    var box = $('preview');
+    box.innerHTML = '';
+    var show = [];
+    for (var i = 0; i < parsed.pairs.length && show.length < 8; i += 1) {
+      show.push('他：' + parsed.pairs[i].user + '<br>你：' + parsed.pairs[i].assistant);
+    }
+    for (var j = 0; j < parsed.lines.length && show.length < 14; j += 1) {
+      show.push('你：' + parsed.lines[j]);
+    }
+    show.forEach(function (html) {
+      var d = document.createElement('div');
+      d.innerHTML = html.replace(/</g, '&lt;').replace(/&lt;br&gt;/g, '<br>');
+      box.appendChild(d);
+    });
+    var kept = stored.lines.length + stored.pairs.length;
+    $('saved').textContent = kept ? '已经存着 ' + kept + ' 条（' + stored.pairs.length + ' 组问答、' + stored.lines.length + ' 句台词）' : '';
+  }
+
+  function reparse() {
+    var all = $('raw').value;
+    for (var i = 0; i < files.length; i += 1) all += '\n' + files[i];
+    parsed = parse(all);
+    render();
+  }
+
+  function load() {
+    if (!modelId) return;
+    $('state').textContent = '读…';
+    fetch('/dsh-live2d/api/corpus?id=' + encodeURIComponent(modelId))
+      .then(function (r) { return r.json(); })
+      .then(function (d) {
+        stored = { lines: d.lines || [], pairs: d.pairs || [] };
+        $('state').textContent = '读好了';
+        render();
+      })
+      .catch(function (e) { $('state').textContent = '读不了：' + e.message; });
+  }
+
+  function save(mode) {
+    if (!modelId) return;
+    var payload;
+    if (mode === 'replace') {
+      payload = parsed;
+    } else {
+      payload = {
+        lines: stored.lines.concat(parsed.lines),
+        pairs: stored.pairs.concat(parsed.pairs)
+      };
+    }
+    if (!payload.lines.length && !payload.pairs.length) {
+      $('state').textContent = '没有可存的内容';
+      return;
+    }
+    $('state').textContent = '存…';
+    fetch('/dsh-live2d/api/corpus', {
+      method: 'POST',
+      headers: { 'content-type': 'application/json' },
+      body: JSON.stringify({ id: modelId, lines: payload.lines, pairs: payload.pairs })
+    })
+      .then(function (r) { return r.json(); })
+      .then(function (d) {
+        if (!d.ok) { $('state').textContent = '没存上：' + (d.error || '?'); return; }
+        $('state').textContent = '存好了，共 ' + d.total + ' 条';
+        $('raw').value = '';
+        files = [];
+        load();
+      })
+      .catch(function (e) { $('state').textContent = '没存上：' + e.message; });
+  }
+
+  $('raw').addEventListener('input', reparse);
+  $('pick').addEventListener('click', function () { $('files').click(); });
+  $('drop').addEventListener('click', function () { $('files').click(); });
+  $('append').addEventListener('click', function () { save('append'); });
+  $('replace').addEventListener('click', function () { save('replace'); });
+
+  $('files').addEventListener('change', function (e) {
+    var list = Array.prototype.slice.call(e.target.files || []);
+    readFiles(list);
+  });
+
+  function readFiles(list) {
+    if (!list.length) return;
+    $('state').textContent = '读文件…';
+    var left = list.length;
+    list.forEach(function (f) {
+      var reader = new FileReader();
+      reader.onload = function () {
+        files.push(String(reader.result || ''));
+        left -= 1;
+        if (left === 0) { $('state').textContent = '读好了 ' + files.length + ' 个文件'; reparse(); }
+      };
+      reader.onerror = function () { left -= 1; if (left === 0) reparse(); };
+      reader.readAsText(f, 'utf-8');
+    });
+  }
+
+  var drop = $('drop');
+  ['dragenter', 'dragover'].forEach(function (ev) {
+    drop.addEventListener(ev, function (e) { e.preventDefault(); drop.classList.add('hot'); });
+  });
+  ['dragleave', 'drop'].forEach(function (ev) {
+    drop.addEventListener(ev, function (e) { e.preventDefault(); drop.classList.remove('hot'); });
+  });
+  drop.addEventListener('drop', function (e) {
+    var dt = e.dataTransfer;
+    if (dt && dt.files && dt.files.length) readFiles(Array.prototype.slice.call(dt.files));
+  });
+
+  fetch('/dsh-live2d/api/models')
+    .then(function (r) { return r.json(); })
+    .then(function (d) {
+      var list = d.models || [];
+      var sel = $('models');
+      sel.innerHTML = '';
+      list.forEach(function (m) {
+        var o = document.createElement('option');
+        o.value = m.id;
+        o.textContent = (m.name || m.id) + (m.id === d.selected ? '（正在显示）' : '');
+        sel.appendChild(o);
+      });
+      modelId = d.selected || (list[0] && list[0].id) || '';
+      if (modelId) sel.value = modelId;
+      if (!list.length) { $('state').textContent = '还没有模型'; return; }
+      load();
+    })
+    .catch(function (e) { $('state').textContent = '读不了模型表：' + e.message; });
+
+  $('models').addEventListener('change', function (e) {
+    modelId = e.target.value;
+    load();
+  });
+})();
+</script>
+</body>
+</html>

