fix(flagd-provider): update fractional targeting test expectations for flagd v0.15.4

[jonathannorris]

Summary

• Updates the color-palette-experiment fractional targeting test expectations to match the bucketing behavior in flagd v0.15.4
• The fractional evaluator changed to high-precision integer arithmetic ((hash * totalWeight) >> 32), which rebucketed all users — tracked in [flagd] [FEATURE] Harden Hashing Consistency And Add Support For Non-string Attributes in Fractional Evaluation #73
• Expected values verified by running the exact twmb/murmur3 + distributeValue algorithm from flagd's source at core/v0.15.4
• Originally identified as a pre-existing failure in chore: resolve open dependabot security alerts #165
• Replaced targeting keys with ones that resolve to all four distinct variants (red, blue, green, grey), ensuring the test validates fractional bucketing rather than just the default fallback

[gemini-code-assist]

Code Review

This pull request updates the expected hex color values in the flagd provider's test suite for fractional targeting. Feedback suggests that the updated expectations are less effective because most keys now resolve to the default variant, making it difficult to verify if the fractional logic is functioning correctly or simply falling back to defaults. It is recommended to use targeting keys that map to a more diverse set of variants to ensure proper test coverage.