feat(sim): add Batumi radio physics for DCR

[Komzpa]

Summary

This draft remains the rollup/reference branch for the Meshtasticator radio-physics and policy work split from the original large patch. Please review the smaller draft PRs below in order; each slice has its own focused commit, docs/tests, and validation notes.

The stack keeps default simulator behavior unchanged unless the new CLI flags or packaged preset are used. The main goal is to make DCR/DTP experiments measurable in Meshtasticator without asking reviewers to swallow one giant patch at once.

Split Tracker

• PR1 CLI/import hygiene: refactor(sim): make lora mesh cli importable #67
• PR2 scenario input hardening and terrain/SRTM inputs: feat(sim): add map and terrain inputs #68
• PR3 land-cover clutter and OSM exporter: feat(sim): add land-cover clutter inputs #69
• PR4 RF link budget, packet-loss, and capture-collision physics: feat(sim): add capture-aware rf physics #70
• PR5 Dynamic Coding Rate simulator policy: feat(sim): add dynamic coding rate policy #71
• PR6 Dynamic TX Power simulator policy: feat(sim): add dynamic tx power policy #72
• PR7 Batumi preset, calibration, and packaged data: feat(sim): add batumi radio preset #73
• PR8 user/CI comparison workflow and quickstart: feat(sim): add radio policy compare workflow #74
• PR9 reproducibility and interactive link-budget cleanup: fix(sim): make radio comparisons reproducible #75

Review Notes

• The rollup is not intended to be merged as one patch.
• The split PRs are stacked because later slices need the simulator hooks and preset from earlier slices.
• Each split PR body names the intended review commit so the actual new diff is easy to inspect even while stacked.
• The packaged Batumi scenario is sanitized: simulator geometry, terrain/clutter grids, and aggregate calibration only; no original node names or source identifiers are included.
• Remaining rollup-only differences are old duplicate documentation/test reshuffles and are intentionally not part of the review stack.

Validation

The split PRs carry focused validation for their own files. The final cleanup slice was checked with focused tests, ruff, compileall, a static/DCR/DCR+DTP smoke comparison, full unittest, diff whitespace checks, and a private-marker scan.

[CLAassistant]

All committers have signed the CLA.

[zandi]

I'm not a maintainer of this repo and don't claim or intend to be, but respectfully this is a colossal PR (11k lines added with many new files & features) for a repo with, as far as I've seen, a single maintainer who isn't maintaining this repo as their day job. IMO this would need to at minimum be split into separate digestible PRs with clear intention behind the changes to be accepted, but that's not my call. I'll not go on an anti-AI rant for now.

[GUVWAF]

I agree with the above; while I think there might be useful additions in the PR, it is currently way too big to be reviewed.
I see that you/your AI now created a list with "mergeable slices", which is good and PR #67 looks reasonable in size and affected changes.

Besides that, I would like to keep the main branch clean by only simulating what current Meshtastic devices do or can be configured to. Dynamic Coding Rate is not yet there (there is this open PR though: meshtastic/firmware#10359), so I would propose to keep those on your fork, at least for now.

[Komzpa]

Thanks @GUVWAF @zandi. I mainly did this to check my hypotheses on meshtastic/firmware#10359 and dynamic power, so I don't offer to main branch something that won't work in practice. DCR seems a win and dynamic power seems a fail. That part is done for me, but I'll be happy to spend some AI tokens to document and clean it up make it useful for future experimenters. Will fix any comments there.

I hope it's not too annoying with the AI-side - without AI in this case I'll still be dreaming to make dynamic power work, here now I have model that it's not a helpful thing on our real mesh. I do not have intention to make your life harder as maintainer, just sharing results that I happen to have.