package/assets/viewer.html

@@ -78,6 +78,7 @@
   <div id="settings"></div>
   <div id="actions">
     <button id="import">导入模型</button>
+    <button id="corpus">语料</button>
     <button id="close">关掉</button>
   </div>
   <div id="tip">选整个文件夹，浏览器会自己把里面所有文件读进来。</div>

package/assets/viewer.js

@@ -44693,6 +44693,15 @@ void main(void)\r
   });
   window.__live2d = { say, mood: applyMood, refresh };
   $2("import").onclick = () => $2("files").click();
+  var corpusBtn = $2("corpus");
+  if (corpusBtn) {
+    corpusBtn.onclick = () => {
+      try {
+        window.parent.postMessage({ type: "dsh-live2d/open", sheet: "corpus" }, "*");
+      } catch {
+      }
+    };
+  }
   $2("close").onclick = () => $2("drawer").classList.remove("show");
   $2("files").onchange = (e2) => importFolder(e2.target.files);
   function moodValue(id) {

package/lib/client.js

@@ -8,6 +8,7 @@
     const BASE = '/dsh-live2d';
     const BGS = ['none', 'dark', 'light', 'paper'];
     const BG_LABEL = { none: '透明', dark: '深色', light: '浅色', paper: '纸质' };
+    const SHEET_TITLE = { tone: '音色', persona: '人设卡', corpus: '语料' };
 
     const CSS = `
       .dsh-l2d-card{position:fixed;right:18px;bottom:18px;z-index:9000;border-radius:16px;overflow:hidden;
@@ -96,6 +97,14 @@
       const [draft, setDraft] = react.useState('');
       const [busy, setBusy] = react.useState(false);
       const [sheet, setSheet] = react.useState(null);
+      react.useEffect(() => {
+        const onMsg = (e) => {
+          const d = e.data || {};
+          if (d.type === 'dsh-live2d/open' && d.sheet) setSheet(String(d.sheet));
+        };
+        window.addEventListener('message', onMsg);
+        return () => window.removeEventListener('message', onMsg);
+      }, []);
       const [nonce, setNonce] = react.useState(0);
       const ref = react.useRef(null);
 
@@ -255,7 +264,7 @@
             h(
               'div',
               { className: 'dsh-l2d-sheet-bar' },
-              h('span', null, sheet === 'tone' ? '音色' : '人设卡'),
+              h('span', null, SHEET_TITLE[sheet] || sheet),
               h('button', { onClick: () => setSheet(null) }, '关掉'),
             ),
             h('iframe', { className: 'dsh-l2d-sheet-body', src: BASE + '/' + sheet }),

package/lib/index.js

@@ -173,6 +173,57 @@ const TALK_SAMPLES = [
   '他：明天要早起 —— 你：那我今晚早点关灯。',
 ].join('\n');
 
