[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_334":335,"_336":337},"routes/_app.x.$handle.$name",{"_110":109,"_38":51,"_118":96,"_33":119,"_13":14,"_120":-5,"_11":121,"_122":123,"_124":125,"_126":127,"_128":129,"_26":-5,"_130":131,"_132":133,"_134":31,"_135":136,"_137":65,"_138":-5,"_139":-5,"_140":141},"routes/_app.x.$handle.$name.branch.$branch.version.$version.diff",{"_11":12,"_13":14,"_15":16,"_17":18,"_19":-5},"project",{"_49":50,"_38":51,"_44":52,"_53":-5,"_54":31,"_55":56,"_57":56,"_58":59,"_60":-5,"_19":61,"_62":63,"_64":65,"_66":-5,"_67":-5,"_68":69,"_70":71,"_72":73,"_74":75,"_76":77,"_78":-5,"_79":80,"_81":80,"_82":83,"_84":83,"_85":83,"_86":80,"_87":80,"_88":89,"_90":91,"_92":93,"_94":31},"version",12,"changes",[25],"aiChatInfo",{"_20":21,"_22":-5,"_23":-5,"_24":-5},"userId","aiChatMessageId","a2ecfc63-ca04-4396-a8e0-30f8e6f12b5b","aiChatThreadId","aiChatThreadUserId","aiChatThreadTitle",{"_26":27,"_28":29,"_30":31,"_32":31,"_33":34,"_35":34},"file",{"_36":37,"_38":39,"_40":26,"_41":43,"_44":45,"_46":14,"_47":48,"_30":31},"previous",{"_36":37,"_38":39,"_40":26,"_41":42},"deleted",false,"deletedHere","path","frontend/index.html","latestPath","fileId","019e551a-5303-7301-b8d9-c0cc68bae9d4","name","index.html","fileType","content","\n\n \n \n \n React Hono Val Town Starter\n \n \n \n \n

\n \n \n \n","\n\n \n \n \n 回归分析平台\n \n \n \n \n \n

\n \n \n \n\n","createdAt",["D",1779544426880],"branchVersion","branchId","67c85a3c-56ae-11f1-8e14-ee650bb23af1","id","67c409f0-56ae-11f1-8e14-ee650bb23af1","regression-platform",["D",1779544249028],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","Starter template with client-side React & Hono server","imageUrl","6edec025-959e-47c2-ad6f-9a1be3b237f3","pinnedFileId","019e551a-5303-7301-b8d9-d25f0061fe9c","pinPreview",true,"pinMigrated","screenshotVersion","likeCount",0,"orgId","024eb136-bcd1-41e7-a89c-4a73af7907c6","tursoName","67c409f056ae11f18e14ee650bb23af1","tursoHostname","67c409f056ae11f18e14ee650bb23af1-val-town.aws-us-east-2.turso.io","tursoMeasuredAt",["D",1789492855745],"pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_110":109,"_117":113},"org",{"_110":109,"_111":-5,"_112":113,"_19":61,"_114":115,"_116":31},"branch",{"_13":95,"_38":96,"_49":48,"_97":50,"_98":99,"_100":101,"_102":56,"_103":104,"_105":106,"_107":51,"_108":109},"isOwned",26,"main","projectId","forkedBranchId","961078ea-f528-11ef-aa68-569c3dd06744","forkedBranchVersion",88,"forkedProjectPrivacy","forkedProjectId","96102714-f528-11ef-aa68-569c3dd06744","forkedProjectOrgId","019aa807-52c2-709c-928b-2029bae9a558","projectName","userHandle","jashon83","handle","displayName","avatar","https://img.clerk.com/eyJ0eXBlIjoicHJveHkiLCJzcmMiOiJodHRwczovL2ltYWdlcy5jbGVyay5kZXYvb2F1dGhfZ2l0aHViL2ltZ18ycGhUZFFQaG9weDBrQXdjMkZZVlNVRmh4cHkifQ","tier","free","canRestrictVals","profileImageUrl","branchName","","adoptOrgId",{"_49":50,"_38":51,"_44":329,"_53":-5,"_54":31,"_55":56,"_57":56,"_58":59,"_60":-5,"_19":61,"_62":63,"_64":65,"_66":-5,"_67":-5,"_68":69,"_70":71