Add local model support: Ollama and HuggingFace Transformers

- Add ProviderType.OLLAMA and ProviderType.HUGGINGFACE
- Implement OllamaProvider with HTTP API support
  - Sync/async call and streaming
  - list_models() to enumerate available models
  - Tool calling via prompt simulation
- Implement TransformersProvider for local HuggingFace models
  - Lazy model loading
  - Sync/async call and streaming with TextIteratorStreamer
  - Tool calling via prompt simulation
- Update ProviderFactory to register new providers
- Update create_provider() with ollama/huggingface support
- Add 11 new tests for local model providers
- Fix mypy type annotations with TYPE_CHECKING

Author: ChidcGithub

codegnipy/__init__.py

@@ -86,6 +86,8 @@
     BaseProvider,
     OpenAIProvider,
     AnthropicProvider,
+    OllamaProvider,
+    TransformersProvider,
     ProviderFactory,
     create_provider
 )
@@ -169,6 +171,8 @@
     "BaseProvider",
     "OpenAIProvider",
     "AnthropicProvider",
+    "OllamaProvider",
+    "TransformersProvider",
     "ProviderFactory",
     "create_provider",
     # Validation