feat: Add Edge tts

[keyboardstaff]

Integrate Edge TTS engine and build a unified multi-engine TTS system supporting Kokoro (local) and Edge TTS (online) with seamless switching.

---

Changes:

• New Edge TTS engine (edge_tts.py) — wraps edge-tts SDK with cached voice listing and synthesis outputting base64 WAV
• New voice list API (tts_voices.py) — POST /tts_voices returns engine-specific voices in unified [{id, name, language, gender}] format
• Unified synthesize API (synthesize.py) — routes to Kokoro or Edge TTS based on tts_engine setting, passes voice/rate params
• Parameterized Kokoro engine (kokoro_tts.py) — voice/speed now accept function arguments instead of hardcoded globals
• Settings upgrade (settings.py) — replaced tts_kokoro (bool) with tts_enabled, tts_engine, tts_voice, tts_rate; includes automatic migration of legacy field
• Frontend Settings UI (speech.html) — TTS toggle → engine selector → dynamic voice dropdown → speed slider
• Frontend TTS logic (speech-store.js) — renamed speakWithKokoro to speakWithServer for unified server-side TTS dispatch; browser TTS as fallback
• Preload (preload.py) — adapted to new tts_enabled + tts_engine fields
• Dependency (requirements.txt) — added edge-tts>=7.2.7