The args variable was not accessible in the create_transcription function,
causing a NameError when using --whisper-cpp CLI option. This fix adds
global_args to store the parsed arguments for access in endpoint functions.

- Modified build.sh to build whispercpp with Vulkan support
- Added --audio-vulkan-device argument to specify GPU device for Whisper
- Added Vulkan detection and logging for Whisper transcription
- Set GGML_VULKAN_DEVICE environment variable for GPU selection

faster-whisper doesn't support GGUF format (it's llama.cpp format).
Now detects GGUF files by extension and goes directly to whispercpp.

- Add faster_whisper_failed flag to properly track failures
- When faster-whisper throws non-ImportError (e.g., GGUF not supported),
  now falls back to whispercpp instead of failing
- Applies to both pre-loading and transcription endpoint

- Add specific detection for 'invalid ELF' / 'Mach-O' architecture mismatch errors
- Improve error messages to mention both options:
  - Install PyTorch + faster-whisper
  - Use built-in whispercpp model (tiny/base/small/medium/large)
- Fix critical bug: now raises HTTPException instead of returning None

- Recognize built-in model names: tiny, base, small, medium, large-v1, large
- Allow pre-loading these models directly without file path

- Add better error detection for 'not a valid preconverted model' errors
- Provide clear guidance to users about whispercpp limitations
- Suggest installing faster-whisper with PyTorch or using built-in model names
- Update both transcription endpoint and pre-loading code

- Update transcription endpoint to try faster-whisper first, then whispercpp
- Update pre-loading code to support both backends
- Add whispercpp to all requirements files (vulkan, nvidia, default)
- Remove broken llama.cpp fallback (llama.cpp cannot transcribe Whisper)

When audio model is in GGUF format, use llama.cpp instead of faster-whisper
for pre-loading. This allows using Vulkan backend for audio transcription.

When only one model type is specified (e.g., only --audio-model with no
--model), automatically pre-load it even in on-demand mode. This ensures
the model is downloaded and ready for use.

- Add --loadall flag to pre-load all models at startup
- Add --loadswap flag to keep models in RAM, swap active to VRAM
- Fix bug where load_mode was used before being defined in audio model section
- Remove duplicate load_mode determination code
- Improve error message for no main model specified to include TTS

- Add --tts-model option for Kokoro TTS models
- Add /v1/audio/speech endpoint (OpenAI-compatible)
- Add model caching to prevent redundant downloads
- Replace MD5 with SHA-256 for cache keys
- Move hashlib and pathlib imports to module level

- --model is now optional if using audio or image models only
- Shows helpful error message with examples if no model specified
- Prints available models at startup

- Accept full HTTPS URLs for --model (Vulkan/GGUF models)
- Accept full HTTPS URLs for --audio-model (faster-whisper models)
- Downloads file to temp directory before loading
- Shows download progress percentage

- Add --debug CLI argument to enable debug mode
- When enabled, dumps full request body (no truncation)
- When enabled, dumps full generated text (no truncation)
- When enabled, dumps extracted tool calls in JSON format
- Useful for troubleshooting tool call issues

- Replace class-based Config with model_config = ConfigDict() in all Pydantic models
- Fix Jinja2 crash by ensuring all messages have content key that is never None
- Enhanced message cleaning in generate_chat and generate_chat_stream to create copies and ensure content is always a string
- Add final safety check in chat_completions endpoint for content handling

- Add explicit check for missing content key in message dictionaries
- Use more aggressive regex patterns in strip_tool_calls_from_content
- Handle tool call tags in various formats (JSON, XML, tool names)
- Add checks in format_messages, _manual_format_messages, and chat_completions endpoint
- Fixes: 'dict object' has no attribute 'content' error in Jinja2 templates

- Added safety check in generate_chat_stream to replace None content with empty string
- Added same check in generate_chat for consistency
- This prevents 'dict object has no attribute content' error when
  processing messages with tool_calls that have no text content

- Add --audio-model and --image-model CLI arguments
- Add --loadall, --audio-ctx, --audio-offload, --vision-ctx, --vision-offload args
- Implement MultiModelManager class for dynamic model switching
- Add POST /v1/audio/transcriptions endpoint (OpenAI-compatible)
- Add POST /v1/images/generations endpoint (OpenAI-compatible)
- Update endpoints to use multi_model_manager for model selection
- Audio uses faster-whisper for local transcription
- Images use Stable Diffusion via diffusers

- Fix: Handle None content in messages to prevent Jinja2 'dict object has no attribute content' error
  - Added safety check in chat_completions function
  - Fixed _manual_format_messages to explicitly check for None
  - Fixed format_messages in VulkanBackend to ensure content is never None

- Fix: Always filter tool call format from output
  - Changed filter to run unconditionally (not just when tools are present)
  - Added extra regex patterns for JSON format tool calls like <tool>{...}</tool>

- Also fixed: Minor typos in comments (cket ->cket)

- Add seen_signatures set to extract_tool_calls() to prevent duplicates
- Add strip_tool_calls_from_content() method to remove <tool>...</tool> tags
- Filter tool format from each chunk in real-time during streaming
- Simplify post-stream tool call handling since content is already cleaned
- Also handle non-streaming responses for tool call content cleanup