feat(interactions): expose SpeakerConfig.mode and SpeakerConfigMode in SDK

[adri22235]

Summary

Exposes SpeakerConfig.mode and SpeakerConfigMode (verbatim, conversational) in the Interactions types to support dual-speaker screenplay and voice mode controls for Gemini 3.8 Flash TTS models.

Changes Included

• Updates google/genai/_gaos/types/interactions/__init__.py to export SpeakerConfigMode.
• Updates google/genai/_gaos/types/interactions/speakerconfig.py with SpeakerConfigMode union and adds mode to SpeakerConfig and SpeakerConfigParam.

Verification

• Fully tested across Python 3.10, 3.11, 3.12, and 3.13.
• Verified with active Google CLA signed by @adri22235.