refactor: provider-agnostic LLMJudge with auto-detection for OpenAI (#1103)

[TanujaNair03]

Description

This PR refactors the LLMJudge to be truly provider-agnostic by implementing automatic environment-based detection. While #477 added the ability to manually inject a provider, the judge still defaulted to a hardcoded Anthropic implementation that would fail for users without an Anthropic key or the anthropic package. This PR introduces a "Smart Default" system that prioritizes OPENAI_API_KEY when available, creating a seamless experience for non-Anthropic users without requiring manual code changes to existing tests.

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Refactoring (no functional changes)

Related Issues

Fixes #1103

Changes Made

• Automatic Provider Detection: Implemented _get_fallback_provider() to automatically detect and prioritize OPENAI_API_KEY followed by ANTHROPIC_API_KEY.
• Backward Compatibility: Maintained the legacy _get_client() method and implemented logic to detect mocked clients, ensuring 100% compatibility with existing unit tests.
• Robust JSON Parsing: Added a centralized _parse_json_result helper that robustly handles markdown code blocks and "chatter" from various LLM providers.
• Dependency Optimization: Lazy-loaded the anthropic client so the package is no longer a strict requirement for users running only OpenAI models.

Testing

• Unit tests pass (cd core && pytest tests/test_llm_judge.py)
• Lint passes (cd core && ruff check framework/testing/llm_judge.py)
• Manual testing performed (Verified on Python 3.14.2 with OpenAI keys)

Checklist

• My code follows the project's style guidelines
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[github-actions]

PR Closed - Requirements Not Met

This PR has been automatically closed because it doesn't meet the requirements.

PR Author: @TanujaNair03
Found issues: #1103 (assignees: none), #477 (assignees: pradyten)
Problem: The PR author must be assigned to the linked issue.

To fix:

1. Assign yourself (@TanujaNair03) to one of the linked issues
2. Re-open this PR

Why is this required? See #472 for details.

[TimothyZhang7]

lgtm