dsh-voice-mode 0.7.14 → 0.7.16

package/README.en.md

@@ -30,7 +30,6 @@ interrupts playback and the running turn. No API key.
 - **Hardening**: session-existence check, loopback + Origin guards, per-endpoint rate limits, model SHA256 pinning, download-host allowlist.
 
 > ℹ️ The `demo-voice-flow.gif` at the top is a **real recording of the current UI** (voice mode → live transcription → pause auto-send → sentence-by-sentence read-aloud with live captions), produced by driving the real pipeline via `screenshots/scripts/capture-demo.mjs`.
-> ⚠️ `assets/demo.gif` and the older screenshots under the repo-root `assets/` still show the **upstream legacy single-button UI**; the current UI adds a mode-switch button next to the mic.
 
 ---
 
@@ -283,7 +282,7 @@ flowchart LR
 | Hotword biasing | None | None (removed in v0.7.7; see version note) |
 | Caption a11y | None | **4 font tiers + 3 width tiers + theme-following** |
 | Wake word | None | **Lightweight streaming match + prefix filler whitelist** |
-| dsh compatibility | — | **0.1.1-rc.2 → 0.1.5-rc.2 full range** |
+| dsh compatibility | — | **full range since 0.1.1-rc.2 (incl. 0.1.7-rc.1)** |
 
 ---
 

package/README.md

@@ -35,7 +35,6 @@ DeepSeek Harness 语音双工对话模式：会话内一键进入 → 边说边
 - **安全加固**：会话存在性校验 / 回环+Origin 校验 / 全端点限流 / **ASR+TTS 全模型 SHA256 固定** / 下载域名白名单 / 重定向守卫。
 
 > ℹ️ 顶部 `demo-voice-flow.gif` 是**当前界面的真实录制**（语音模式 → 边说边出字 → 停顿自动发送 → 按句朗读 + 实时字幕），由 `screenshots/scripts/capture-demo.mjs` 驱动真实链路产出。
-> ⚠️ 仓库内 `assets/demo.gif` 与根目录 `assets/` 下的旧截图仍为**上游旧版界面**（单按钮时期）；当前界面在语音按钮旁多一颗「模式切换」按钮。
 
 ---
 
@@ -248,7 +247,7 @@ flowchart LR
 | 热词偏置 | 无 | 无（已移除，详见 v0.7.7 文档说明） |
 | 字幕 a11y | 无 | **4 档字号 + 3 档宽度 + 主题跟随** |
 | 唤醒词 | 无 | **轻量流式匹配 + 前缀语气词白名单** |
-| 兼容 dsh | — | **0.1.1-rc.2 → 0.1.5-rc.2 全版本（+ 0.1.6-alpha.2 预览）** |
+| 兼容 dsh | — | **0.1.1-rc.2 起全版本（含 0.1.7-rc.1）** |
 
 ---
 

package/lib/client.js

@@ -2815,7 +2815,7 @@ var interruptFirstAt = 0;
 var isSpeechFalseRun = 0;
 var lastReenterAt = 0;
 var INT_CONFIRM_FRAMES = { 0: 3, 1: 2, 2: 1 };
-var inject = ["slots", "sessions", "settingsScope"];
+var inject = ["slots", "sessions"];
 var TELEMETRY_VIEW = [
   { stage: "utterance-end", key: "telUtteranceEnd" },
   { stage: "endpoint-fired", key: "telEndpoint" },
@@ -2825,7 +2825,7 @@ var TELEMETRY_VIEW = [
   { stage: "first-tts-chunk", key: "telFirstChunk" },
   { stage: "first-audio-played", key: "telFirstPlayed" }
 ];
-var BUILD_TAG = "4a95ba9";
+var BUILD_TAG = "37c709b";
 var TELEMETRY_FLAG = "dsh-voice-mode.telemetry";
 var telemetryEnabled = typeof localStorage !== "undefined" && localStorage.getItem(TELEMETRY_FLAG) === "1";
 console.log("[dsh-voice] build=" + BUILD_TAG);
@@ -2950,7 +2950,8 @@ function apply(ctx) {
       VoiceOverlay
     )
   );
-  if (ctx.settingsScope) {
+  const settingsScope = ctx.get?.("settingsScope");
+  if (settingsScope) {
     ctx.slots.inject(
       "settings.plugin.item",
       () => ctx.slots.register(
@@ -2961,7 +2962,7 @@ function apply(ctx) {
           order: 100,
           label: t("stateVoiceMode")
         },
-        () => React.createElement(VoiceSettingsCard, { scope: ctx.settingsScope.bind({ namespace: "voice-mode" }) })
+        () => React.createElement(VoiceSettingsCard, { scope: settingsScope.bind({ namespace: "voice-mode" }) })
       )
     );
   }

