Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@ Optional provider configuration is described in `.env.example` and `app/llm/`. K

Use **Settings → Check inputs** or `POST /api/validate` with `{"model_id":"..."}` or `{"model":{...}}`. Inspection covers names, references, syntax, initial values, flow endpoints, tables, settings and a short numerical run. Units and provenance are retained. Scientific validity is not established by these checks or by finite simulation output.

The curated public library contains **996 examples: 993 dynamic models and three explicitly labeled static calculators**. Curation retired 155 failing, incomplete, degenerate or redundant definitions from the served library while preserving their exact bytes as test evidence. Five flowless examples were reconstructed from attributable equations and checked independently. Every active example must complete its full default run; numerical completion is not empirical validation. See [curation decisions](docs/LIBRARY-CURATION.md), [reconstructions](docs/MODEL-RECONSTRUCTIONS-2026-10.md), and the historical [failure register](docs/FAILURE-REGISTER-1.9.md). Additional user models are preserved during release.
The curated public library contains **1000 examples: 997 dynamic models and three explicitly labeled static calculators**. Curation retired 155 failing, incomplete, degenerate or redundant definitions from the served library while preserving their exact bytes as test evidence. Five flowless examples were reconstructed from attributable equations and checked independently. Every active example must complete its full default run; numerical completion is not empirical validation. See [curation decisions](docs/LIBRARY-CURATION.md), [reconstructions](docs/MODEL-RECONSTRUCTIONS-2026-10.md), and the historical [failure register](docs/FAILURE-REGISTER-1.9.md). Additional user models are preserved during release.

## API and MCP

Expand Down
4 changes: 2 additions & 2 deletions docs/LIBRARY-ADMISSIONS-2026-10.md
Original file line number Diff line number Diff line change
@@ -1,8 +1,8 @@
# Reviewed model admissions — October 2026

The public source library contains 996 models: 993 dynamic examples and three explicitly static calculators. Four batches add 80 independently reviewed examples to the 923-model curated baseline. Source-quality reviews retire seven unsupported or redundant definitions in total, including two May parameter variants, and replace reviewed canonical examples: 923 + 80 - 7 = 996. The 26 later citation corrections change descriptions only. Historical admissions, the earlier five quality retirements and the original 155 retirements retain their exact evidence and archives. Model count does not justify duplicates, unsupported equations or nominal variations. Local source adoption is separate from public deployment; see the May migration document and quality-follow-up manifest.
The public source library contains 1000 models: 997 dynamic examples and three explicitly static calculators. Six batches add 84 independently reviewed examples to the 923-model curated baseline. Source-quality reviews retire seven unsupported or redundant definitions in total, including two May parameter variants, and replace reviewed canonical examples: 923 + 84 - 7 = 1000. The 26 later citation corrections change descriptions only. Historical admissions, the earlier five quality retirements and the original 155 retirements retain their exact evidence and archives. Model count does not justify duplicates, unsupported equations or nominal variations. Local source adoption is separate from public deployment; see the May migration document and quality-follow-up manifest.

The additions comprise five mechanistic numerical benchmarks, twenty biological systems, thirty-one feedback/control examples, ten environmental process-law adaptations and fourteen further physical/scientific examples. Each admitted file is an exact copy of its independently reviewed candidate, with a SHA-256 approval and source path in `operations/library-admissions-20261024.json`. The second twenty are separately recorded in `operations/library-admissions-second-20261024.json`. Candidate source, license notices, numerical evidence and independent review are retained alongside each batch. Admission tests enforce exact bytes and IDs, and cross-batch duplicate checks exclude only the candidate being checked.
The additions comprise five mechanistic numerical benchmarks, twenty biological systems, thirty-one feedback/control examples, ten environmental process-law adaptations, fourteen further physical/scientific examples, two ecological/adsorption systems and two finite-state/control systems. Each admitted file is an exact copy of its independently reviewed candidate, with a SHA-256 approval and source path in `operations/library-admissions-20261024.json`. The second twenty are separately recorded in `operations/library-admissions-second-20261024.json`. Candidate source, license notices, numerical evidence and independent review are retained alongside each batch. Admission tests enforce exact bytes and IDs, and cross-batch duplicate checks exclude only the candidate being checked.

