fix: handle null metric values from platform API changes

[rdheekonda]

Issue: SDK failing with Pydantic validation errors after platform API hotfix for NaN handling in metrics.

Root Cause:

• Platform API now converts NaN → null in JSON responses (hotfix #1920)
• SDK Metric model expects float but receives None for undefined metrics
• Pydantic validation fails: "Input should be a valid number [type=float_type, input_value=None]"

Solution:

• Updated Metric.value type from float to float | None
• Maintains backward compatibility while supporting new API behavior

Testing:
✅ Verified SDK can parse responses with null metric values
✅ Existing functionality preserved for valid float values
✅ Authentication and project listing work correctly

Related: Platform API hotfix #1920 - NaN JSON serialization fix

Compatible with platform API changes that convert NaN evaluation metrics to null.

Generated Summary:

• Changed the type of value in the Metric class from float to float | None.
• This allows value to also be None, providing greater flexibility in representing metrics without a defined value.
• Impact: Potentially avoids issues where a metric may not be applicable or is unknown, improving data integrity.

This summary was generated with ❤️ by rigging