feat: compose Q10 map content from grouped traits

[tubededentifrice]

Summary

Completes the Q10 map integration on top of #885 with a small low-level/high-level trait boundary matching the three incoming data streams.

• MapDpsTrait owns restricted-zone and virtual-wall DPS values and exposes their decoded overlays.
• MapContentTrait explicitly depends on that DPS trait, retains the latest map and trace packets, and calls the pure render_q10_map function whenever renderable source data changes.
• The high-level trait exposes only caller-facing map data: the composed image, rooms, path, robot position, and heading.
• Calibration, layers, projected map data, and other rendering intermediates stay inside the renderer.
• A rendering failure clears the derived image instead of pairing stale image bytes with newer source state.
• DPS updates received before any map are retained by MapDpsTrait without emitting a no-op high-level map notification.
• Q10PropertiesApi wires the DPS trait into both DPS fanout and the dependent map-content trait.

The redundant path-specific CLI command from the staging stack is intentionally excluded, since normal image_content is already the single composed output. Trait tests cover source ownership and renderer handoff; calibration and pixel behavior remain in the map-module renderer tests.

Validation

• pre-commit run --show-diff-on-failure --color=always --all-files
• focused map/trait suite — 87 passed
• pytest -q — 796 passed, 86 snapshots passed

Stack

• feat: compose Q10 maps in a pure renderer #885 — merged pure Q10 renderer
• This PR — trait ownership and integration

[allenporter]

Looks great, just a minor change nit -- perhaps you can remove in a follow up PR? Thank you!

[allenporter]

Lets not test internal details like this, just exercise behavior through the "api" methods.