The numerical checks use handwritten independent derivatives, SciPy reference solvers, matrix exponentials, exact recurrences, analytic equilibria, conservation identities and step refinement where applicable. These tests check the described mathematics and implementation; they do not establish empirical accuracy or suitability for forecasting. Illustrative parameters, simplified closures, inaccessible original sources and incomplete unit metadata remain disclosed in the batch documentation.

Expand Down
18 changes: 18 additions & 0 deletions docs/LIBRARY-FIFTH-SIXTH-ADMISSIONS-2026-10.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
# Four reviewed systems and control admissions

Root authorized local source admission of the exact two fifth-systems and two sixth-control candidates on October 24, 2026. The canonical catalog now contains 1,000 definitions: 997 dynamic examples and three explicitly static calculators. Accounting is 923 curated baseline + 84 reviewed admissions - 7 later retirements = 1,000. Local adoption is separate from public publication or deployment.

| Example | Mathematical scope and limits |
| --- | --- |
| Holt shared predator | Two logistic prey share one predator; the conversion factor applies to recruitment and maintenance. Normalized densities and defaults are illustrative. This is a shared-enemy interaction, not fitted ecological prediction. |
| Three-stage adsorption breakthrough | Three authored constant-volume mixed cells with reversible equilibrium-loading-based linear driving force and single-component Langmuir equilibrium. It is neither a plug-flow discretization claim nor a CADET replication. Positive volumes and nonnegative physical parameters/capacity-bounded initial loading are required. |
| Finite-source machine repair | Four ensemble probabilities for three independent exponential-failure machines and one repair server. Rates depend on operating machines; waiting states remain. It is not an event simulation or an aging/common-cause model. Initial probabilities must be nonnegative and sum to one. |
| PI back-calculation | Signed PI integral contribution with actual-minus-command actuator tracking and an illustrative first-order plant. Saturation, antiwindup and switching are explicit. It is not a calibration or stability guarantee for arbitrary edited gains, delays or physical plants. |

Primary equations, rights/access qualifications, original defaults, units, deduplication and independent full-series/analytic/invariant checks remain in the two frozen research packages. Their original pending-admission status describes the historical review stage. No source PDF, figure or upstream software is redistributed. No previous candidate manifest, numerical receipt, retirement archive or admission receipt was rewritten. Finite numerical completion is not empirical validation; coarse edited timesteps may violate positivity even when the continuous equations preserve it.

The new transaction manifest `operations/library-admissions-fifth-sixth-20261024.json` binds every one of the previous 996 model hashes, the four exact candidate/active copies and both frozen review manifests. The exact pre-admission cumulative ledger is retained as a hash-bound test fixture. All 80 earlier admission rows and every other ledger field remain unchanged except the new four rows and admitted/active totals. No runtime module or user definition is changed. Editor/library interfaces retain their established features and change count text only.

The new test collector runs the frozen mathematical checks unchanged. The sixth-batch historical duplicate screen is reproduced against its hash-bound 996-model baseline, avoiding newly admitted self-copies; cross-batch candidate screening is retained. Historical follow-up/fourth-admission accounting uses the exact frozen 80-row ledger and still independently checks current models through the new four-entry transaction. No historical count/assertion or original receipt is weakened.