package/lib/index.js

@@ -2010,8 +2010,81 @@ var Config = z.object({
   rate: z.number().default(1.1),
   interruptLevel: z.union([z.const(0), z.const(1), z.const(2)]).default(0),
   silenceMs: z.number().default(1500),
-  idleTimeoutMinutes: z.number().default(5)
+  idleTimeoutMinutes: z.number().default(5),
+  // 设置面板字段（与 VoiceSettingsValue 一一对应；0.1.7+ 由插件自身 Config 派生读取）。
+  // 注意：刻意不标 .volatile() —— schemastery 3.18.4 的 volatile 会破坏 schema 函数
+  // 调用形态（sv({}) → {field:{}}），而 0.1.5-rc.3 起的设置分层会调用插件 Config 做
+  // merge，{} 透过 mergeLayers 污染并触发 ValidationError（2026-09-23 实测）。
+  // 0.1.7 功能读取走 config 直接读，不依赖 volatile（仅官方 UI 自动投影受影响，
+  // 本插件自带 settings-form 设置面板 + /voice-mode/config，不受影响）。
+  autoSend: z.boolean().default(true),
+  autoResume: z.boolean().default(false),
+  mode: z.union([z.const("toggle"), z.const("hold")]).default("toggle"),
+  bargeInMode: z.union([z.const("auto"), z.const("manual"), z.const("detect")]).default("detect"),
+  echoGateDb: z.number().default(6),
+  shortcut: z.string().default("Ctrl+Shift+V"),
+  spokenFormat: z.boolean().default(true),
+  senseVoice: z.boolean().default(true),
+  wakeWord: z.string().default(""),
+  toolBeep: z.boolean().default(false),
+  senseITN: z.boolean().default(true),
+  captionFontSize: z.union([z.const(0), z.const(1), z.const(2), z.const(3)]).default(0),
+  captionMaxWidth: z.union([z.const(0), z.const(1), z.const(2)]).default(1),
+  backchannelYield: z.boolean().default(true),
+  yieldMs: z.number().default(1500)
 });
+function voiceSettingsFromConfig(config) {
+  const {
+    ttsEngine,
+    kokoroModel,
+    voice,
+    rate,
+    interruptLevel,
+    silenceMs,
+    idleTimeoutMinutes,
+    modelHost,
+    autoSend,
+    autoResume,
+    mode,
+    bargeInMode,
+    echoGateDb,
+    shortcut,
+    spokenFormat,
+    senseVoice,
+    wakeWord,
+    toolBeep,
+    senseITN,
+    captionFontSize,
+    captionMaxWidth,
+    backchannelYield,
+    yieldMs
+  } = config;
+  return {
+    ttsEngine,
+    kokoroModel,
+    voice,
+    rate,
+    interruptLevel,
+    silenceMs,
+    idleTimeoutMinutes,
+    modelHost,
+    autoSend,
+    autoResume,
+    mode,
+    bargeInMode,
+    echoGateDb,
+    shortcut,
+    spokenFormat,
+    senseVoice,
+    wakeWord,
+    toolBeep,
+    senseITN,
+    captionFontSize,
+    captionMaxWidth,
+    backchannelYield,
+    yieldMs
+  };
+}
 function apply(ctx, config) {
   let activeVoiceSession = null;
   let activeTabId = null;
@@ -2056,22 +2129,32 @@ function apply(ctx, config) {
       }
     }
   };
-  const settingsScope = ctx.settings.register(
-    NS_VOICE_MODE,
-    createVoiceSettingsSchema(),
-    {
-      base: {
-        ttsEngine: config.ttsEngine,
-        voice: config.voice,
-        rate: config.rate,
-        interruptLevel: config.interruptLevel,
-        silenceMs: config.silenceMs,
-        idleTimeoutMinutes: config.idleTimeoutMinutes,
-        modelHost: config.modelHost
+  const legacySettings = ctx.settings;
+  const useLegacySettings = typeof legacySettings?.register === "function";
+  let settingsScopeRef = null;
+  let vset;
+  if (useLegacySettings && legacySettings && typeof legacySettings.register === "function") {
+    const settingsScope = legacySettings.register.call(
+      legacySettings,
+      NS_VOICE_MODE,
+      createVoiceSettingsSchema(),
+      {
+        base: {
+          ttsEngine: config.ttsEngine,
+          voice: config.voice,
+          rate: config.rate,
+          interruptLevel: config.interruptLevel,
+          silenceMs: config.silenceMs,
+          idleTimeoutMinutes: config.idleTimeoutMinutes,
+          modelHost: config.modelHost
+        }
       }
-    }
-  );
-  let vset = settingsScope.get();
+    );
+    vset = settingsScope.get();
+    settingsScopeRef = settingsScope;
+  } else {
+    vset = voiceSettingsFromConfig(config);
+  }
   const asr = createAsrRuntime({
     cacheDir: config.cacheDir,
     modelHost: () => vset.modelHost,
@@ -2118,23 +2201,26 @@ function apply(ctx, config) {
   const unsubscribe = queue.subscribe((frame) => broadcast("audio", frame));
   ctx.effect(() => unsubscribe);
   ctx.effect(() => () => void queue.close());
-  ctx.effect(
-    () => settingsScope.watch((next) => {
-      const prev = vset;
-      vset = next;
-      if (next.ttsEngine !== engineKind) {
-        engineKind = next.ttsEngine;
-        queue.setEngine(makeEngine(engineKind));
-      } else if (engineKind === "kokoro" && next.kokoroModel !== activeKokoroModel) {
-        activeKokoroModel = next.kokoroModel;
-        queue.setEngine(makeEngine("kokoro"));
-      }
-      queue.updateVoice(next.voice, next.rate);
-      if (next.senseITN !== prev.senseITN || next.senseVoice !== prev.senseVoice) {
-        asr.markStale();
-      }
-    })
-  );
+  if (settingsScopeRef) {
+    const scopeRef = settingsScopeRef;
+    ctx.effect(
+      () => scopeRef.watch((next) => {
+        const prev = vset;
+        vset = next;
+        if (next.ttsEngine !== engineKind) {
+          engineKind = next.ttsEngine;
+          queue.setEngine(makeEngine(engineKind));
+        } else if (engineKind === "kokoro" && next.kokoroModel !== activeKokoroModel) {
+          activeKokoroModel = next.kokoroModel;
+          queue.setEngine(makeEngine("kokoro"));
+        }
+        queue.updateVoice(next.voice, next.rate);
+        if (next.senseITN !== prev.senseITN || next.senseVoice !== prev.senseVoice) {
+          asr.markStale();
+        }
+      })
+    );
+  }
   const currentVoice = () => vset.voice;
   const currentRate = () => vset.rate;
   const currentInterrupt = () => vset.interruptLevel;
@@ -2571,7 +2657,8 @@ function apply(ctx, config) {
             res.end(JSON.stringify({ error: "mode must be toggle or hold" }));
             return;
           }
-          void settingsScope.update({ mode }).then(() => {
+          const persistMode = settingsScopeRef ? settingsScopeRef.update({ mode }) : ctx.settings.mutate ? ctx.settings.mutate(NS_VOICE_MODE, [{ op: "set", path: ["mode"], value: mode }]) : Promise.reject(new Error("mode persistence unsupported on this host"));
+          void persistMode.then(() => {
             res.statusCode = 200;
             res.setHeader("content-type", "application/json");
             res.end(JSON.stringify({ ok: true, mode }));

