From 5f65fa0547a202e4d142247d8953e790342dd745 Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Tue, 8 Sep 2026 11:23:43 -0400 Subject: [PATCH 1/3] cortex: birth slope_hierarchy_scale (graphical-ep phase 3); retire slope_hierarchy MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The `graphical-ep` campaign's science half gets its first Cortex science project. - projects.yaml: new `slope_hierarchy_scale` row — active, `assistant: autolens_assistant`, local `/mnt/c/Users/Jammy/Science/slope_hierarchy_scale`, RAL `/mnt/ral/jnightin/slope_hierarchy_scale`, ledger `wiki/project/state.md`, `remote: none` until a human creates the private org repo. - projects.yaml: `slope_hierarchy` retired (`cortex.py retire`) — it answered its four N=5 goals and was wrapped up 2026-07-22 — and its `local_path` hand-edited to the vault, `/mnt/c/Users/Jammy/Science/z_projects_complete/slope_hierarchy`, so the row still says where the N=5 data lives. The tree itself was `mv`d intact, git history and remote untouched. - tasks: `git mv slope_hierarchy/n25_scale_up.md slope_hierarchy_scale/`, `Project:` retargeted, `Gates: PyAutoFit#1405, #1558, #1560, #1562`, the stale "Ready when …" line replaced by the phase-2 gate evidence and a "Borrowed from" note, then `cortex.py move … gated`. - dashboard.md / dashboard.html re-rendered. `tasks/slope_hierarchy/methods_writeup.md` stays `planned` under the retired row (a retired project keeps planned tasks) — flagged for the human to move or drop. cortex check OK; pytest 126 passed; dashboard --check exit 0. Closes #24 Co-Authored-By: Claude Fable 5.1 Claude-Session: https://claude.ai/code/session_01JcyWRpcbRiAAYFqgFJarEr --- dashboard.html | 14 ++++++-- dashboard.md | 32 +++++++++++++------ projects.yaml | 20 ++++++++++-- .../n25_scale_up.md | 32 ++++++++++++++----- 4 files changed, 75 insertions(+), 23 deletions(-) rename tasks/{slope_hierarchy => slope_hierarchy_scale}/n25_scale_up.md (68%) diff --git a/dashboard.html b/dashboard.html index b5d6b6d..8933bae 100644 --- a/dashboard.html +++ b/dashboard.html @@ -237,6 +237,7 @@ +
ProjectNext taskAwaitingRunningLast ruling
euclid_dr1_prelimTen real DR1 lenses fitted from the pipeline repo alone01-
slope_hierarchy_scaleDoes hierarchical slope recovery hold at N=25 to 5000-
subhalo_validationDoes rectangular_adapt false-detect a subhalo on the clumpy source052026-09-07

Nothing open

@@ -248,11 +249,11 @@

Nothing open

ic50_workspacedormantplanned 1 pj011646dormantnone profilingdormantnone -slope_hierarchydormantplanned 2 subhalo_simulationsdormantnone -
1 retired
    +
    2 retired
    • inference_programme — retired 2026-09-07: restarting from scratch: the project setup was unsatisfactory and it informed how the Cortex manages projects; RAL outputs stashed at /mnt/ral/jnightin/inference_programme_retired_2026-09-07
      • +
    • slope_hierarchy — retired 2026-09-08: wrapped up 2026-07-22 at N=5; succeeded by slope_hierarchy_scale; tree vaulted under z_projects_complete/

    Projects markdown version

    @@ -264,6 +265,15 @@

    euclid_dr1_prelim

    tasks/euclid_dr1_prelim/

    +

    slope_hierarchy_scale

    +

    Local /mnt/c/Users/Jammy/Science/slope_hierarchy_scale

    +

    RAL /mnt/ral/jnightin/slope_hierarchy_scale

    +

    active · both partition · tasks: gated 1

    +

    Does hierarchical slope recovery hold at N=25 to 50gated — PyAutoFit#1405, PyAutoFit#1558, PyAutoFit#1560, PyAutoFit#1562

