feat: video-to-video endpoints + v2m ducking/preserve_speech/output_format

[spencer-zqian]

Summary

Brings the Python SDK to parity with the current public API.

• Add video_to_video_music and video_to_video_sfx resources (sync + async, video output) with submit()/generate(), a new VideoResult dataclass, and parse_video_result.
• Extend video_to_music.submit/generate_async: add preserve_speech (primary; isolate_vocals kept as a forwarded alias), output_format (m4a|wav), and ducking. ducking is omitted when None so the backend default-on applies; MusicResult gains a ducked field.
• Add async submit()/generate_async() + mode/output_format to text_to_music.
• Version 0.3.0 → 0.4.0 (pyproject.toml + _version.py); widen sonilo-video-kit's core pin to >=0.3,<0.5.
• Not published — no tags pushed.

Testing

• python -m pytest: core 157/157 pass; sonilo-video-kit 52/52 pass.
• Production E2E: v2v-music, v2v-sfx (valid mp4s); video_to_music(ducking=True) returned a populated ducked entry; text_to_music(output_format="wav") produced real pcm_s16le wav.

🤖 Generated with Claude Code

https://claude.ai/code/session_013rzBraznKwUMK2JQwTwf5r

[sapient-app]

Sapient docs evals

Waiting for the staging docs URL before running evals.

Sapient will start the selected PR evals automatically when GitHub reports a successful docs deployment for this PR. This usually happens within 15 minutes.

Commit: 9e39afc
Status: waiting for staging docs URL