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "dsh-voice-mode",
-  "description": "Full-duplex voice plugin for DeepSeek Harness: local zipformer2 streaming ASR (no API key) → editable draft; Edge TTS or local VITS / Kokoro read-aloud with live captions; true barge-in; hardened HTTP surface + model SHA256 pinning; compatible with all dsh releases since 0.1.1-rc.2 (verified 0.1.1-rc.2 → 0.1.5-rc.2, plus 0.1.6-alpha.2 preview). · DSH 语音双工插件：流式识别入草稿、按句朗读+实时字幕、开口即打断；免 API Key、安全加固、全版本兼容（0.1.1-rc.2 起，含 0.1.5-rc.2 端到端验证 + 0.1.6-alpha.2 预览核验）。",
-  "version": "0.7.14",
+  "description": "Full-duplex voice plugin for DeepSeek Harness: local zipformer2 streaming ASR (no API key) → editable draft; Edge TTS or local VITS / Kokoro read-aloud with live captions; true barge-in; hardened HTTP surface + model SHA256 pinning; compatible with all dsh releases since 0.1.1-rc.2 (verified 0.1.1-rc.2 → 0.2.0-rc.2). · DSH 语音双工插件：流式识别入草稿、按句朗读+实时字幕、开口即打断；免 API Key、安全加固、全版本兼容（0.1.1-rc.2 起，已验证至 0.2.0-rc.2 端到端）。",
+  "version": "0.7.16",
   "type": "module",
   "main": "lib/index.js",
   "repository": {
@@ -29,7 +29,6 @@
     "lib/sense-worker.mjs",
     "cordis.patch.yml",
     "assets/architecture.svg",
-    "assets/demo.gif",
     "scripts/prefetch.mjs",
     "scripts/list-voices.mjs",
     "scripts/bench-asr.mjs",
@@ -79,7 +78,7 @@
     }
   },
   "dependencies": {
-    "@deepseek-ai/schemastery": "^3.18.1",
+    "@deepseek-ai/schemastery": "^3.18.4",
     "msedge-tts": "^2.0.7",
     "sherpa-onnx": "^1.13.6",
     "sherpa-onnx-node": "1.13.6"