,"_72":73,"_74":75,"_76":330,"_78":-5,"_79":80,"_81":80,"_82":83,"_84":83,"_85":83,"_86":80,"_87":80,"_88":331,"_90":332,"_92":333,"_94":31},"tree",[196,205,212,213],"files",[196,197,198,199,200,201,202,203,204,205,206,207,208,209,210,211,212,213],"counts",{"_186":69,"_187":163,"_188":69,"_189":190,"_191":192},"contents",[153,154,155,156],"upstreamProject",{"_49":104,"_38":149,"_58":59,"_60":150,"_55":56,"_88":151},"upstreamBranch",{"_49":99,"_44":142,"_143":144,"_38":96,"_97":104,"_13":145,"_98":-5,"_100":-5,"_146":147,"_53":-5,"_94":31,"_55":56,"_108":148,"_70":106,"_107":149},"emptyEnvVars","flags",[],"hasHttpFiles","grants","orgs","customDomains",{},["D",1740674013689],"updatedAt",["D",1785420707976],91,"lockfileKey","1785420708123-esm.town_v_templates_react-hono-starter@91-main","templates","react-hono-starter","https://imagedelivery.net/iHX6Ovru0O7AjmyT5yZRoA/64dc00cc-94d9-47dd-81b1-dde02610d900/square512?exp=1789541999&sig=8fad4d8ed521ff338d1e7bb443f46241d1f9e0e657d24a41f82306e6eb92823c",{"_110":148,"_117":152},"https://imagedelivery.net/iHX6Ovru0O7AjmyT5yZRoA/129cab32-3bd7-4c3e-f39b-f12900c76100/square512?exp=1789541999&sig=156ae2b8ac752fc06f52d45eab428b512e0d31e2aab99a830d8bb32149c64f0f",{"_49":180,"_38":181,"_36":182,"_159":-5,"_47":48,"_46":183,"_161":65,"_44":184,"_13":69,"_40":178,"_41":-5,"_30":31,"_166":-5,"_33":185},{"_49":174,"_38":175,"_36":176,"_159":-5,"_47":48,"_46":163,"_161":65,"_44":177,"_13":69,"_40":178,"_41":-5,"_30":31,"_166":-5,"_33":179},{"_49":168,"_38":169,"_36":170,"_159":-5,"_47":48,"_46":163,"_161":65,"_44":171,"_13":69,"_40":26,"_41":172,"_30":31,"_166":-5,"_33":173},{"_49":157,"_38":158,"_36":63,"_159":-5,"_47":48,"_46":160,"_161":65,"_44":162,"_13":163,"_40":164,"_41":165,"_30":31,"_166":-5,"_33":167},"70420fc8-56ae-11f1-913f-ee650bb23af1","index.ts","parentId",2,"latest",["D",1779544263276],1,"http","import { parseVal, serveFile } from \"https://esm.town/v/std/utils/index.ts\";\nimport { Hono } from \"npm:hono\";\nimport { dataRoutes } from \"./backend/routes/data.ts\";\nimport { preprocessRoutes } from \"./backend/routes/preprocess.ts\";\nimport { trainRoutes } from \"./backend/routes/train.ts\";\nimport { regressionRoutes } from \"./backend/routes/regression.ts\";\nimport { predictRoutes } from \"./backend/routes/predict.ts\";\nimport { initDB } from \"./backend/db.ts\";\n\nconst app = new Hono();\n\n// Initialize database on startup\nawait initDB();\n\n// API routes\napp.route(\"/api/data\", dataRoutes);\napp.route(\"/api/preprocess\", preprocessRoutes);\napp.route(\"/api/train\", trainRoutes);\napp.route(\"/api/regression\", regressionRoutes);\napp.route(\"/api/predict\", predictRoutes);\n\n// Serve frontend\napp.get(\"/\", () => serveFile(\"/frontend/index.html\"));\napp.get(\"/frontend/**/*\", (c) => serveFile(c.req.path));\napp.get(\"/source\", (c) => c.redirect(parseVal().links.self.val));\n\napp.onError((err) => Promise.reject(err));\n\nexport default app.fetch;\n","label",[158],"67cdbe78-56ae-11f1-8e14-ee650bb23af1","README.md","019e551a-5303-7301-b8d9-cf7bf407ea4d",["D",1779544249028],"# React + Hono Starter\n\nA minimal starter template for building client-side React apps with a Hono\nbackend on Val Town. Click **Remix** to get your own copy.