Add Inkplate 6COLOR (7-color ACeP) support

[brandon-dacrib]

Summary

Adds support for the Soldered Inkplate 6COLOR (600×448, 7-color ACeP, panel_ic_type 0x0043):

• DISPLAY_PALETTE_MAP[(0x0043, ColorScheme.BWGBRYO)] = INKPLATE_6COLOR in display_palettes.py.
• encode_image: handle ColorScheme.BWGBRYO as 4bpp. The palette is ordered to the panel's native nibbles, so indices map directly (identity), unlike BWGBRY which remaps.

Dependency

Requires the Bwgbryo scheme + INKPLATE_6COLOR palette from OpenDisplay/epaper-dithering#43.

Testing

Validated end-to-end on real hardware via the CLI: provisioned an Inkplate 6COLOR (panel_ic_type=0x0043, color_scheme=8, 600×448) and opendisplay upload rendered a dithered image — full transfer + ACeP refresh, correct geometry. Color fidelity depends on the (currently nominal) palette calibration tracked in #43.

Related

• Device side: Add Inkplate 6COLOR (EP585C_600x448) ACeP 7-color panel bitbank2/bb_epaper#33 (panel driver), Add Inkplate 6COLOR panel mapping (mapEpd 0x0043 → EP585C_600x448) Firmware#41 (panel mapping)

🤖 Generated with Claude Code

[jonasniesner]

Please don't hardcore specific panels to a different color_scheme, if you look into the toolbox advanced view, I already reserved 7 for that

[brandon-dacrib]

Oh. Fair. A little clarification. What is the right thing to do here technically?
Thanks. I want to align this with the protocol/toolbox convention before updating the PRs.

1. Is color_scheme = 7 intended to be the generic 7-color ACeP/BWGBRYO mode, so I should replace my proposed BWGBRYO value 8 with 7?

2. In py-opendisplay, is it OK to keep a measured palette lookup keyed by (panel_ic_type, color_scheme) for calibration, or should this PR avoid
   adding the Inkplate-specific (0x0043, BWGBRYO) -> INKPLATE_6COLOR mapping entirely?

I noticed epaper-dithering currently defines GRAYSCALE_8 as value 7, so I want to avoid changing the wrong protocol value.

[brandon-dacrib]

Closing this because we are stopping the OpenDisplay path for Inkplate 6COLOR. This host-library mapping worked for the proof of concept, but production use would still depend on Firmware direct-write changes to avoid the BLE DIRECT_WRITE_START reset. Rather than carrying a partial OpenDisplay stack, we are moving Home Assistant integration to Soldered's ESPHome component: https://github.com/SolderedElectronics/Soldered-Inkplate-ESPHome.