fix(voice): add service status validation, safe dictionary access, and fallback attributes in voice_status

[Vishaaallll]

Summary

Hardens the voice_status endpoint in ods/extensions/services/dashboard-api/routers/voice.py:

• Validates cfg object type (isinstance(cfg, dict)) before evaluating health checks for Whisper, TTS, and LiveKit services.
• Uses getattr(result, "status", "unknown") to safely extract service status from health check return objects.
• Replaces direct key lookups (entry["status"]) in installed_healthy with safe dictionary .get() calls (entry.get("status")) and isinstance(entry, dict) type guards.
• Enriches status payloads with explicit configured boolean metadata for downstream frontend rendering.

Why It Is Useful

• Prevents HTTP 500 Dashboard Errors: Replaces unhandled exceptions (KeyError, AttributeError, TypeError) with graceful fallback status representations ("unavailable", "not_configured").
• Supports Optional Voice Stack Topologies: Allows installations without optional voice components (or partial voice configurations) to query readiness endpoints reliably.
• Provides Richer Status Context: Adds explicit configured flags to response dictionaries, enabling the UI to distinguish between unconfigured services and failed service health checks.

Verification

• Ran test suite pytest ods/extensions/services/dashboard-api/tests/ -k voice -v (8 passed).