Port “Compass view” bottom sheet from Meshtastic-Apple PR #1504

[jakevis]

This ports the “Compass view” from Meshtastic-Apple PR pull/1504 (thanks @RCGV1) to Android. It keeps the phone’s location as origin and guides you to the selected node via a modal bottom sheet on Node Detail → Position → “Open Compass.”

What changed

• Add localized strings for compass title/action, distance/bearing display, and sensor/permission/location warnings.
• New providers: CompassHeadingProvider (rotation vector with accel+mag fallback) and PhoneLocationProvider (LocationManager flow with permission/provider state).
• CompassViewModel/CompassUiState combine heading + phone location with the target node, computing distance, bearing, alignment, and elapsed “last position update.”
• Compose CompassSheetContent: dial ~2/3 width, red north reference, black target bearing, larger cardinal labels + quarter/eighth ticks, distance/bearing rows, elapsed time, and warnings with CTAs (request permission, open location settings, request position).
• Node Detail integration: Position section gets “Open Compass”; sheet starts/stops providers with visibility, caches target node, and includes “Request Position” inside the sheet.

Why
Per the iOS PR: adds a compass view to quickly navigate to a node—useful off-grid when locating a friend, dog, or tracker.

Testing

• Built and ran on Android emulator (API 35); opened Node Detail → Position → Open Compass; verified sheet renders and warnings appear when sensors/permissions are missing. (No automated tests run in this PR.)

[CLAassistant]

All committers have signed the CLA.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[jakevis]

Screenshots of the changes:

[DaneEvans]

Why lines not a compass rose? And why are there 3 lines for a single bearing?

[jakevis]

Why lines not a compass rose? And why are there 3 lines for a single bearing?

Personal preference and re-using some old code I had from another project. Red is fixed line on the N marker; the thin under it is a heading line. Align the black lines and your walking directly to the node (accessibility feature mostly from the old project).

Orignal code and design I was playing with had support for multiple nodes on the rose (and that might a future PR - with a rose accessible from somewhere else). Totem Compass style - show where you stared nodes are both distance and direction from you.

For the moment; happy to drop the heading and red lines if folks prefer.

[mdecourcy]

Should there perhaps be a message displayed when the target node has a high degree of precision bits? So we don't send the users on a wild goosechase

[jakevis]

Should there perhaps be a message displayed when the target node has a high degree of precision bits? So we don't send the users on a wild goosechase

Let me see how hard it would be to tell the user the degree of error on the screen.. if its communicated somewhere..

[jakevis]

@DaneEvans @mdecourcy - ok simplified the rose and did implement an accuracy cone-

• Data source: We use gps_accuracy (mm) and DOP (PDOP, or HDOP/VDOP fallback) from the node’s Position. If these are missing, uncertainty is “unknown.”
• Visual: A translucent blue wedge around the bearing shows ±angular error (error radius vs. your distance). Close range + poor accuracy = wide wedge; good accuracy or long distance = narrow wedge.
• Messaging (beneath the dial, next to distance/bearing):
  • Known: “Estimated area: ±X (±Y°)” — rendered from compass_uncertainty.
  • Unknown: “Estimated area: unknown accuracy” — rendered from compass_uncertainty_unknown.

I need to test this on some more physical devices though- only tested this in the emulator right now- but would love feedback.

[jakevis]

Converting to draft pending physical device testing of the error cone.

[RCGV1]

@jakevis I made a pr into your fork repo to improve the view of the compass.

[jakevis]

@jakevis I made a pr into your fork repo to improve the view of the compass.

Looks amazing! Thank you!

[jamesarich]

How y'all feeling about this, @jakevis @RCGV1 ?

Should we let copilot do a review pass?

[RCGV1]

The only thing I have not fully tested is the uncertainty feature, also I did not figure out how to localize distance since right now it only shows km

[jakevis]

I have a todo item tomorrow to test this on a couple physical devices now I'm back from the weekend. I want to check the uncertainty cone some more- it's a tad hard to do in the emulator.

I have ran codex-max over it before I pushed to my branch; but yes - I would certainly say copilot should triple check 😀 - I can ask codex to take a pass at the localization as well.