chore(sceptre): remove the unused SunSpec protocol

[adamsrnmsu]

Description

Remove SunSpec support from the SCEPTRE app.

SunSpec had been broken for years without anyone noticing: until #137, every inverter on the sunspec protocol raised KeyError. #137 asked whether SunSpec was still wanted. @GhostofGoes voted to remove it ("It was never used in all my time working on this stack"), and @mgaliar agreed ("I am good with keeping this fix or burning sunspec. We haven't used it in years.").

What's removed:

• Protocol code and data: protocols/sunspec/, meaning the SunSpecDevice register generator and about 1 MB of vendored SMDX models (99 files).
• Templates: templates/sunspec.mako, the gobennu branch of sceptre_start.mako, and the four sunspec branches of fep_template.mako.
• Protocol registrations: sunspec comes out of the metadata parser, the validation model, the field-device port table and the register builder.
• Exclusions: the SunSpec entries in package-data, the ruff exclude and per-file ignores, and .codespellrc.
• Docs: the SunSpec row in the app README's Known Issues table, the SunSpec note in src/python/AGENTS.md, and the (unreleased) CHANGELOG entry for the fix(sceptre): normalise the SunSpec infrastructure key #137 fix. A new ### Removed entry replaces it.

A field device whose metadata still has a sunspec key now fails validation with an error saying SunSpec was removed. Without that check, extra="allow" would silently ignore the key.

Related Issues/PRs

Follows up the discussion on #137. It doesn't depend on #152–#155 and doesn't conflict with #152, #153 or #155. #154 no longer touches the sunspec/__init__.py this PR deletes, so the two don't conflict.

Type of Change

• Chore (CI, build, dependencies, etc.) (chore)
• Other (please describe): removal of an unused protocol

Checklist

• This PR conforms to the process detailed in the Contributing Guide.
• I have included no proprietary/sensitive information in my code or the PR.
• I have commented my code, particularly in hard-to-understand areas.
• I have made corresponding changes to the documentation.
• I have tested my code (describe below).

Testing

• make check and make test in src/python on Python 3.12: 698 passed. That's main's 697 plus a new validation case for a leftover sunspec key.
• test_sceptre_golden.py passes unchanged, so the FEP and field-device configs rendered for the golden scenario are byte-identical after the template edits.
• infrastructures_golden.txt was regenerated, and the diff is exactly the 50 sunspec rows removed.
• A built wheel has no SunSpec files and still ships infrastructures.yaml and the mydesigner tree.

Additional Notes

An open question, deliberately left alone here: both inverters in configs/infrastructures.yaml still list SunSpec model numbers ([1, 103, 120, 123] and [1, 101, 123, 126]) as their analog-read-write fields. On DNP3 and Modbus those were already exposed as registers literally named 1, 103 and so on (visible in the golden file before this PR). That's pre-existing and unchanged here, and the table comment now says so. Should those become real field names, or be dropped?

🤖 Generated with Claude Code

https://claude.ai/code/session_016KAfcDUSerQCCWwBM9BxAo

[GhostofGoes]

Fix the slop then this is perfect. Feels good to burn away 25,000 lines of long dead code. Good riddance.

[GhostofGoes]

Instead of the slop, use this: ' Sceptre app: SunSpec has been burned away by the SCORCHing heat of 3000 Suns and thus no longer exists in this repo.'

[adamsrnmsu]

Done, verbatim.

[GhostofGoes]

LOL