New anthropic client to support new generator output design

[phi-jkim]

What does this PR do?

These are some of the things this PR does and a link to the design document

Design document: https://www.notion.so/21c294805781808f898ae1800297b043?v=21c294805781801badc1000ca12d6f74&p=22629480578180cbaadde27d0574e135&pm=s

• Updated AnthropicAPIClient to use OpenAI SDK with Anthropic's compatibility
endpoint instead of the native Anthropic API client
• Added Response API compatibility converter (from OpenAI agents SDK) - Created
ChatCompletionToResponseConverter utility to transform ChatCompletion objects into
OpenAI's Response API format for raw_response of the generator output which should be model-agnostic (follow the OpenAI Response API standard)

Design: Anthropic Client Response API Support

The new Anthropic client implementation supports AdalFlow's model-agnostic
raw_response field design through the following architecture:

OpenAI SDK Compatibility Layer

• Base URL: Uses https://api.anthropic.com/v1/ with OpenAI SDK to call Anthropic's
  compatibility endpoint
• Authentication: Maintains ANTHROPIC_API_KEY environment variable support
• API Format: Converts inputs to OpenAI's ChatCompletion format while supporting
  Anthropic-specific features

Response API Converter (ChatCompletionToResponseConverter)

• Purpose: Transforms OpenAI ChatCompletion objects into Response API format for
  consistency across model providers
• Streaming Support: Handles both sync and async streaming with proper event
  sequencing
• Features Supported:
  • Text content streaming with ResponseTextDeltaEvent
  • Refusal handling with ResponseRefusalDeltaEvent
  • Tool calls with ResponseFunctionCallArgumentsDeltaEvent
  • Usage tracking with detailed token counts (including reasoning tokens)

Model-Agnostic Generator Output

• raw_response Field: Contains the processed Response API compatible output (text or
  stream events)
• api_response Field: Preserves the original ChatCompletion object for direct API
  access
• Usage Tracking: Converts ChatCompletion usage to AdalFlow's ResponseUsage format
  with detailed token breakdowns
• Backward Compatibility: Maintains existing parser interfaces for legacy code

Before submitting

• Was this discussed/agreed via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?

[liyin2015]

use adalflow setup_env

[liyin2015]

overall, it looks fine, will look closer tomorrow

[liyin2015]

lgtm. thx!