+/** 语料：lines = 她的台词，pairs = 一问一答 */
+function readCorpus(id) {
+  try {
+    const d = JSON.parse(readFileSync(join(MODELS, id, 'corpus.json'), 'utf8'));
+    return {
+      lines: Array.isArray(d.lines) ? d.lines.filter((x) => typeof x === 'string' && x.trim()).slice(0, 500) : [],
+      pairs: Array.isArray(d.pairs)
+        ? d.pairs
+            .filter((p) => p && typeof p.assistant === 'string' && p.assistant.trim())
+            .slice(0, 500)
+            .map((p) => ({ user: String(p.user || '').slice(0, 200), assistant: String(p.assistant).slice(0, 400) }))
+        : [],
+    };
+  } catch {
+    return { lines: [], pairs: [] };
+  }
+}
+
+function writeCorpus(id, corpus) {
+  try {
+    writeFileSync(
+      join(MODELS, id, 'corpus.json'),
+      JSON.stringify({ lines: corpus.lines.slice(0, 500), pairs: corpus.pairs.slice(0, 500) }, null, 2),
+      'utf8',
+    );
+  } catch {}
+}
+
+/** 从语料里挑几条最贴当前对话的，拼成样例 */
+function pickCorpusSamples(corpus, text, toneN = 4) {
+  const out = [];
+  const chars = new Set(String(text || ''));
+  const scored = (corpus.pairs || [])
+    .map((p) => {
+      let hit = 0;
+      for (const ch of String(p.user || '')) if (chars.has(ch)) hit += 1;
+      return { p, hit };
+    })
+    .sort((a, b) => b.hit - a.hit);
+  for (const s of scored.slice(0, 3)) {
+    if (s.p.user) out.push('他：' + s.p.user + ' —— 你：' + s.p.assistant);
+    else out.push('你：' + s.p.assistant);
+  }
+  const pool = (corpus.lines || []).slice();
+  for (let i = 0; i < toneN && pool.length; i += 1) {
+    const k = Math.floor(Math.random() * pool.length);
+    out.push('你：' + pool.splice(k, 1)[0]);
+  }
+  return out;
+}
+
 function personaOf(meta) {
   const p = typeof meta?.persona === 'string' ? meta.persona.trim() : '';
   if (p) return p;
@@ -587,6 +638,7 @@ async function handle(req, res, ctx) {
       } catch {}
     }
     const facts = st.memory ? pickFacts(store.facts, text) : [];
+    const samples = st.memory ? pickCorpusSamples(readCorpus(id), text) : [];
     const hist = store.messages;
     const lines = hist
       .slice(-8)
@@ -720,6 +772,11 @@ async function handle(req, res, ctx) {
     return sendFile(res, join(ASSETS, 'mobile.html'), MIME['.html']);
   }
 
+  // 语料页
+  if (path === '/corpus' || path === '/corpus.html') {
+    return sendFile(res, join(ASSETS, 'corpus.html'), MIME['.html']);
+  }
+
   // 人设卡页
   if (path === '/persona' || path === '/persona.html') {
     return sendFile(res, join(ASSETS, 'persona.html'), MIME['.html']);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "dsh-live2dwaifu",
-  "version": "1.0.6",
+  "version": "1.1.0",
   "description": "给 DSH 网页右下角请一个 Live2D 角色：会眨眼、会呼吸、视线跟着鼠标、点她有反应、说话时嘴动、心情变了眉毛跟着动。模型不用内置——谁装谁放自己的文件夹；支持自定义 .exp3.json 表情和 .motion3.json 动作，卡片能拖能缩，背景可换。",
   "type": "module",
   "main": "lib/index.js",