@goodandready/dsh-voice 0.8.32 → 0.8.34

package/README.md

@@ -91,7 +91,6 @@ graph LR
 * 🗣️ **Hands-free Voice Actions** *(0.8.32)*: Spoken commands ("send", "cancel", "clear", "new line") trigger UI actions directly.
 * 🧩 **Structured Prompt Voice Selection** *(0.8.32)*: Spoken options automatically select buttons in interactive assistant choice prompts.
 * 📖 **Developer Lexicon & IT Jargon Correction** *(0.8.32)*: Phonetic normalization for developer slang (GitHub, Docker, Kubernetes, pnpm) with customizable settings dictionary.
-* 🌐 **Realtime Audio Streaming** *(0.8.11)*: Low-latency WebSocket bridge (`/dsh-voice/realtime`) for OpenAI Realtime API or local Sherpa-ONNX streaming. API keys stay securely on the host.
 * 🌊 **Liquid Wave & Dynamic Orb Visualizer** *(0.8.12)*: Smooth animated audio visualization in the recording pill with real-time mic volume reactivity. Switch between organic multi-layer liquid waves, pulsating radiant orb, classic bars, or off.
 
 ---
@@ -217,8 +216,7 @@ Registers `transcribe_audio(file_path, language?)` in `ctx.tools`, allowing agen
 ### Internal HTTP Endpoints
 * `POST /dsh-voice/transcribe` — `{ dataBase64, mimeType, mode }` → `{ ok, text, provider, tookMs }`
 * `POST /dsh-voice/polish` — `{ text }` → `{ ok, text }`
-* `GET /dsh-voice/status` — Returns daemon status, active chains, SenseVoice and realtime config.
-* `GET /dsh-voice/realtime` — **WebSocket upgrade** for low-latency audio streaming (OpenAI Realtime API / Sherpa-ONNX). Accepts binary audio chunks, returns JSON text deltas.
+* `GET /dsh-voice/status` — Returns daemon status, active chains, SenseVoice.
 
 ---
 

package/README.ru.md

@@ -91,7 +91,6 @@ graph LR
 * 🗣️ **Голосовые команды управления (Hands-free Actions)** *(0.8.32)*: фразы «отправь», «отмена», «очисти», «с новой строки» выполняют действия напрямую, не попадая в текст сообщения.
 * 🧩 **Голосовой выбор в интерактивах (Structured Prompts)** *(0.8.32)*: произнесение вариантов ответов нажимает соответствующие кнопки в активных модальных окнах ассистента.
 * 📖 **Словарь IT-жаргона и сленга разработчиков** *(0.8.32)*: фонетическое исправление сленга (гитхаб -> GitHub, кубер -> Kubernetes, пнпм -> pnpm) с возможностью добавления своих правил в настройках.
-* 🌐 **Потоковое аудио в реальном времени** *(0.8.11)*: низколатентный WebSocket-мост (`/dsh-voice/realtime`) для OpenAI Realtime API или локального Sherpa-ONNX. API-ключи остаются на хосте.
 * 🌊 **Анимированные визуализаторы Liquid Wave и Dynamic Orb** *(0.8.12)*: живая интерактивная анимация звуковой волны в полоске записи с откликом на громкость микрофона. Выбор стиля: текучая волна, пульсирующая сфера, классические столбики или выключен.
 
 ---
@@ -218,7 +217,6 @@ dsh plugin --profile web add @goodandready/dsh-voice
 * `POST /dsh-voice/transcribe` — `{ dataBase64, mimeType, mode }` → `{ ok, text, provider, tookMs }`
 * `POST /dsh-voice/polish` — `{ text }` → `{ ok, text }`
 * `GET /dsh-voice/status` — состояние демонов, цепочки, конфигурация SenseVoice и реалтайма.
-* `GET /dsh-voice/realtime` — **WebSocket upgrade** для низколатентного аудио-стриминга (OpenAI Realtime API / Sherpa-ONNX). Принимает бинарные аудио-чанки, возвращает JSON-дельты текста.
 
 ---
 

package/README.zh.md

@@ -85,7 +85,6 @@ graph LR
 * 🔒 **API 密钥安全隔离**：密钥由服务端 `ctx.credentials` 统一解析，绝不向浏览器前端泄漏。
 * 🖥️ **本地 whisper.cpp 服务端直连**：自动管理 `whisper-server` 进程，结合 `ffmpeg` 实现实时音频转码。
 * ⚡ **SenseVoice-ONNX / Sherpa-ONNX** *(0.8.11)*：超快速（~50–100ms）非自回归本地语音识别引擎，自动清除情绪/事件标签。支持 Sherpa-ONNX HTTP 和 OpenAI 兼容端点。
-* 🌐 **实时音频流** *(0.8.11)*：低延迟 WebSocket 桥接（`/dsh-voice/realtime`），支持 OpenAI Realtime API 或本地 Sherpa-ONNX 流式传输。API 密钥安全保留在服务端。
 * 🌊 **Liquid Wave 与 Dynamic Orb 动态可视化** *(0.8.12)*：录音面板内的生动波形动态可视化，实时响应麦克风音量。可选平滑流动波浪、发光脉冲球体、经典频条或关闭。
 