All 1,000 tracked default runs passed in 93.66 seconds with fallback seed 42 against the frozen 22 core modules, including parser SHA256 72070b639fd2fe664d713a6cf4b53cf25e9dc653692e3ed028f84e97173c9e56. Full private receipt SHA256 3f3cb61719348a08e9d546f697bad29fddb79b8a369f55c2c86161276d7f0efd; compact receipt SHA256 dfa824143d70d13b76e17a9dd59dab59578d10c3d146b8ff0c48346194825096; minimal public validation receipt SHA256 200cca04903c9ad03c051c89d04d58aa4ef05f49d9a80dfa5feab285dea4825e. The compact receipt retains per-model path/hash/status and the full receipt digest; its full_receipt_private_path field is private metadata and must be omitted from a separately reviewed public export. All 71 focused candidate/adoption/history checks passed in 21.21 seconds. Final full suite passed: 3,216 passed, one skipped, three known expected failures, 18 passing subtests and seven warnings in 381.13 seconds. Independent review additionally ran 72 focused checks (19.84 seconds) and verified all prior 996 model bytes, 80 ledger rows, source manifests and complete audit projections. See LIBRARY-FIFTH-SIXTH-ADOPTION-REVIEW.md. This records local source adoption; public publication and deployment are separate steps.
38 changes: 38 additions & 0 deletions docs/LIBRARY-FIFTH-SIXTH-ADOPTION-REVIEW.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
# Independent review: four exact admissions to canonical1,000

Reviewed October24,2026. Approve the exact local four-create adoption and accounting package described below, subject to the still-running full-suite passing without runtime/model/test changes. This approval does not publish or deploy; public PR11 remains a separate996-model parser release and live data were not accessed or modified by this review.

I independently read the original committed model snapshot at the pre-admission committed baseline using Git archive, rather than trusting the proposed before-hash map. All996 paths and bytes match that independent snapshot and remain exact on disk. The only model differences are the four approved creates. Every new active file is byte-identical to its frozen research original and exact prior reviewed hash; both research manifests and all18 bound original assets remain unchanged. All1,000 current model IDs are unique. No app/core, other app module, script or dependency change belongs to this adoption.

I independently retrieved the pre-admission cumulative ledger from that committed base. The new historical fixture matches its exact bytes and SHA2569cbb8c583e3aa369828616a3481f88e444ce47a01876d7ea7eb5828905224812. All80 old rows and all other fields are unchanged; only the exact four rows and admitted84/active1000 totals are added. Accounting remains923+84-7=1000. Exactly three models remain explicitly labeled static calculators; the four new examples are dynamic. Static editor and companion-library HTML match the prior files after only996-to1000 count substitutions; README changes only996/993 to1000/997. Existing controls/navigation/features remain intact.

The new collector runs the frozen mathematical tests without rewriting them. Only the sixth-batch duplicate test that would compare its new active self-copies is replaced by the same expanded/constant-abstracted algorithm against the bound historical996 catalog, still including prior/fifth candidate cross-checks. The fifth duplicate check already uses its frozen996 inventory. Fourth-admission/follow-up tests retain the exact77/993 and80/996 historical ledger assertions and now separately bind the current four-create transaction; no historical count or evidence assertion is weakened. The new exact-adoption test protects old bytes, old ledger fields, source/review hashes and the entire1000-file set.

Independently ran the read-only adoption verification checks and `.venv/bin/python -m pytest -q tests/test_library_candidates_fifth_sixth.py tests/test_quality_followup.py tests/test_public_library_validation.py`:72passed in19.84s. The integration tests retain analytical equilibria/invariants/limits,120 random independent Euler RHS cases, varied matrix-exponential repair trajectories, saturated/unsaturated exact PI curves, DOP853 switching comparisons and step refinement. Prior independent peer evidence remains bound by the unchanged candidates; no calibration or universal positivity/stability claim is made.

Verified all1000 rows in the author's complete default audit against the current exact file paths/IDs/hashes/statuses, all22 engine hashes against current disk, and parser72070b639fd2fe664d713a6cf4b53cf25e9dc653692e3ed028f84e97173c9e56. Full audit1000passed93.66s with fallback seed42 was not redundantly rerun by this reviewer. Its SHA256 is3f3cb61719348a08e9d546f697bad29fddb79b8a369f55c2c86161276d7f0efd. Compact rows/metadata and full-receipt digest match exactly. Public receipt is the exact minimal ordered projection; all996 old public rows/order and all22 engine bindings are unchanged. It contains no private path or model bodies. The compact private receipt's full_receipt_private_path must be omitted from any separately reviewed public export.

