@goodandready/dsh-voice 0.8.12 → 0.8.14

package/lib/client.js

@@ -343,10 +343,10 @@ window.__ModuleLoader__.load({
       '.dvo-dash-name{font-size:12px;font-weight:600;color:var(--dsw-alias-label-primary)}' +
       '.dvo-dash-row{display:flex;justify-content:space-between;align-items:center;font-size:11px;color:var(--dsw-alias-label-secondary)}' +
       '.dvo-badge{display:inline-flex;align-items:center;font-size:11px;padding:1px 6px;border-radius:6px;font-weight:600}' +
-      '.dvo-badge-fast{background:rgba(16,185,129,0.15);color:#10b981}' +
-      '.dvo-badge-norm{background:rgba(245,158,11,0.15);color:#f59e0b}' +
-      '.dvo-badge-slow{background:rgba(249,115,22,0.15);color:#f97316}' +
-      '.dvo-badge-err{background:rgba(239,68,68,0.15);color:#ef4444}' +
+      '.dvo-badge-fast{background:var(--dsw-alias-bg-layer-2);color:var(--dsw-alias-state-success-primary);border:1px solid var(--dsw-alias-state-success-primary)}' +
+      '.dvo-badge-norm{background:var(--dsw-alias-bg-layer-2);color:var(--dsw-alias-state-warning-primary);border:1px solid var(--dsw-alias-state-warning-primary)}' +
+      '.dvo-badge-slow{background:var(--dsw-alias-bg-layer-2);color:var(--dsw-alias-state-warning-primary);border:1px solid var(--dsw-alias-border-l2)}' +
+      '.dvo-badge-err{background:var(--dsw-alias-bg-layer-2);color:var(--dsw-alias-state-error-primary);border:1px solid var(--dsw-alias-state-error-primary)}' +
       '.dvo-badge-idle{background:var(--dsw-alias-bg-layer-2);color:var(--dsw-alias-label-tertiary)}'
     const cssId = 'dsh-voice/client.module.css'
     if (typeof document !== 'undefined' && !document.querySelector('style[data-plugin-css="' + cssId + '"]')) {
@@ -736,7 +736,8 @@ window.__ModuleLoader__.load({
         rec.cutting = false
         if (blob.size < 1200) return       // слишком короткий кусок — это не речь
         try {
-          const text = await sendAudio(blob, rec.mime, 'dictation')
+          const out = await sendAudio(blob, rec.mime, 'dictation')
+          const text = out && out.text ? out.text : ''
           const delay = Number(voice.settings.sendDelayMs) || 0
           if (text && delay > 0 && !voice.holding) {
             // Отложенная вставка с окном отмены (#29-5): текст уже вставлен,

package/lib/index.js

@@ -158,7 +158,7 @@ export const Config = z.object({
     .description('Realtime streaming provider: "openai" (OpenAI Realtime API) or "sherpa-onnx" (local streaming).'),
   realtimeModel: z.string().default('gpt-4o-realtime-preview')
     .description('Realtime model identifier.'),
-  visualizerStyle: z.enum(['liquid-wave', 'dynamic-orb', 'bars', 'off']).default('liquid-wave')
+  visualizerStyle: z.union(['liquid-wave', 'dynamic-orb', 'bars', 'off']).default('liquid-wave')
     .description('Audio visualizer animation style in the recording pill: "liquid-wave", "dynamic-orb", "bars", or "off".'),
 })
 
@@ -442,7 +442,7 @@ export function apply(ctx, baseConfig) {
         realtimeProvider: cfg.realtimeProvider || 'openai',
         realtimeModel: cfg.realtimeModel || 'gpt-4o-realtime-preview',
         visualizerStyle: cfg.visualizerStyle || 'liquid-wave',
-        providerStats: getProviderStats(),
+        providerStats: statsTracker.get(),
       })
     },
   }), 'dsh-voice: /status route')
@@ -565,7 +565,7 @@ export function apply(ctx, baseConfig) {
         if (info.size > cfg.maxFileBytes) {
           throw new Error(`transcribe_audio: file too large (${info.size} bytes, max ${cfg.maxFileBytes})`)
         }
-        if (info.size < 100) throw new Error('transcribe_audio: file is empty or too small')
+        if (info.size < 44) throw new Error('transcribe_audio: file is empty or too small')
         const mime = MIME_BY_EXT[path.extname(filePath).toLowerCase()] || 'audio/wav'
         const bytes = await readFile(filePath)
         const modeCfg = { ...cfg.message, language: String(args.language || cfg.message.language) }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@goodandready/dsh-voice",
-  "version": "0.8.12",
+  "version": "0.8.14",
   "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",