 ---

package/lib/client.js

@@ -521,7 +521,7 @@ window.__ModuleLoader__.load({
         window.dispatchEvent(new CustomEvent('dsh:tts:stop', { bubbles: true }))
         const audios = document.querySelectorAll('audio')
         for (const a of audios) {
-          if (!a.paused) { try { a.pause() } catch {} }
+          if (!a.paused) { try { a.pause() } catch (err) { /* best effort audio pause */ } }
         }
       } catch { /* best effort */ }
       voice.set({ isTtsSpeaking: false })
@@ -3109,7 +3109,23 @@ window.__ModuleLoader__.load({
     }
 
     function registerSettings(ctx) {
-      // Row seat first (the seat the current core renders), legacy card after it.
+      // List seat (plugins.item): the seat the Plugins page renders as the plugin's
+      // own page with its configuration — the host draws the title, icon and crumb
+      // and asks for view 'summary' (the card's one-liner) or view 'page' (the form).
+      // The label is a static string on purpose: it is resolved while the page
+      // renders, and a locale lookup there would take the whole client batch down.
+      ctx.slots.inject('plugins.item', () => ctx.slots.register(
+        {
+          name: 'plugins.item',
+          id: ROW_ID,
+          order: 60,
+          label: () => 'Voice input',
+          locale: NS,
+          inject: () => ({ ctx: ctx }),
+        },
+        PluginCard,
+      ))
+      // Row seat and the legacy card stay as fallbacks.
       ctx.slots.inject('plugins.row.config', () => ctx.slots.register(
         {
           name: 'plugins.row.config',

package/lib/index.js

@@ -213,16 +213,15 @@ export function apply(ctx, baseConfig) {
   }
 
   async function whisperAlive() {
+    const rawUrl = String(live().whisperUrl || '').trim()
+    const base = rawUrl.includes('/inference') ? rawUrl.split('/inference')[0] : rawUrl.replace(/\/+$/, '')
+    if (!base) return false
     try {
-      const controller = new AbortController()
-      const t = setTimeout(() => controller.abort(), 2000)
-      const rawUrl = String(live().whisperUrl || '').trim()
-      const base = rawUrl.includes('/inference') ? rawUrl.split('/inference')[0] : rawUrl.replace(/\/+$/, '')
-      if (!base) return false
-      const res = await fetch(base + '/', { signal: controller.signal })
-      clearTimeout(t)
+      const res = await fetch(base + '/', { signal: AbortSignal.timeout(2000) })
       return res.ok
-    } catch { return false }
+    } catch {
+      return false
+    }
   }
 
   let whisperError = null
@@ -270,16 +269,15 @@ export function apply(ctx, baseConfig) {
   let sensevoiceError = null
 
   async function sensevoiceAlive() {
+    const cfg = live()
+    const base = String(cfg.sensevoiceUrl || '').replace(/\/api\/.*$/, '').replace(/\/v1\/.*$/, '').replace(/\/+$/, '')
+    if (!base) return false
     try {
-      const cfg = live()
-      const base = String(cfg.sensevoiceUrl || '').replace(/\/api\/.*$/, '').replace(/\/v1\/.*$/, '').replace(/\/+$/, '')
-      if (!base) return false
-      const controller = new AbortController()
-      const t = setTimeout(() => controller.abort(), 2000)
-      const res = await fetch(base + '/', { signal: controller.signal })
-      clearTimeout(t)
+      const res = await fetch(base + '/', { signal: AbortSignal.timeout(2000) })
       return res.ok || res.status === 404 // sherpa-onnx answers 404 on / but is alive
-    } catch { return false }
+    } catch {
+      return false
+    }
   }
 
   let startingSensevoice = false

package/lib/normalize.js

@@ -87,15 +87,38 @@ export const DEFAULT_JARGON_DICTIONARY = [
   { from: 'ворктри', to: 'worktree' },
 ]
 
+function jargonRegex(from) {
+    const escaped = from.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')
+  return new RegExp(`(?<![a-zA-Zа-яА-ЯёЁ0-9_-])${escaped}(?![a-zA-Zа-яА-ЯёЁ0-9_-])`, 'gi')
+}
+
+const DEFAULT_JARGON_REGEX = DEFAULT_JARGON_DICTIONARY
+  .filter((item) => item && item.from && item.to)
+  .map((item) => ({ re: jargonRegex(item.from), to: item.to }))
+
+const customJargonRegex = new Map()
+
+function cachedJargonRegex(from) {
+  let re = customJargonRegex.get(from)
+  if (!re) {
+    re = jargonRegex(from)
+    customJargonRegex.set(from, re)
+  }
+  return re
+}
+
 export function applyJargonDictionary(text, customReplacements = []) {
   if (!text || typeof text !== 'string') return text || ''
-  const dict = [...(customReplacements || []), ...DEFAULT_JARGON_DICTIONARY]
   let result = text
-  for (const item of dict) {
+  for (const item of customReplacements || []) {
     if (!item?.from || !item?.to) continue
-    const escaped = item.from.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')
-    const regex = new RegExp(`(?<![a-zA-Zа-яА-ЯёЁ0-9_-])${escaped}(?![a-zA-Zа-яА-ЯёЁ0-9_-])`, 'gi')
-    result = result.replace(regex, item.to)
+    const re = cachedJargonRegex(item.from)
+    re.lastIndex = 0
+    result = result.replace(re, item.to)
+  }
+  for (const item of DEFAULT_JARGON_REGEX) {
+    item.re.lastIndex = 0
+    result = result.replace(item.re, item.to)
   }
   return result
 }

