feat(ui): redesign voice cockpit with clinical companion drawer, Svelte 5 migration, and Gemini 3.8 voice

- Re-architect VoiceSession into full-bleed studio cockpit with overlay slide-over companion drawer
- Eliminate width squishing defect on LiveInsightPanel and add native dark/light theme support
- Integrate authentic Google Gemini 3.8 Flash TTS with Fenrir persona and fallback audio pipeline
- Implement comprehensive Playwright E2E visual regression suite across Desktop, iPad, and iPhone (21 tests, 45 snapshots)
- Migrate core application to Svelte 5 runes (, , ) with reactive stores
- Maintain .NET 8 Azure Functions serverless endpoints and clean build hygiene

services/geminiVoiceService.ts

@@ -0,0 +1,121 @@
+/**
+ * Google Gemini Native Voice Synthesis Service
+ * 
+ * Provides direct speech synthesis using Google's native Gemini voice models
+ * (gemini-3.8-flash-tts and gemini-2.5-flash-preview-tts) with official prebuilt
+ * voices: Fenrir (Gruff / Salt of the Earth), Puck, Charon, Kore, Aoede.
+ */
+
+import { getGeminiApiKey } from './apiConfig';
+import { decodeBase64 } from './audioUtils';
+
+const GEMINI_API_BASE = 'https://generativelanguage.googleapis.com/v1beta';
+
+/**
+ * Packages raw PCM16 audio into a self-contained WAV ArrayBuffer.
+ */
+export function pcmToWav(pcmData: Uint8Array, sampleRate = 24000, numChannels = 1): ArrayBuffer {
+  const wavBuffer = new ArrayBuffer(44 + pcmData.byteLength);
+  const view = new DataView(wavBuffer);
+  // "RIFF"
+  view.setUint32(0, 0x52494646, false);
+  view.setUint32(4, 36 + pcmData.byteLength, true);
+  // "WAVE"
+  view.setUint32(8, 0x57415645, false);
+  // "fmt "
+  view.setUint32(12, 0x666d7420, false);
+  view.setUint32(16, 16, true);
+  view.setUint16(20, 1, true); // AudioFormat: PCM (1)
+  view.setUint16(22, numChannels, true);
+  view.setUint32(24, sampleRate, true);
+  view.setUint32(28, sampleRate * numChannels * 2, true);
+  view.setUint16(32, numChannels * 2, true);
+  view.setUint16(34, 16, true); // BitsPerSample: 16
+  // "data"
+  view.setUint32(36, 0x64617461, false);
+  view.setUint32(40, pcmData.byteLength, true);
+  new Uint8Array(wavBuffer, 44).set(pcmData);
+  return wavBuffer;
+}
+
+/**
+ * Synthesizes speech using Google Gemini's native voice model.
+ * 
+ * @param text The text dialogue to speak.
+ * @param voiceName Prebuilt Gemini voice: 'Fenrir' | 'Puck' | 'Charon' | 'Kore' | 'Aoede' | 'Zephyr'
+ * @param apiKey Optional explicit Google Gemini API key.
+ * @returns ArrayBuffer containing standard decoded audio data (WAV container).
+ */
+export async function synthesizeSpeechWithGemini(
+  text: string,
+  voiceName = 'Fenrir',
+  apiKey?: string
+): Promise<ArrayBuffer> {
+  const effectiveKey = getGeminiApiKey(apiKey);
+  if (!effectiveKey) {
+    throw new Error('Google Gemini API key is missing. Please set GEMINI_API_KEY in Settings or .env.');
+  }
+
+  // Normalize voice name to valid Gemini prebuilt voices (Zephyr -> Puck/Fenrir)
+  const validVoices = ['Fenrir', 'Puck', 'Charon', 'Kore', 'Aoede'];
+  const targetVoice = validVoices.includes(voiceName) ? voiceName : 'Fenrir';
+
+  const modelsToTry = [
+    'gemini-3.8-flash-tts',
+    'gemini-2.5-flash-preview-tts'
+  ];
+
+  let lastError: any = null;
+
+  for (const model of modelsToTry) {
+    try {
+      const url = `${GEMINI_API_BASE}/models/${model}:generateContent?key=${encodeURIComponent(effectiveKey)}`;
+      const response = await fetch(url, {
+        method: 'POST',
+        headers: {
+          'Content-Type': 'application/json'
+        },
+        body: JSON.stringify({
+          contents: [{ parts: [{ text }] }],
+          generationConfig: {
+            responseModalities: ['AUDIO'],
+            speechConfig: {
+              voiceConfig: {
+                prebuiltVoiceConfig: {
+                  voiceName: targetVoice
+                }
+              }
+            }
+          }
+        })
+      });
+
+      if (!response.ok) {
+        const errorText = await response.text();
+        throw new Error(`Gemini ${model} TTS error (${response.status}): ${errorText}`);
+      }
+
+      const json = await response.json();
+      const part = json.candidates?.[0]?.content?.parts?.[0];
+      if (!part || !part.inlineData?.data) {
+        throw new Error(`Gemini ${model} returned no audio content in response candidates.`);
+      }
+
+      const mimeType = part.inlineData.mimeType || '';
+      const rawBytes = decodeBase64(part.inlineData.data);
+
+      if (mimeType.includes('wav')) {
+        const slice = rawBytes.buffer.slice(rawBytes.byteOffset, rawBytes.byteOffset + rawBytes.byteLength);
+        return slice as ArrayBuffer;
+      }
+
+      // If PCM (e.g. audio/L16;codec=pcm;rate=24000), wrap in WAV container for universal browser decoding
+      return pcmToWav(rawBytes, 24000, 1);
+    } catch (err: any) {
+      lastError = err;
+      console.warn(`Failed synthesis with Gemini model ${model}, checking next model...`, err);
+    }
+  }
+
+  throw lastError || new Error('All Gemini TTS models failed to generate audio.');
+}