Add client.videos.generate() for /v1/videos/generations endpoint

[flyworker]

Context

mega-nova-api is adding POST /v1/videos/generations (MeganovaAI/mega-nova-api#435). The SDK needs a matching client method.

Requirements

• Create meganova/resources/videos.py with VideosResource
• Add client.videos.generate() method
• Support parameters: prompt, model, image (for i2v), duration, aspect_ratio
• Handle async polling pattern (video gen takes >30s)
  • generate() returns a task/job ID
  • client.videos.get(job_id) polls for completion
  • Optional: generate(wait=True) blocks until done
• Create Pydantic models in models/videos.py:
  • VideoGenerationRequest
  • VideoGenerationResponse(id, status, video_url, ...)
• Wire into main MegaNova client class
• Add example: examples/video_generation.py

Response Model

class GeneratedVideo(BaseModel):
    url: str
    duration: Optional[float]
    model: str

class VideoGenerationResponse(BaseModel):
    id: str
    status: str  # pending, processing, completed, failed
    data: Optional[List[GeneratedVideo]]
    created: int

Depends On

• mega-nova-api#435 (video endpoint implementation)

[flyworker]

📋 Reference Report: AI API Compatibility Best Practices Research (29 items, 7 categories — protocols, gateways, SDKs, cloud platforms, local inference)

Tracking Epic: MeganovaAI/mega-nova-api#447

[flyworker]

Implemented in PR #15