\n\n**Want to build UI?** Start writing React components in\n[`frontend/components/`](/frontend/components/). Import them into `App.tsx` and\nyou're good to go. Styling is [Twind](https://twind.style/) — just use Tailwind\nutility classes, zero config.\n\n**Want to add an API?** Add routes in [`index.ts`](/index.ts) (there's a\ncommented-out example). Fetch them from your React components.\n\n## Architecture\n\n```mermaid\ngraph LR\n Browser -->|\"GET /\"| Hono[\"index.ts
(Hono server)\"]\n Hono -->|serves| HTML[\"index.html\"]\n HTML -->|loads| Entry[\"index.tsx\"]\n Entry -->|renders| App[\"App.tsx\"]\n App -->|renders| Counter[\"Counter.tsx\"]\n```\n\n## File structure\n\n```\nindex.ts ← Hono backend (serves frontend + API routes)\nfrontend/\n index.html ← HTML shell (loads Twind + React)\n index.tsx ← React entrypoint (mounts )\n favicon.svg ← App icon\n components/\n App.tsx ← Root component (layout + view source link)\n Counter.tsx ← Example component with state\n```\n\n## How it works\n\n1. **`index.ts`** is the HTTP entrypoint — a Hono server that serves the\n frontend and any API routes you add.\n2. **`frontend/index.html`** loads Twind and the React entrypoint.\n3. **`frontend/index.tsx`** mounts the React app.\n4. **`frontend/components/App.tsx`** is the root layout — add new components\n alongside it.\n5. **`frontend/components/Counter.tsx`** is an example stateful component —\n replace it with your own!\n\n## Next steps\n\n- Add API routes in `index.ts`\n- Add new components in `frontend/components/`\n- Add a SQLite database — see the\n [React Hono Example](https://www.val.town/x/stevekrouse/reactHonoExample) for\n a full-stack example with database, queries, and shared types",[169],"67cdbc66-56ae-11f1-8e14-ee650bb23af1","frontend","019e551a-5303-7301-b8d9-c990a2e05367",["D",1779544249028],"directory",[175],"709cfbc2-56ae-11f1-913f-ee650bb23af1","backend","709cfc62-56ae-11f1-913f-ee650bb23af1",3,["D",1779544263876],[181],"pulls","branches","forks","envVars",{"_193":69,"_195":69},"connections",{"_193":69,"_194":69},"count","unfilledCount","emptyCount",{"_49":180,"_38":181,"_36":182,"_159":-5,"_47":48,"_46":183,"_161":65,"_44":326,"_13":69,"_40":178,"_41":-5,"_30":31,"_166":-5,"_215":327,"_33":328},{"_49":321,"_38":288,"_36":282,"_159":182,"_47":48,"_46":322,"_161":65,"_44":323,"_13":69,"_40":178,"_41":-5,"_30":31,"_166":-5,"_215":324,"_33":325},{"_49":313,"_38":314,"_36":315,"_159":282,"_47":48,"_46":316,"_161":31,"_44":317,"_13":69,"_40":225,"_41":318,"_30":31,"_166":-5,"_215":319,"_33":320},{"_49":305,"_38":306,"_36":307,"_159":282,"_47":48,"_46":308,"_161":31,"_44":309,"_13":69,"_40":225,"_41":310,"_30":31,"_166":-5,"_215":311,"_33":312},{"_49":297,"_38":298,"_36":299,"_159":282,"_47":48,"_46":300,"_161":31,"_44":301,"_13":69,"_40":225,"_41":302,"_30":31,"_166":