diff --git a/README.ja.md b/README.ja.md index 8517e5c..0f59849 100644 --- a/README.ja.md +++ b/README.ja.md @@ -6,7 +6,7 @@ - Input: GCPs plus a single CRS - Output: `world_file_basic` and post-shear refinement -- Runtime: Node.js と browser。browser は**行ごとに実績が違う**ため 3 段に分けて書く（m3-t1）。**(1) バンドル可能性 = 検証済**: Teiresias の `pnpm run verify:demo-bundle`（`scripts/verify-demo-browser-bundle.mjs`）が demo build の出力チャンクを走査し、Node 固有指定子が残っていないことと Daedalus 由来コードが実在することを機械検証する（CI から継続的に走る）。**(2) 実ブラウザでの実行 = 観測済（「検証済」ではない）**: 2026-08-29 に Teiresias デモを実ブラウザで **1 回観測**した（m3-t1/AC-16）。1 回の観測は自動検査ではない ∴ 回帰を守らない。**(3) CI 上での継続的・自動のブラウザ実行と canvas ドラッグの自動操作 = 未検証**（Cycle 1 の既知欠落）。`Daedalus/src` は本件のために 1 バイトも変更していない（m0 凍結を解いていない） +- Runtime: Node.js とブラウザ。ブラウザでの対応は 3 段に分けて書きます。**(1) バンドルできること = 検証済み**: Teiresias の `pnpm run verify:demo-bundle`（`scripts/verify-demo-browser-bundle.mjs`）がデモのビルドの出力チャンクを走査し、Node 固有の指定子が残っていないことと Daedalus 由来のコードが実在することを機械で検証します（CI から継続的に走ります）。**(2) 実ブラウザでの実行 = 手動で 1 回確認済み**: 2026-08-29 に Teiresias のデモを実ブラウザで動かして確かめました。1 回の手動確認は自動テストではないため、回帰は防げません。**(3) CI での継続的・自動のブラウザ実行 = 未対応**（予定。[#4](https://github.com/code4history/Daedalus/issues/4) を参照）。 - Scope: fitting only, not CRS candidate lookup English README is [here](./README.md). @@ -81,7 +81,7 @@ affine の残差自由度は `df2 = 2n-6` である。採用は次の**両方** `holdout` を指定した場合はさらに `holdoutRmsAffine <= holdoutRmsSimilarity` を要求する。 -**SSE の数値的ゼロ**は厳密比較ではなく許容つきで判定する（仕様正本の「数値許容内の厳密適合」）。 +**SSE の数値的ゼロ**は厳密比較ではなく許容つきで判定する（「数値許容内の厳密適合」）。 ``` epsSSE = 1024 · Number.EPSILON² · Σ(X² + Y²) （和は training GCP の geo 座標） @@ -92,7 +92,7 @@ Z(s) = (s <= epsSSE) - `Z(SSE6)` かつ `¬Z(SSE4)`: `F = +Infinity` として F gate を通す `alpha = 0.05` / `minRelativeImprovement = 0.01` / 安全係数 `1024` は実データで較正した値ではなく、 -Cycle 1 の過適合抑制のための初期ヒューリスティック（仮仕様）である。 +過適合抑制のための初期ヒューリスティックである。 ### `debug`（`options.debug` が真のときのみ） @@ -104,21 +104,14 @@ Cycle 1 の過適合抑制のための初期ヒューリスティック（仮仕 | `fStatistic` / `criticalF` / `relativeImprovement` | F gate を評価したとき | | `holdoutRmsSimilarity` / `holdoutRmsAffine` | `holdout` を指定して検証に到達したとき | -### 非目標（Cycle 1） +### 非目標 + +次のものには対応していません。将来の版の候補として [#3](https://github.com/code4history/Daedalus/issues/3)（検証・重み・ロバスト推定・広域など）と [#4](https://github.com/code4history/Daedalus/issues/4)（CI でのブラウザ実行）で扱います。 - LOO（leave-one-out）validation（未指定 holdout は `validationSkipped`） - 一般形 `criticalF`（df1≠2 の不完全ベータ） - weight / robust fitting / 逆写像 RMS の高度化 / アンチメリディアン・極域 - -## 関連文書 - -- [CRS 推定ライブラリ群 仮仕様正本](../docs/superpowers/specs/2026-08-27-c1-crs-libs-provisional-spec.md) — - 本 package の観測契約・数値挙動が集約されている（**仮仕様**であり人間承認済みの確定仕様ではない）。 -- [mock transformer 契約文書](../docs/superpowers/specs/2026-08-28-mock-transformer-contract.md) — - `src/testkit/` が提供する mock adapter の契約。 - -上記リンクは**親リポジトリ（monorepo）を checkout したときにのみ解決する**。 -本 repository を単体で clone した場合、`../docs/` は存在しない。 +- CI での継続的・自動のブラウザ実行（上の「Runtime」を参照） ## ライセンス diff --git a/README.md b/README.md index 60c6027..0e0684c 100644 --- a/README.md +++ b/README.md @@ -6,7 +6,7 @@ - Input: GCPs plus a single CRS - Output: `world_file_basic` and post-shear refinement -- Runtime: Node.js and browser. The browser story differs line by line, so it is written in three tiers (m3-t1). **(1) Bundlability = verified**: Teiresias' `pnpm run verify:demo-bundle` (`scripts/verify-demo-browser-bundle.mjs`) scans the output chunks of the demo build and mechanically verifies that no Node-specific specifier remains and that Daedalus-derived code is actually present (it runs continuously from CI). **(2) Execution in a real browser = observed (not "verified")**: the Teiresias demo was observed **once** in a real browser on 2026-08-29 (m3-t1/AC-16). A single observation is not an automated check, so it does not guard against regressions. **(3) Continuous, automated browser execution on CI and automated canvas drag operation = unverified** (a known gap of Cycle 1). `Daedalus/src` has not been changed by a single byte for this work (the m0 freeze has not been lifted). +- Runtime: Node.js and browser. Browser support is described in three levels. **(1) Bundlability = verified**: Teiresias' `pnpm run verify:demo-bundle` (`scripts/verify-demo-browser-bundle.mjs`) scans the output chunks of the demo build and mechanically verifies that no Node-specific specifier remains and that Daedalus-derived code is actually present (it runs continuously from CI). **(2) Execution in a real browser = checked manually once**: the Teiresias demo was run in a real browser on 2026-08-29. A single manual check is not an automated test, so it does not guard against regressions. **(3) Continuous, automated browser execution on CI = not available yet** (planned; see [#4](https://github.com/code4history/Daedalus/issues/4)). - Scope: fitting only, not CRS candidate lookup 日本語のREADMEは[こちら](./README.ja.md)。 @@ -84,7 +84,7 @@ the following hold. When `holdout` is specified, `holdoutRmsAffine <= holdoutRmsSimilarity` is required in addition. **Numerical zero of the SSE** is judged with a tolerance rather than by strict comparison -(the "exact fit within numerical tolerance" of the provisional specification). +("exact fit within numerical tolerance"). ``` epsSSE = 1024 · Number.EPSILON² · Σ(X² + Y²) (the sum is over the geo coordinates of the training GCPs) @@ -96,8 +96,7 @@ Z(s) = (s <= epsSSE) - `Z(SSE6)` and `¬Z(SSE4)`: the F gate is passed by treating `F = +Infinity` `alpha = 0.05`, `minRelativeImprovement = 0.01` and the safety factor `1024` are not values calibrated -against real data; they are initial heuristics (a provisional specification) for suppressing overfitting -in Cycle 1. +against real data; they are initial heuristics for suppressing overfitting. ### `debug` (only when `options.debug` is truthy) @@ -109,22 +108,16 @@ in Cycle 1. | `fStatistic` / `criticalF` / `relativeImprovement` | when the F gate was evaluated | | `holdoutRmsSimilarity` / `holdoutRmsAffine` | when `holdout` was specified and validation was reached | -### Non-goals (Cycle 1) +### Non-goals + +The following are not supported. They are candidates for a future release and are tracked in the +issues [#3](https://github.com/code4history/Daedalus/issues/3) (validation, weights, robust fitting, wide areas and more) and +[#4](https://github.com/code4history/Daedalus/issues/4) (browser execution on CI). - LOO (leave-one-out) validation (an unspecified holdout gives `validationSkipped`) - The general form of `criticalF` (the incomplete beta function for df1≠2) - Weights / robust fitting / more advanced inverse-mapping RMS / antimeridian and polar regions - -## Related documents - -- [Provisional specification of the CRS estimation libraries](../docs/superpowers/specs/2026-08-27-c1-crs-libs-provisional-spec.md) — - the observation contract and the numerical behaviour of this package are collected there - (it is a **provisional specification**, not a finalized specification approved by a human). -- [Mock transformer contract](../docs/superpowers/specs/2026-08-28-mock-transformer-contract.md) — - the contract of the mock adapter provided by `src/testkit/`. - -The links above **resolve only when the parent repository (the monorepo) is checked out**. -When this repository is cloned on its own, `../docs/` does not exist. +- Continuous, automated browser execution on CI (see "Runtime" above) ## License diff --git a/openspec/changes/bootstrap-worldfile-fitting/tasks.md b/openspec/changes/bootstrap-worldfile-fitting/tasks.md index 654675f..d24c46b 100644 --- a/openspec/changes/bootstrap-worldfile-fitting/tasks.md +++ b/openspec/changes/bootstrap-worldfile-fitting/tasks.md @@ -1,4 +1,12 @@ ## 1. Specification -- [ ] 1.1 Define `world_file_basic` fitting. -- [ ] 1.2 Define post-selection shear refinement. -- [ ] 1.3 Define residual metrics and helper contracts. +- [x] 1.1 Define `world_file_basic` fitting. +- [x] 1.2 Define post-selection shear refinement. +- [x] 1.3 Define residual metrics and helper contracts. + +> Implemented. +> - 1.1: closed-form similarity (Helmert) fitting behind the public `fitWorldFile` (`src/fitting/similarity.ts`, `src/index.ts`). +> Test: `tests/similarity.test.ts` "similarity closed-form Helmert fitting". +> - 1.2: affine (shear) refinement after selection, adopted only through the F gate and the relative improvement gate, +> with optional holdout validation (`src/fitting/affine.ts`). Test: `tests/affine.test.ts`. +> - 1.3: `rmsMeters` / `rmsPixels`, the `debug` SSE and MSE, and the warning codes (`src/types.ts`). +> Tests: `tests/affine.test.ts`, `tests/contract-freeze.test.ts`, `tests/api-smoke.test.ts`. diff --git a/openspec/project.md b/openspec/project.md index bfc1426..9f9ad49 100644 --- a/openspec/project.md +++ b/openspec/project.md @@ -1,7 +1,7 @@ # Project Context ## Purpose -Daedalus provides world-file fitting for a known target CRS in Maplat V3.1. +Daedalus provides world-file fitting for a known target CRS in Maplat. ## Tech Stack - TypeScript (planned) diff --git a/package.json b/package.json index 9d3b499..1d005bb 100644 --- a/package.json +++ b/package.json @@ -12,7 +12,7 @@ "default": "./dist/index.js" } }, - "description": "World-file fitting helpers for Maplat V3.1.", + "description": "World-file fitting helpers for Maplat.", "repository": { "type": "git", "url": "git+https://github.com/code4history/Daedalus.git"