New full-default point counts are Holt3201, adsorption1501, repair1501, PI2001. All four change their intended states. The full private audit honestly retains generic all_stocks_zero warnings for initially empty adsorption and zero-start PI: feed and reference/error drive those systems away from zero, as the analytical/full-series checks show. No errors or unit issues are reported for these four. The warnings are not suppressed or converted into empirical-validation claims.

One nonblocking admission-doc comma typo was reported and corrected by its author. Documentation may append final full-suite/release evidence; any change to the exact model/runtime/ledger/test/audit bytes below needs a fresh review. Remaining gates are the author full suite, root approval, separately reviewed public packaging, exact CI and release checks. No Git index, public checkout or remote operation was changed by this read-only adoption review.

## Reviewed file hashes

- README.md: 7157dac01d804d44eab2280555d18e6f08d365bc1a584b74879deb12072d6c07
- docs/LIBRARY-ADMISSIONS-2026-10.md: 681a879db84a05e491d13333ef9cb8df9f76fc4dcc2b3be86ebb2f230e25e36f
- models/ecology/holt-shared-predator-fifth-systems.json: 7866c48ce562e80c6335d4f6e796e00c3b6e88656713deef3d009d781cddc6c6
- models/engineering/finite-source-machine-repair-sixth-control.json: 340761ca086934eae01c9efa199e94fa079494d225f4c4cb9739441147d10dac
- models/engineering/pi-back-calculation-sixth-control.json: 1b5a386e5e9463d83dc2b42de0d1782e08dcae37d84339ae4613d368970f6837
- models/engineering/three-stage-adsorption-breakthrough-fifth-systems.json: f27abdf55694a3794f67f93bb133dcad3575cba1456a1eb03819a3c9dfa8fa9b
- operations/library-admissions-20261024.json: 639d24fcb4a00e2d8f3b205474a0830b3bc8dacb3754e314adebc35e1710d1c6
- sites/library/index.html: 6cda1db2fcc2cec480f5f195a38c88ac08ba8fa05edc5bb9481129e05adf14f1
- static/index.html: dba894503d736240bfd4f8e945ff5f99771fa71d9388f3fdd01e6af6f80ca81d
- static/library-validation.json: 200cca04903c9ad03c051c89d04d58aa4ef05f49d9a80dfa5feab285dea4825e
- tests/test_public_library_validation.py: e06d14c132381f841acaa848e16e3c0f915d8767cd76253163b052417cef71a9
- tests/test_quality_followup.py: 7ee957df84b88f4451903a1b8ffcc07429c6a3276ee3dcd3f904757896da5b5d
- operations/library-admissions-fifth-sixth-20261024.json: 3defc87d3ff4059129ecfe4b6c53ac322b2a3d9d5d54d9b1f802908e201419fc
- operations/library-fifth-sixth-full-audit-20261024.json: 3f3cb61719348a08e9d546f697bad29fddb79b8a369f55c2c86161276d7f0efd
- operations/library-fifth-sixth-audit-20261024.json: dfa824143d70d13b76e17a9dd59dab59578d10c3d146b8ff0c48346194825096
- tests/fixtures/fifth-sixth-admissions-20261024/ledger-before.json: 9cbb8c583e3aa369828616a3481f88e444ce47a01876d7ea7eb5828905224812
- tests/test_library_candidates_fifth_sixth.py: 024dd433a9c7f1e6b7167159d034b622b80c88ba5db59e3e682250ead87fe906
- docs/LIBRARY-FIFTH-SIXTH-ADMISSIONS-2026-10.md: 9b2ab951f3af39a7b65ba3a8c86673e7fff40ffd1397edddbcacf68fd9a82ea6
Loading
Loading