    +

    tasks/slope_hierarchy_scale/

    +
    +

    subhalo_validation

    Local /mnt/c/Users/Jammy/Science/subhalo_validation

    RAL /mnt/ral/jnightin/subhalo_validation

    diff --git a/dashboard.md b/dashboard.md index 303e97e..228940b 100644 --- a/dashboard.md +++ b/dashboard.md @@ -30,6 +30,7 @@ One row per active project — the next thing, what it is holding, and when it l | Project | Next task | Awaiting | Running | Last ruling | |---|---|---:|---:|---| | euclid_dr1_prelim | Ten real DR1 lenses fitted from the pipeline repo alone | 0 | 1 | - | +| slope_hierarchy_scale | Does hierarchical slope recovery hold at N=25 to 50 | 0 | 0 | - | | subhalo_validation | Does rectangular_adapt false-detect a subhalo on the clumpy source | 0 | 5 | 2026-09-07 | #### Nothing open @@ -42,7 +43,6 @@ One row per active project — the next thing, what it is holding, and when it l | ic50_workspace | dormant | planned 1 | retire ↓ | | pj011646 | dormant | none | retire ↓ | | profiling | dormant | none | retire ↓ | -| slope_hierarchy | dormant | planned 2 | retire ↓ | | subhalo_simulations | dormant | none | retire ↓ |
    📋 retire concr @@ -93,14 +93,6 @@ One row per active project — the next thing, what it is holding, and when it l
    -
    📋 retire slope_hierarchy - -``` -/cortex — retire the science project slope_hierarchy: confirm it is holding nothing live, then in PyAutoCortex run `python3 scripts/cortex.py retire slope_hierarchy --why ""`, `python3 scripts/cortex.py check`, `pyauto-brain cortex dashboard --apply`, and push the ledger. The row, its tasks and its rulings all stay — only the status and the note change. -``` - -
    -
    📋 retire subhalo_simulations ``` @@ -109,9 +101,10 @@ One row per active project — the next thing, what it is holding, and when it l
    -
    1 retired +
    2 retired - inference_programme — retired 2026-09-07: restarting from scratch: the project setup was unsatisfactory and it informed how the Cortex manages projects; RAL outputs stashed at /mnt/ral/jnightin/inference_programme_retired_2026-09-07 +- slope_hierarchy — retired 2026-09-08: wrapped up 2026-07-22 at N=5; succeeded by slope_hierarchy_scale; tree vaulted under z_projects_complete/
    @@ -137,6 +130,25 @@ Report where the runs of the PyAutoCortex task tasks/euclid_dr1_prelim/dr1_preli [tasks/euclid_dr1_prelim/](https://github.com/PyAutoLabs/PyAutoCortex/tree/main/tasks/euclid_dr1_prelim/) +### slope_hierarchy_scale + +**Local** `/mnt/c/Users/Jammy/Science/slope_hierarchy_scale` + +**RAL** `/mnt/ral/jnightin/slope_hierarchy_scale` + +active · both partition · tasks: gated 1 + +
    📋 Does hierarchical slope recovery hold at N=25 to 50 — gated — PyAutoFit#1405, PyAutoFit#1558, PyAutoFit#1560, PyAutoFit#1562 + +``` +python3 scripts/cortex.py gates # then, once they have closed: move tasks/slope_hierarchy_scale/n25_scale_up.md ready +# gates: PyAutoFit#1405, PyAutoFit#1558, PyAutoFit#1560, PyAutoFit#1562 +``` + +
    + +[tasks/slope_hierarchy_scale/](https://github.com/PyAutoLabs/PyAutoCortex/tree/main/tasks/slope_hierarchy_scale/) + ### subhalo_validation **Local** `/mnt/c/Users/Jammy/Science/subhalo_validation` diff --git a/projects.yaml b/projects.yaml index d6891e0..cc046b4 100644 --- a/projects.yaml +++ b/projects.yaml @@ -151,7 +151,7 @@ subhalo_simulations: slope_hierarchy: remote: Jammy2211/slope_hierarchy - local_path: /mnt/c/Users/Jammy/Science/slope_hierarchy + local_path: /mnt/c/Users/Jammy/Science/z_projects_complete/slope_hierarchy ral_root: /mnt/ral/jnightin/slope_hierarchy mirror: none sync_cli: hpc/sync @@ -160,8 +160,22 @@ slope_hierarchy: assistant: none witness_file: results/**/*.json partition: both - status: dormant - note: "personal remote, recorded as a fact — not a PyAutoLabs