feat(simulation): planet_class temperature envelope (T-1033, D-240) #164

Closed

jpmschweitzer wants to merge 0 commits from class-envelope-temperature into main

pull from: class-envelope-temperature

merge into: :main

:main

Conversation 2 Commits - Files Changed -

jpmschweitzer commented 2026-06-08 18:30:13 +02:00

Owner

Copy Link

Copy Source

Summary

T-1033 implements D-240: temperature derives from the authored planet_class envelope, not astrophysics. This fixes the unusable sun-driven model that produced +356 °C "temperate" worlds (NULL spectral_class → G-Sun default on an M-dwarf catalog).

Model (D-240)

• New [planet_class_temperature] envelope table in climate_constants.toml — the approved bands (frozen [−90,−25] … hot_arid [20,58] … volcanic [30,90]), with cold_*/hot_*/warm_* prefix parsing and a temperate fallback.
• derive_temperature_c(params, constants, body_seed): band from planet_class; latitude lerps across the band (equator → warm, pole → cold); atmosphere greenhouse + elevation lapse modulate within; a small seed nudge for per-world variety; clamped to the band — a body can never derive outside its class. Airless → None.

Removed (the non-canonical inputs, per D-240)

• Deleted derive_distance_au, ClimateConstants::luminosity(), the [star_luminosity] table.
• Stripped orbital_period_days / spectral_class / star_type / axial_tilt_deg from BodyParams + BodyParamsReader (now reads 3 columns from bodies, no star_systems join). DB columns left in place — no schema change, no systems.db regen.

Verification

• New every_planet_class_derives_within_its_band test — 13 classes × 5 atmospheres × 19 latitudes × 4 seeds, all in band (the D-240 consistency guard, true by construction via the final clamp).
• Acceptance check on 25 diverse real bodies (lead-run): every world now sits inside its class band — the +356 °C bodies derive to ≤28 °C, frozen → −25/−70, volcanic → 90/58, airless → none, latitude gradient holds everywhere.
• cargo test passes, clippy --all-targets -D warnings clean, fmt clean. D-010 integer-disciplined (the f32 temp feeds the integer climate ladders as before).

Ticket: T-1033 (epic T-974, decision D-240). After merge, the T-1031 harness rebases onto this and regenerates its golden against the corrected temperatures.

🤖 Generated with Claude Code

## Summary **T-1033** implements **D-240**: temperature derives from the authored `planet_class` envelope, not astrophysics. This fixes the unusable sun-driven model that produced +356 °C "temperate" worlds (NULL `spectral_class` → G-Sun default on an M-dwarf catalog). ### Model (D-240) - New `[planet_class_temperature]` envelope table in `climate_constants.toml` — the approved bands (`frozen [−90,−25]` … `hot_arid [20,58]` … `volcanic [30,90]`), with `cold_*`/`hot_*`/`warm_*` prefix parsing and a `temperate` fallback. - `derive_temperature_c(params, constants, body_seed)`: band from `planet_class`; **latitude lerps across the band** (equator → warm, pole → cold); atmosphere greenhouse + elevation lapse modulate within; a small **seed nudge** for per-world variety; **clamped to the band** — a body can never derive outside its class. Airless → `None`. ### Removed (the non-canonical inputs, per D-240) - Deleted `derive_distance_au`, `ClimateConstants::luminosity()`, the `[star_luminosity]` table. - Stripped `orbital_period_days` / `spectral_class` / `star_type` / `axial_tilt_deg` from `BodyParams` + `BodyParamsReader` (now reads 3 columns from `bodies`, no `star_systems` join). **DB columns left in place** — no schema change, no `systems.db` regen. ### Verification - New `every_planet_class_derives_within_its_band` test — 13 classes × 5 atmospheres × 19 latitudes × 4 seeds, **all in band** (the D-240 consistency guard, true by construction via the final clamp). - **Acceptance check on 25 diverse real bodies** (lead-run): every world now sits inside its class band — the +356 °C bodies derive to ≤28 °C, frozen → −25/−70, volcanic → 90/58, airless → none, latitude gradient holds everywhere. - `cargo test` passes, `clippy --all-targets -D warnings` clean, `fmt` clean. D-010 integer-disciplined (the f32 temp feeds the integer climate ladders as before). Ticket: T-1033 (epic T-974, decision D-240). After merge, the T-1031 harness rebases onto this and regenerates its golden against the corrected temperatures. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

jpmschweitzer added 1 commit 2026-06-08 18:30:13 +02:00

feat(simulation): planet_class temperature envelope, drop orbit/star inputs (T-1033, D-240) ... b6be9f5758

Replace the sun-driven Stefan-Boltzmann temperature with the D-240 class-envelope
model. Temperature derives only from authored fields — no orbit/star physics.

- [planet_class_temperature] envelope table in climate_constants.toml (approved
  bands: frozen [-90,-25] .. hot_arid [20,58] .. volcanic [30,90]); cold_/hot_/
  warm_ prefix parsing; temperate fallback for unknowns.
- derive_temperature_c(params, constants, body_seed): band = envelope(planet_class);
  latitude lerps across it (equator=warm, pole=cold); atmosphere greenhouse +
  elevation lapse modulate within; small seed nudge for per-body variety; CLAMP to
  band — a body can never escape its class. Airless -> None.
- Delete the dead astrophysics: derive_distance_au, ClimateConstants::luminosity(),
  the [star_luminosity] table. Strip orbital_period_days/spectral_class/star_type/
  axial_tilt_deg from BodyParams + BodyParamsReader (read 3 cols from bodies, no
  star_systems join). DB columns left in place (no regen).
- New every_planet_class_derives_within_its_band test (13 classes x 5 atmo x 19 lat
  x 4 seeds, all in band) — the D-240 consistency guard.

Verified on 25 diverse real bodies: the worlds that derived to +356C now sit inside
their class bands (temperate capped at 28C). cargo test passes, clippy -D warnings
clean, fmt clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

jpmschweitzer referenced this issue from a commit 2026-06-08 18:36:53 +02:00

test+docs(simulation): address PR #164 review (both APPROVE)

jpmschweitzer added 1 commit 2026-06-08 18:36:53 +02:00

test+docs(simulation): address PR #164 review (both APPROVE) ... 54609eab1a

- Band-invariant test now sweeps high altitude (6 km) as well as sea level, so
  the lapse term + cold-end clamp are exercised jointly, not lat alone. (Hoshe #1, Tyre #3)
- Clarify the temp() helper comment: seed 0 is deterministic but still applies a
  constant non-zero nudge. (Hoshe #2)
- (Hoshe #3: confirmed tectonic_activity is genuinely absent from systems-schema.sql;
  the reader's 'not in schema' comment is accurate — no change.)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

jpmschweitzer commented 2026-06-08 18:37:57 +02:00

Author

Owner

Copy Link

Copy Source

Review: class-envelope-temperature → main (type: code) — APPROVED

Reviewers: Hoshe (code quality) + Tyre (architecture). Both APPROVE.

Tyre (Architecture): APPROVE

D-240 faithfully realized. Orbit/star removal complete + clean (BodyParams carries none; derive_distance_au/luminosity() deleted; reader drops the star_systems join, selects 3 columns). Temperature = planet_class envelope + latitude lerp + greenhouse-as-band-fraction + lapse + seed nudge, clamped to [cold,warm] — the band invariant is structural, not conventional. No schema change (correct call). Determinism preserved.

Hoshe (Code Quality): APPROVE

Zero orbit/star data reaches derive_temperature_c; remaining references are only DB-authoring tools/schema, correctly untouched. Clamp makes out-of-band impossible by construction; the parametric sweep is a genuine regression guard.

Tidy-ups applied (54609eab1)

• Band-invariant test now sweeps high altitude (6 km) as well as sea level, exercising lapse + cold-end clamp jointly. (Hoshe #1 / Tyre #3)
• Clarified the temp() helper comment re: seed-0 nudge. (Hoshe #2)
• Confirmed tectonic_activity is genuinely absent from systems-schema.sql — the reader comment is accurate. (Hoshe #3)

One open design question (non-blocking) — Tyre #4

D-240 lists hydrosphere among the canonical modulators ("atmosphere / hydrosphere / elevation modulate within the bounds"). In the implementation, hydrosphere drives moisture (→ precipitation/glaciation/vegetation) but does not directly modulate temperature (e.g. ocean → maritime moderation / reduced swing). Tyre reads D-240's wording as permitting the moisture-mediated interpretation; flagged for the author to confirm intent. Tracked as a possible follow-up, not a blocker.

Verdict: APPROVED

Verified on 25 diverse real bodies — the +356 °C worlds now derive in-band. cargo test passes, clippy -D warnings clean, fmt clean.

🤖 Generated with Claude Code

## Review: class-envelope-temperature → main (type: code) — APPROVED Reviewers: **Hoshe** (code quality) + **Tyre** (architecture). Both **APPROVE**. ### Tyre (Architecture): APPROVE D-240 faithfully realized. Orbit/star removal complete + clean (`BodyParams` carries none; `derive_distance_au`/`luminosity()` deleted; reader drops the `star_systems` join, selects 3 columns). Temperature = `planet_class` envelope + latitude lerp + greenhouse-as-band-fraction + lapse + seed nudge, **clamped to `[cold,warm]`** — the band invariant is structural, not conventional. No schema change (correct call). Determinism preserved. ### Hoshe (Code Quality): APPROVE Zero orbit/star data reaches `derive_temperature_c`; remaining references are only DB-authoring tools/schema, correctly untouched. Clamp makes out-of-band impossible by construction; the parametric sweep is a genuine regression guard. ### Tidy-ups applied (`54609eab1`) - Band-invariant test now sweeps **high altitude (6 km)** as well as sea level, exercising lapse + cold-end clamp jointly. (Hoshe #1 / Tyre #3) - Clarified the `temp()` helper comment re: seed-0 nudge. (Hoshe #2) - Confirmed `tectonic_activity` is genuinely absent from `systems-schema.sql` — the reader comment is accurate. (Hoshe #3) ### One open design question (non-blocking) — Tyre #4 D-240 lists `hydrosphere` among the canonical modulators ("atmosphere / hydrosphere / elevation modulate within the bounds"). In the implementation, hydrosphere drives **moisture** (→ precipitation/glaciation/vegetation) but does **not** directly modulate **temperature** (e.g. ocean → maritime moderation / reduced swing). Tyre reads D-240's wording as permitting the moisture-mediated interpretation; flagged for the author to confirm intent. Tracked as a possible follow-up, not a blocker. ### Verdict: APPROVED Verified on 25 diverse real bodies — the +356 °C worlds now derive in-band. `cargo test` passes, clippy `-D warnings` clean, fmt clean. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

jpmschweitzer added 1 commit 2026-06-08 19:00:08 +02:00

feat(simulation): hydrosphere maritime moderation of temperature (T-1033, D-240) ... b894a721ad

Add the D-240 hydrosphere modulator: water-rich worlds compress the equator->pole
temperature gradient toward the band midpoint (milder at both ends); dry worlds
swing the full class band.

- [hydrosphere_maritime] table in climate_constants.toml + ClimateConstants
  field/method (maritime_factor). Keyed on the ACTUAL bodies.hydrosphere vocab
  (liquid_water/ocean/extensive=0.6, rivers/moderate=0.8, ice=0.85, subsurface*
  =0.9-0.95; minimal/trace/none/NULL -> 1.0).
- derive_temperature_c step 2 now lerps across a maritime-compressed sub-band
  (mid +/- band_half*factor) instead of the full band. Clamp still guarantees the
  class-band invariant.
- New maritime_hydrosphere_compresses_the_gradient test (ocean delta < dry delta).

Verified on real worlds: liquid_water/ocean temperate worlds now delta ~24 (milder)
vs dry ~40, all in-band. cargo test passes, clippy -D warnings clean, fmt clean.

Note: discovered derive_moisture_q has the SAME vocab mismatch (expects 'ocean',
DB uses 'liquid_water' for 175 bodies -> falls to default moisture). Pre-existing,
out of scope here — filing a follow-up.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

jpmschweitzer referenced this issue from a commit 2026-06-08 19:09:56 +02:00

test+docs(simulation): address PR #164 re-review (maritime)

jpmschweitzer added 1 commit 2026-06-08 19:09:56 +02:00

test+docs(simulation): address PR #164 re-review (maritime) ... bafed6ad9a

Re-review of the maritime moderation addition (Hoshe REQUEST_CHANGES, Tyre APPROVE):

- every_planet_class_derives_within_its_band now sweeps hydrosphere
  (none/ocean/liquid_water/rivers/ice) as well, so the maritime-compressed
  gradient path is covered by the band invariant — not just the None/1.0 case.
  (Hoshe #1, Tyre #1)
- Update the BodyParams.hydrosphere doc comment to the real systems.db vocabulary
  (liquid_water/ocean-coastal/extensive/etc.) the maritime table keys on. (Hoshe #2)

cargo test passes (the band sweep now ~59k in-band assertions), clippy -D warnings
clean, fmt clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

jpmschweitzer commented 2026-06-08 19:14:50 +02:00

Author

Owner

Copy Link

Copy Source

Re-review: class-envelope-temperature → main (maritime moderation) — APPROVED

Covers the post-approval addition (b894a721 hydrosphere maritime moderation). The class-envelope core was approved in the prior round.

