feat: add OpenAI-compatible Bedrock provider with error handling

Implements AWS Bedrock inference provider using OpenAI-compatible endpoint
for Llama models available through Bedrock.

Changes:
- Add BedrockInferenceAdapter using OpenAIMixin base
- Configure region-specific endpoint URLs
- Add NotImplementedError stubs for unsupported endpoints
- Implement authentication error handling with helpful messages
- Remove unused models.py file
- Add comprehensive unit tests (12 total)
- Add provider registry configuration

src/llama_stack/distributions/ci-tests/run.yaml

@@ -46,6 +46,9 @@ providers:
       api_key: ${env.TOGETHER_API_KEY:=}
   - provider_id: bedrock
     provider_type: remote::bedrock
+    config:
+      api_key: ${env.AWS_BEDROCK_API_KEY:=}
+      region_name: ${env.AWS_DEFAULT_REGION:=us-east-2}
   - provider_id: ${env.NVIDIA_API_KEY:+nvidia}
     provider_type: remote::nvidia
     config: