demo: add video voice semantic search smoke

[qiffang]

Summary

Adds a minimal Drive9-native video voice semantic-search demo:

• documents the GitHub repo survey and why transcript-first is the fastest safe path
• adds scripts/verify_local_video_voice_search_demo.py for local stub and provider smoke tests
• adds helper unit tests for supported media/path validation
• advertises the new video voice smoke in scripts/drive9-server-local-env.sh

This PR is intentionally a draft and should not be merged until qiffang reviews the direction. It proves voice-in-video search through the existing audio_extract_text -> files.content_text -> grep/FTS/vector path, not visual scene retrieval or timestamped segment search.

Validation

• python3 -m py_compile scripts/verify_local_video_voice_search_demo.py scripts/verify_local_video_voice_search_demo_test.py scripts/verify_local_audio_extract.py
• python3 -m unittest scripts/verify_local_video_voice_search_demo_test.py
• bash -n scripts/drive9-server-local-env.sh
• python3 scripts/verify_local_video_voice_search_demo.py --help
• git diff --cached --check
• go test ./cmd/drive9-server -run Audio -count=1
• go test ./cmd/drive9-server-local -run Audio -count=1

Notes

I did not run the live drive9-server-local stub demo because it requires local TiDB/server runtime setup. The script is wired to use the existing local audio verifier helpers once that runtime is started.

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: ef7cd004-e2f5-4453-b0a9-65f382a2f97f

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch adversary-1/video-semantic-demo

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[qiffang]

Closing: superseded by PR #751 (video visual extraction). Voice-based audio path remains available via existing audio_extract_text if needed later.