package/lib/sensevoice-installer.js

@@ -6,6 +6,8 @@ import { mkdir, stat } from 'node:fs/promises'
 import { homedir } from 'node:os'
 import path from 'node:path'
 import { spawn } from 'node:child_process'
+import { Readable, Transform } from 'node:stream'
+import { pipeline } from 'node:stream/promises'
 import { writeJson, isTrustedCaller } from './http-util.js'
 
 export const SENSEVOICE_MODEL_URL =
@@ -78,29 +80,21 @@ export async function getSensevoiceStatus(cfg, isAliveFn) {
 export async function downloadArchive(url, destPath, onProgress) {
   const res = await fetch(url, { signal: AbortSignal.timeout(600_000) })
   if (!res.ok) throw new Error(`Download failed with HTTP ${res.status}`)
+  if (!res.body) throw new Error('Download failed with an empty body')
 
   const totalBytes = Number(res.headers.get('content-length')) || 150_000_000
   let receivedBytes = 0
-
-  const fileStream = createWriteStream(destPath)
-  const reader = res.body.getReader()
-
-  while (true) {
-    const { done, value } = await reader.read()
-    if (done) break
-    fileStream.write(value)
-    receivedBytes += value.length
-    if (typeof onProgress === 'function') {
-      const pct = Math.min(95, Math.round((receivedBytes / totalBytes) * 90))
-      onProgress(pct, `downloading (${Math.round(receivedBytes / 1024 / 1024)} MB)`)
-    }
-  }
-
-  await new Promise((resolve, reject) => {
-    fileStream.end()
-    fileStream.on('finish', resolve)
-    fileStream.on('error', reject)
+  const progress = new Transform({
+    transform(chunk, _enc, cb) {
+      receivedBytes += chunk.length
+      if (typeof onProgress === 'function') {
+        const pct = Math.min(95, Math.round((receivedBytes / totalBytes) * 90))
+        onProgress(pct, `downloading (${Math.round(receivedBytes / 1024 / 1024)} MB)`)
+      }
+      cb(null, chunk)
+    },
   })
+  await pipeline(Readable.fromWeb(res.body), progress, createWriteStream(destPath))
 }
 
 export function extractTarBz2(archivePath, targetDir) {
@@ -191,7 +185,7 @@ export async function startSensevoiceInstall(ctx, cfg, updateConfigFn, startSens
 export function registerSensevoiceInstaller(ctx, { liveConfig, isAlive, updateConfig, startSensevoice }) {
   if (!ctx.webServer || typeof ctx.webServer.register !== 'function') return
 
-  ctx.webServer.register({
+  return ctx.webServer.register({
     kind: 'exact',
     path: '/dsh-voice/sensevoice-installer',
     handler: async (req, res) => {

package/lib/transcribe-core.js

@@ -11,8 +11,9 @@
  * @returns {{order: string[], models: Record<string, string>}}
  */
 export function buildProviderOrder({ localOnly, chain, customKeys, knownKeys, defaultModels }) {
+  const localEngines = new Set(['local-whisper', 'sensevoice'])
   const source = localOnly
-    ? (chain || []).filter((e) => e && e.provider === 'local-whisper')
+    ? (chain || []).filter((e) => e && localEngines.has(e.provider))
     : (chain || [])
   const models = {}
   const order = []
@@ -23,7 +24,7 @@ export function buildProviderOrder({ localOnly, chain, customKeys, knownKeys, de
     models[entry.provider] = entry.model || defaultModels[entry.provider] || ''
   }
   if (localOnly && order.length === 0) {
-    throw new Error('localOnly mode is on, but local-whisper is not in the chain')
+    throw new Error('localOnly mode is on, but no local engine (local-whisper or sensevoice) is in the chain')
   }
   return { order, models }
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@goodandready/dsh-voice",
-  "version": "0.8.32",
+  "version": "0.8.34",
   "description": "Voice input for DeepSeek Harness: dictation chunked by pauses and voice messages, each with its own provider fallback chain (Deepgram, Groq, HuggingFace, local whisper.cpp, plus any OpenAI-compatible API of your own).",
   "license": "MIT",
   "type": "module",