-5,"_215":303,"_33":304},{"_49":289,"_38":290,"_36":291,"_159":282,"_47":48,"_46":292,"_161":65,"_44":293,"_13":69,"_40":225,"_41":294,"_30":31,"_166":-5,"_215":295,"_33":296},{"_49":279,"_38":280,"_36":281,"_159":282,"_47":48,"_46":283,"_161":31,"_44":284,"_13":69,"_40":225,"_41":285,"_30":31,"_166":-5,"_215":286,"_33":287},{"_49":271,"_38":272,"_36":273,"_159":182,"_47":48,"_46":274,"_161":65,"_44":275,"_13":69,"_40":225,"_41":276,"_30":31,"_166":-5,"_215":277,"_33":278},{"_49":263,"_38":264,"_36":265,"_159":182,"_47":48,"_46":266,"_161":65,"_44":267,"_13":69,"_40":225,"_41":268,"_30":31,"_166":-5,"_215":269,"_33":270},{"_49":174,"_38":175,"_36":176,"_159":-5,"_47":48,"_46":163,"_161":65,"_44":260,"_13":69,"_40":178,"_41":-5,"_30":31,"_166":-5,"_215":261,"_33":262},{"_49":256,"_38":248,"_36":243,"_159":176,"_47":48,"_46":163,"_161":65,"_44":257,"_13":69,"_40":178,"_41":-5,"_30":31,"_166":-5,"_215":258,"_33":259},{"_49":249,"_38":250,"_36":251,"_159":243,"_47":48,"_46":163,"_161":31,"_44":252,"_13":69,"_40":225,"_41":253,"_30":31,"_166":-5,"_215":254,"_33":255},{"_49":240,"_38":241,"_36":242,"_159":243,"_47":48,"_46":163,"_161":31,"_44":244,"_13":69,"_40":225,"_41":245,"_30":31,"_166":-5,"_215":246,"_33":247},{"_49":233,"_38":234,"_36":235,"_159":176,"_47":48,"_46":163,"_161":65,"_44":236,"_13":69,"_40":26,"_41":237,"_30":31,"_166":-5,"_215":238,"_33":239},{"_49":229,"_38":39,"_36":37,"_159":176,"_47":48,"_46":14,"_161":65,"_44":230,"_13":163,"_40":26,"_41":43,"_30":31,"_166":-5,"_215":231,"_33":232},{"_49":221,"_38":222,"_36":223,"_159":176,"_47":48,"_46":163,"_161":31,"_44":224,"_13":69,"_40":225,"_41":226,"_30":31,"_166":-5,"_215":227,"_33":228},{"_49":168,"_38":169,"_36":170,"_159":-5,"_47":48,"_46":163,"_161":65,"_44":218,"_13":69,"_40":26,"_41":172,"_30":31,"_166":-5,"_215":219,"_33":220},{"_49":157,"_38":158,"_36":63,"_159":-5,"_47":48,"_46":160,"_161":65,"_44":214,"_13":163,"_40":164,"_41":165,"_30":31,"_166":-5,"_215":216,"_33":217},["D",1779544263276],"children",[],[158],["D",1779544249028],[],[169],"67cdb75c-56ae-11f1-8e14-ee650bb23af1","index.tsx","019e551a-5303-7301-b8d9-c4a05227290d",["D",1779544249028],"script","/** @jsxImportSource https://esm.sh/react@18.2.0 */\nimport { createRoot } from \"https://esm.sh/react-dom@18.2.0/client\";\nimport { App } from \"./components/App.tsx\";\n\nconst root = document.getElementById(\"root\");\nif (!root) {\n throw new Error(\"No root element found\");\n}\n\ncreateRoot(root).render();",[],[175,222],"d1c61a0a-56ae-11f1-9595-ee650bb23af1",["D",1779544426880],[],[175,39],"67cdaf0a-56ae-11f1-8e14-ee650bb23af1","favicon.svg","019e551a-5303-7301-b8d9-bc0e8f5aabf9",["D",1779544249028],"\n",[],[175,234],"67cda8b6-56ae-11f1-8e14-ee650bb23af1","Counter.tsx","019e551a-5303-7301-b8d9-b7a8d40bac4e","019e551a-5303-7301-b8d9-b9274695fab5",["D",1779544249028],"/** @jsxImportSource https://esm.sh/react@18.2.0 */\nimport { useState } from \"https://esm.sh/react@18.2.0\";\n\nexport function Counter() {\n const [count, setCount] = useState(0);\n return (\n

