refactor: remove dead code, deduplicate backends

Phase 1 - delete dead code:
- studio.py, migrate_add_instruct.py, utils/validation.py
- duplicate _profile_to_response in main.py, duplicate asyncio import
- pointless _get_profiles_dir/_get_generations_dir wrappers
- duplicate LANGUAGE_CODE_TO_NAME and WHISPER_HF_REPOS constants

Phase 2 - extract backends/base.py with shared utilities:
- is_model_cached() replaces 7 copy-pasted HF cache checks
- get_torch_device() replaces 5 device detection methods
- combine_voice_prompts() replaces 5 identical implementations
- model_load_progress() ctx manager replaces progress boilerplate in all backends
- patch_chatterbox_f32() replaces identical monkey-patches in both chatterbox backends

net -1078 lines across the backend

backend/main.py

@@ -18,7 +18,6 @@ import tempfile
 import io
 from pathlib import Path
 import uuid
-import asyncio
 import signal
 import os
 
@@ -54,6 +53,7 @@ def _safe_content_disposition(disposition_type: str, filename: str) -> str:
 
 from . import database, models, profiles, history, tts, transcribe, config, export_import, channels, stories, __version__
 from .database import get_db, Generation as DBGeneration, VoiceProfile as DBVoiceProfile
+from .profiles import _profile_to_response
 from .utils.progress import get_progress_manager
 from .utils.tasks import get_task_manager
 from .utils.cache import clear_voice_prompt_cache
@@ -1980,31 +1980,6 @@ async def update_profile_effects(
     return _profile_to_response(profile)
 
 
-def _profile_to_response(profile) -> models.VoiceProfileResponse:
-    """Convert a DB profile to a VoiceProfileResponse with parsed effects_chain."""
-    import json as _json
-    import logging
-
-    effects_chain = None
-    if profile.effects_chain:
-        try:
-            raw = _json.loads(profile.effects_chain)
-            effects_chain = [models.EffectConfig(**e) for e in raw]
-        except Exception as e:
-            logging.warning(f"Failed to parse effects_chain for profile {profile.id}: {e}")
-
-    return models.VoiceProfileResponse(
-        id=profile.id,
-        name=profile.name,
-        description=profile.description,
-        language=profile.language,
-        avatar_path=profile.avatar_path,
-        effects_chain=effects_chain,
-        created_at=profile.created_at,
-        updated_at=profile.updated_at,
-    )
-
-
 # ============================================
 # FILE SERVING
 # ============================================