Tyre (Architecture): APPROVE

The maritime addition soundly completes the D-240 "hydrosphere modulates temperature within bounds" clause. Compressing the gradient toward the class midpoint is the correct model (moderates both ends; the pre-existing clamp catches greenhouse overshoot without widening the band). TOML/embedded in sync; scope clean. Confirmed T-1034 (the derive_moisture_q vocab twin) is correctly filed separately.

Hoshe (Code Quality): REQUEST_CHANGES → resolved

Verified the gradient math (sub-band ⊆ class band, no inversion possible), TOML sync, complete vocab mapping, and that the maritime test is non-trivial (a no-op maritime_factor → 1.0 would fail it). Two contained findings, both fixed.

Findings & resolution (bafed6ad)

#	Reviewer	Finding	Resolution
1	Hoshe + Tyre	every_planet_class_derives_within_its_band only swept hydrosphere: None (maritime=1.0) — the compressed path wasn't covered by the band invariant	Test now sweeps hydrosphere (none/ocean/liquid_water/rivers/ice) too — ~59k in-band assertions covering the maritime path.
2	Hoshe	BodyParams.hydrosphere doc comment listed the stale pre-maritime vocab	Updated to the real systems.db vocabulary the maritime table keys on.

Verdict: APPROVED

Maritime moderation completes D-240's hydrosphere clause, the class-band invariant is now also test-covered for compressed gradients, and the +356 °C → in-band fix is verified across 25 real worlds. cargo test passes, clippy -D warnings clean, fmt clean.

Note: a pre-existing flaky test voice::lookup::cache_miss_returns_base_text (unrelated — this branch doesn't touch voice/) intermittently fails under suite parallelism; passes on retry. Worth a separate CI-hygiene ticket.

🤖 Generated with Claude Code

## Re-review: class-envelope-temperature → main (maritime moderation) — APPROVED Covers the post-approval addition (`b894a721` hydrosphere maritime moderation). The class-envelope core was approved in the prior round. ### Tyre (Architecture): APPROVE The maritime addition soundly completes the D-240 "hydrosphere modulates temperature within bounds" clause. Compressing the gradient toward the **class midpoint** is the correct model (moderates both ends; the pre-existing clamp catches greenhouse overshoot without widening the band). TOML/embedded in sync; scope clean. Confirmed T-1034 (the `derive_moisture_q` vocab twin) is correctly filed separately. ### Hoshe (Code Quality): REQUEST_CHANGES → resolved Verified the gradient math (sub-band ⊆ class band, no inversion possible), TOML sync, complete vocab mapping, and that the maritime test is non-trivial (a no-op `maritime_factor → 1.0` would fail it). Two contained findings, both fixed. ### Findings & resolution (`bafed6ad`) | # | Reviewer | Finding | Resolution | |---|----------|---------|------------| | 1 | Hoshe + Tyre | `every_planet_class_derives_within_its_band` only swept `hydrosphere: None` (maritime=1.0) — the compressed path wasn't covered by the band invariant | Test now sweeps hydrosphere (`none/ocean/liquid_water/rivers/ice`) too — ~59k in-band assertions covering the maritime path. | | 2 | Hoshe | `BodyParams.hydrosphere` doc comment listed the stale pre-maritime vocab | Updated to the real `systems.db` vocabulary the maritime table keys on. | ### Verdict: APPROVED Maritime moderation completes D-240's hydrosphere clause, the class-band invariant is now also test-covered for compressed gradients, and the +356 °C → in-band fix is verified across 25 real worlds. `cargo test` passes, clippy `-D warnings` clean, fmt clean. *Note: a pre-existing flaky test `voice::lookup::cache_miss_returns_base_text` (unrelated — this branch doesn't touch `voice/`) intermittently fails under suite parallelism; passes on retry. Worth a separate CI-hygiene ticket.* 🤖 Generated with [Claude Code](https://claude.com/claude-code)

jpmschweitzer closed this pull request 2026-06-08 19:17:58 +02:00

jpmschweitzer referenced this issue from a commit 2026-06-08 19:18:01 +02:00

chore(meta): T-1033 done (PR #164 merged) — class-envelope temperature

Pull request closed

This pull request cannot be reopened because the branch was deleted.

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

No items

No labels

Milestone

No items

No Milestone

Projects

Clear projects

No projects

Assignees

Clear assignees

jpmschweitzer (Jeroen Schweitzer)

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: jpmschweitzer/settled-reach#164