\n Clicked\n setCount((c) => c + 1)}\n class=\"bg-blue-500 hover:bg-blue-600 text-white font-medium px-4 py-2 rounded-lg transition-colors cursor-pointer\"\n >\n {count}\n \n times\n

\n );\n}\n",[],[175,248,241],"components","67cd954c-56ae-11f1-8e14-ee650bb23af1","App.tsx","019e551a-5302-73ad-8589-62cc6807635a",["D",1779544249028],"/** @jsxImportSource https://esm.sh/react@18.2.0 */\nimport { Counter } from \"./Counter.tsx\";\n\nexport function App() {\n return (\n

React + Hono Starter

\n \n

\n View source\n

\n );\n}\n",[],[175,248,250],"67cdad8e-56ae-11f1-8e14-ee650bb23af1",["D",1779544249028],[207,208],[175,248],["D",1779544249028],[206,209,210,211],[175],"950e41dc-56ae-11f1-9ddb-ee650bb23af1","math.ts","950e4290-56ae-11f1-9ddb-ee650bb23af1",6,["D",1779544325014],"// ============================================================\n// Pure TypeScript statistics & regression engine (no native deps)\n// ============================================================\n\n// ---------- basic stats ----------\nexport function mean(arr: number[]): number {\n return arr.reduce((s, v) => s + v, 0) / arr.length;\n}\nexport function median(arr: number[]): number {\n const sorted = [...arr].sort((a, b) => a - b);\n const mid = Math.floor(sorted.length / 2);\n return sorted.length % 2 ? sorted[mid] : (sorted[mid - 1] + sorted[mid]) / 2;\n}\nexport function mode(arr: number[]): number {\n const freq: Record = {};\n arr.forEach((v) => (freq[v] = (freq[v] || 0) + 1));\n return +Object.entries(freq).sort((a, b) => b[1] - a[1])[0][0];\n}\nexport function variance(arr: number[]): number {\n const m = mean(arr);\n return arr.reduce((s, v) => s + (v - m) ** 2, 0) / (arr.length - 1);\n}\nexport function std(arr: number[]): number {\n return Math.sqrt(variance(arr));\n}\n\n// ---------- matrix helpers ----------\ntype Mat = number[][];\ntype Vec = number[];\n\nfunction zeros(r: number, c: number): Mat {\n return Array.from({ length: r }, () => new Array(c).fill(0));\n}\nfunction transpose(A: Mat): Mat {\n const [r, c] = [A.length, A[0].length];\n const T = zeros(c, r);\n for (let i = 0; i < r; i++) for (let j = 0; j < c; j++) T[j][i] = A[i][j];\n return T;\n}\nfunction matMul(A: Mat, B: Mat): Mat {\n const [ar, ac, bc] = [A.length, A[0].length, B[0].length];\n const C = zeros(ar, bc);\n for (let i = 0; i < ar; i++)\n for (let j = 0; j < bc; j++)\n for (let k = 0; k < ac; k++) C[i][j] += A[i][k] * B[k][j];\n return C;\n}\nfunction matVecMul(A: Mat, v: Vec): Vec {\n return A.map((row) => row.reduce((s, val, j) => s + val * v[j], 0));\n}\n\n// LU decomposition based inverse\nfunction invertMatrix(M: Mat): Mat {\n const n = M.length;\n const aug = M.map((row, i) => [...row, ...Array.from({ length: n }, (_, j) => (i === j ? 