Add HFlow dataset snapshot use case

[kstonekuan]

Summary

• add a multimodal use case for inspecting generic HFlow dataset snapshots in Spotlight
• show how to export copied media, load samples.parquet, resolve relative media paths, and map media_uri to spotlight.Image
• add the guide to the existing Multimodal navigation without changing the use-case index

The integration uses standard Parquet tables and a pandas DataFrame; it does not require an HFlow-specific Spotlight adapter.

Validation

• ran Spotlight pre-commit hooks for the changed files
• built the complete MkDocs site
• launched a headless Spotlight viewer with the documented samples.parquet and copied-image mapping