Refactor and add support for Inky Impression 7.3"

[axelson]

The Inky Impression 7.3" uses a different chip than the Impression 5.7" and 4" so it required a new HAL module.

To fit this new chip support into the library I also undertook a refactor

• Rename the phat to phat_original
  • I would have used the chip name instead of "original" but I haven't been able to find the name of the chip anywhere
• Segregate HAL for phat and impression based on chip type
• Compile test/support when MIX_TARGET=test

I have tested this on (using https://github.com/axelson/inky_tester):

• Inky pHAT SSD1608
• Inky Impression 5.7"
• Inky Impression 7.3"

This merges the contents of the impression branch which also includes code/work from @lawik and @jasonmj ❤️

[jasonmj]

I would have used the chip name instead of "original" but I haven't been able to find the name of the chip anywhere

I did some searching and found IL91874 at the top of the original python code. Maybe that's it?

[jasonmj]

Lookin great! Here's a first pass with plenty of small suggestions and some questions. I don't think anything here is a blocker really, but I'd be happy to see some of these changes.

[jasonmj]

Hmm, I wonder if Pimoroni would be willing to send us one? I went ahead and wrote a quick note to their support team to see if there's a chance.

[jasonmj]

Pimoroni is sending me a 4" Impression for testing 🎉

[axelson]

Wow, that's great!!

[jasonmj]

Since these are re-used, does it make sense to put them into a common helper module?

[jasonmj]

In other places, we're using Logger. Probably best to be consistent and use Logger here too.

[jasonmj]

Move up.

[jasonmj]

Lookin great! Here's a first pass with plenty of small suggestions and some questions. I don't think anything here is a blocker really, but I'd be happy to see some of these changes.

[nyaray]

So glad to see this continues to get usage/updates, nothing constructive to add, really. ✌️

[axelson]

I would have used the chip name instead of "original" but I haven't been able to find the name of the chip anywhere

I did some searching and found IL91874 at the top of the original python code. Maybe that's it?

Well spotted! Updated in 975d63b

[cloud8421]

Just wondering if this PR has fizzled? I’d be able to help from next week if there’s anything I can do. I have some spares I can test on (will have to check which models).

[lawik]

I don't know if my Impression is broken or if the code is :D

@cloud8421 I believe this ran for @axelson at some point. Try and see if you can get it running if you can. I'd love for us to wrap this up but was not involved very much in the later developments :)

[cloud8421]

Bit of a hiccup - I just tried to boot my impression 7" with the original firmware and I think there's some hardware issues as it's not working as it was before. Have to dig more but I'm unable to test for now. Will keep you posted.

[cloud8421]

No luck in reviving the old hardware - I've ordered a new one, it normally takes very little so I HOPE I can test this early in the week.

[lawik]

I would love if this finally goes over the finish line.

I should shove RPi OS on one of my units and just test my hardware.

[cloud8421]

Got the Impression 7.3 device and did initial testing.

As a testbed, I've generated a plain application with nerves.new, using a Raspberry PI Zero 2, and loaded a checked out local copy of this branch as a dependency.

I've then run the following code in the device SSH console:

{:ok, pid} = Inky.start_link(:impression_7_3, name: InkySample)

It throws the following error, and causes a device reboot:

opening DC pin
** (EXIT from #PID<0.1306.0>) shell process exited with reason: an exception was raised:
    ** (MatchError) no match of right hand side value: {:error, :export_failed}
        (inky 1.0.2) lib/hal/io_impression.ex:58: Inky.IO.Impression.init/1
        (inky 1.0.2) lib/hal/hal_impression_ac073tc1a.ex:110: Inky.HAL.ImpressionAC073TC1A.init/1
        (inky 1.0.2) lib/inky.ex:203: Inky.init/1
        (stdlib 6.0.1) gen_server.erl:2057: :gen_server.init_it/2
        (stdlib 6.0.1) gen_server.erl:2012: :gen_server.init_it/6
        (stdlib 6.0.1) proc_lib.erl:329: :proc_lib.init_p_do_apply/3

The error maps to:

{:ok, dc_pid} = gpio.open(pin_mappings[:dc_pin], :output, initial_value: 0)

I haven't touched any configuration, so following the code it should resolve to the default 22 pin, which according to the device schematics seems correct.

The impression device is new, but I don't think they've changed chipset: looking at Pimoroni Python's source, they map this model to the chipset implemented in this branch, and the PIN number seems correct.