repo; ral_root planned, sync_verbs unverified" + status: retired + note: "retired 2026-09-08: wrapped up 2026-07-22 at N=5; succeeded by slope_hierarchy_scale; tree vaulted under z_projects_complete/" + +slope_hierarchy_scale: + remote: none + local_path: /mnt/c/Users/Jammy/Science/slope_hierarchy_scale + ral_root: /mnt/ral/jnightin/slope_hierarchy_scale + mirror: none + sync_cli: hpc/sync + sync_verbs: [push, pull, sync, status, submit, push-submit, jobs, tail, check] + ledger: wiki/project/state.md + assistant: autolens_assistant + witness_file: results/**/*.json + partition: both + status: active + note: "born 2026-09-08 as graphical-ep phase 3, code borrowed from slope_hierarchy (vaulted under z_projects_complete/); remote pending a human gh repo create" pj011646: remote: Jammy2211/PJ011646 diff --git a/tasks/slope_hierarchy/n25_scale_up.md b/tasks/slope_hierarchy_scale/n25_scale_up.md similarity index 68% rename from tasks/slope_hierarchy/n25_scale_up.md rename to tasks/slope_hierarchy_scale/n25_scale_up.md index 42266e8..d104ac8 100644 --- a/tasks/slope_hierarchy/n25_scale_up.md +++ b/tasks/slope_hierarchy_scale/n25_scale_up.md @@ -1,9 +1,9 @@ -# Slope_hierarchy — scale the hierarchical slope recovery to N=25–50 +# slope_hierarchy_scale — scale the hierarchical slope recovery to N=25–50 -Project: slope_hierarchy +Project: slope_hierarchy_scale Summary: Does hierarchical slope recovery hold at N=25 to 50 -State: planned -Gates: +State: gated +Gates: PyAutoFit#1405, PyAutoFit#1558, PyAutoFit#1560, PyAutoFit#1562 Witness: an N=25–50 parity table in the same shape as the N=5 one, committed under `results/`, plus the scaling measurements (VRAM ceiling, JAX compile time vs model size, sampler ladder at 3N+1 dims, gradient-utilisation sanity) Budget: 48:00 Runs: @@ -15,7 +15,20 @@ Migrated-from: PyAutoMind/draft/research/graphical_ep/slope_hierarchy_n25_scale_ ## Question -Ready when: graphical-ep phase 2 (Mind) is issued — add its ref to `Gates:` and move to `gated`. +Gate evidence (phase 2 of the `graphical-ep` epic is issued and shipped): the Mind record +`PyAutoMind/complete/2026/09/ep-scale-collapse-basin-cure-or-caveat.md`, the phase-2 review +bundle PyAutoFit#1572 / #1573 / #1574 / #1576, and the stale-mask fixed-point fix +PyAutoFit#1580. `Gates:` carries PyAutoFit#1405 — the collapse-basin umbrella, deliberately +still open pending the moment-matching cure decision — alongside PyAutoFit#1558, #1560 and +#1562, the phase-2 EP defects this scale-up rests on, so `gates` surfaces all four for the +human's `move … ready`. + +Borrowed from: this task's project was reborn 2026-09-08 as `slope_hierarchy_scale` +(`/mnt/c/Users/Jammy/Science/slope_hierarchy_scale`), scaffolded fresh with the code borrowed +from `slope_hierarchy`, which was wrapped up at N=5 on 2026-07-22 and vaulted intact at +`/mnt/c/Users/Jammy/Science/z_projects_complete/slope_hierarchy`. The N=5 baseline of record +is `results/ep_history_n5_maxsteps12/` inside that vault; the datasets are **re-simulated** at +N=25 in the new tree, never copied. The `slope_hierarchy` project answered all four of its goals at N=5 and was wrapped up on 2026-07-22. This is the *optional* scale-up that was left on the table, not a blocker or an open @@ -51,8 +64,10 @@ CPU `ral` partition for the EP arm (finiteness, not throughput). ## Witness -An N=25–50 parity table in the same shape as the N=5 one, committed under `results/`, plus an -issue comment on slope_hierarchy#1. +An N=25–50 parity table in the same shape as the N=5 one, committed under `results/` in +`slope_hierarchy_scale`. (The old N=5 witness was commented on Jammy2211/slope_hierarchy#1; +the new tree has no remote yet, so the committed table is the whole witness until a human +runs `gh repo create`.) Scaling measurements to capture while the runs are up anyway (the EP campaign needs them): @@ -70,7 +85,8 @@ Each number goes in the committed results table beside the parity numbers. ## Where to look -- `slope_hierarchy` (project row): `results/ep_history_n5_maxsteps12/` — the N=5 baseline +- `slope_hierarchy` (retired project row, vaulted at `/mnt/c/Users/Jammy/Science/z_projects_complete/slope_hierarchy`): `results/ep_history_n5_maxsteps12/` — the N=5 baseline +- `slope_hierarchy_scale` (this project's row): `results/` — where the N=25 parity table and the scaling measurements land - Jammy2211/slope_hierarchy#1 — the project issue the parity table is commented on - Mind `research/graphical_ep/ep_campaign.md` phase 3 — the campaign rows this feeds - Mind `draft/research/graphical_ep/slope_hierarchy_methods_writeup.md` — the write-up it feeds From 3e3e5201e404e830285cbb596f443f582ed50478 Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Tue, 8 Sep 2026 11:57:19 -0400 Subject: [PATCH 2/3] =?UTF-8?q?cortex:=20slope=5Fhierarchy=5Fscale=20remot?= =?UTF-8?q?e=20=E2=86=92=20PyAutoLabs/slope=5Fhierarchy=5Fscale?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The human created the private org repo and pushed the birth commit. Co-Authored-By: Claude Fable 5.1 Claude-Session: https://claude.ai/code/session_01JcyWRpcbRiAAYFqgFJarEr --- projects.yaml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/projects.yaml b/projects.yaml index cc046b4..332c74c 100644 --- a/projects.yaml +++ b/projects.yaml @@ -164,7 +164,7 @@ slope_hierarchy: note: "retired 2026-09-08: wrapped up 2026-07-22 at N=5; succeeded by slope_hierarchy_scale; tree vaulted under z_projects_complete/" slope_hierarchy_scale: - remote: none + remote: PyAutoLabs/slope_hierarchy_scale local_path: /mnt/c/Users/Jammy/Science/slope_hierarchy_scale ral_root: /mnt/ral/jnightin/slope_hierarchy_scale mirror: none @@ -175,7 +175,7 @@ slope_hierarchy_scale: witness_file: results/**/*.json partition: both status: active - note: "born 2026-09-08 as graphical-ep phase 3, code borrowed from slope_hierarchy (vaulted under z_projects_complete/); remote pending a human gh repo create" + note: "born 2026-09-08 as graphical-ep phase 3, code borrowed from slope_hierarchy (vaulted under z_projects_complete/); private remote created 2026-09-08" pj011646: remote: Jammy2211/PJ011646 From bff7124805a62ef89469d43b934c7b98009ddca7 Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Tue, 8 Sep 2026 11:57:40 -0400 Subject: [PATCH 3/3] cortex: re-render board with the slope_hierarchy_scale remote Co-Authored-By: Claude Fable 5.1 Claude-Session: https://claude.ai/code/session_01JcyWRpcbRiAAYFqgFJarEr --- dashboard.html | 2 +- dashboard.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/dashboard.html b/dashboard.html index 8933bae..461723b 100644 --- a/dashboard.html +++ b/dashboard.html @@ -271,7 +271,7 @@

    slope_hierarchy_scale

    active · both partition · tasks: gated 1

    Does hierarchical slope recovery hold at N=25 to 50gated — PyAutoFit#1405, PyAutoFit#1558, PyAutoFit#1560, PyAutoFit#1562

    -

    tasks/slope_hierarchy_scale/

    +

    tasks/slope_hierarchy_scale/ · issues

    subhalo_validation

    diff --git a/dashboard.md b/dashboard.md index 228940b..fcef4db 100644 --- a/dashboard.md +++ b/dashboard.md @@ -147,7 +147,7 @@ python3 scripts/cortex.py gates # then, once they have closed: move tasks/slop
-[tasks/slope_hierarchy_scale/](https://github.com/PyAutoLabs/PyAutoCortex/tree/main/tasks/slope_hierarchy_scale/) +[tasks/slope_hierarchy_scale/](https://github.com/PyAutoLabs/PyAutoCortex/tree/main/tasks/slope_hierarchy_scale/) · [issues](https://github.com/PyAutoLabs/slope_hierarchy_scale/issues) ### subhalo_validation