Generate a final Veo 3.1-compliant ai_video_prompt based on prompt_type.

Request Fields:

• ai_video_prompt: The original video prompt description
• voiceover: The voiceover text that will be spoken (used for narration/lipsync types)
• prompt_type: One of "narration", "lipsync", or "description" (or empty/None)

Response: Returns a dictionary with:

• ai_video_prompt: Final, Veo 3.1-compliant prompt ready for Veo model
• voiceover: The voiceover text (extracted from prompt for description type, or provided for other types)

Prompt Types:

• "narration": Combines ai_video_prompt and voiceover, NO lip sync, NO background sound effects
• "lipsync": Combines ai_video_prompt and voiceover, PROPER lip sync required, NO background sound effects
• "description": Cleans and enhances ai_video_prompt, extracts speaking text as voiceover