1 : 0))]);\n for (let col = 0; col < n; col++) {\n let maxRow = col;\n for (let row = col + 1; row < n; row++)\n if (Math.abs(aug[row][col]) > Math.abs(aug[maxRow][col])) maxRow = row;\n [aug[col], aug[maxRow]] = [aug[maxRow], aug[col]];\n const pivot = aug[col][col];\n if (Math.abs(pivot) < 1e-12) throw new Error(\"Singular matrix\");\n for (let j = 0; j < 2 * n; j++) aug[col][j] /= pivot;\n for (let row = 0; row < n; row++) {\n if (row === col) continue;\n const factor = aug[row][col];\n for (let j = 0; j < 2 * n; j++) aug[row][j] -= factor * aug[col][j];\n }\n }\n return aug.map((row) => row.slice(n));\n}\n\n// ---------- OLS regression ----------\nexport interface RegressionResult {\n coefficients: number[];\n intercept: number;\n rSquared: number;\n adjRSquared: number;\n mse: number;\n mae: number;\n predictions: number[];\n}\n\nexport function olsRegression(X: Mat, y: Vec): RegressionResult {\n const n = X.length;\n const p = X[0].length;\n // Add intercept column\n const Xa = X.map((row) => [1, ...row]);\n const Xt = transpose(Xa);\n const XtX = matMul(Xt, Xa);\n const XtXinv = invertMatrix(XtX);\n const Xty = matVecMul(Xt, y);\n const beta = matVecMul(XtXinv, Xty);\n\n const intercept = beta[0];\n const coefficients = beta.slice(1);\n const predictions = Xa.map((row) => row.reduce((s, v, j) => s + v * beta[j], 0));\n const yMean = mean(y);\n const ssTot = y.reduce((s, v) => s + (v - yMean) ** 2, 0);\n const ssRes = y.reduce((s, v, i) => s + (v - predictions[i]) ** 2, 0);\n const rSquared = 1 - ssRes / ssTot;\n const adjRSquared = 1 - ((1 - rSquared) * (n - 1)) / (n - p - 1);\n const mse = ssRes / n;\n const mae = y.reduce((s, v, i) => s + Math.abs(v - predictions[i]), 0) / n;\n\n return { coefficients, intercept, rSquared, adjRSquared, mse, mae, predictions };\n}\n\n// ---------- Ridge regression ----------\nexport function ridgeRegression(X: Mat, y: Vec, lambda = 1.0): RegressionResult {\n const n = X.length;\n const p = X[0].length;\n const Xa = X.map((row) => [1, ...row]);\n const Xt = transpose(Xa);\n const XtX = matMul(Xt, Xa);\n // Add lambda * I (skip intercept penalty)\n for (let i = 1; i < XtX.length; i++) XtX[i][i] += lambda;\n const XtXinv = invertMatrix(XtX);\n const Xty = matVecMul(Xt, y);\n const beta = matVecMul(XtXinv, Xty);\n\n const intercept = beta[0];\n const coefficients = beta.slice(1);\n const predictions = Xa.map((row) => row.reduce((s, v, j) => s + v * beta[j], 0));\n const yMean = mean(y);\n const ssTot = y.reduce((s, v) => s + (v - yMean) ** 2, 0);\n const ssRes = y.reduce((s, v, i) => s + (v - predictions[i]) ** 2, 0);\n const rSquared = 1 - ssRes / ssTot;\n const adjRSquared = 1 - ((1 - rSquared) * (n - 1)) / (n - p - 1);\n const mse = ssRes / n;\n const mae = y.reduce((s, v, i) => s + Math.abs(v - predictions[i]), 0) / n;\n\n return { coefficients, intercept, rSquared, adjRSquared, mse, mae, predictions };\n}\n\n// ---------- LASSO regression (coordinate descent) ----------\nexport function lassoRegression(X: Mat, y: Vec, lambda = 1.0, maxIter = 1000, tol = 1e-6): RegressionResult {\n const n = X.length;\n const p = X[0].length;\n\n // Standardize features\n const xMeans = Array.from({ length: p }, (_, j) => mean(X.map((row) => row[j])));\n const xStds = Array.from({ length: p }, (_, j) => {\n const s = std(X.map((row) => row[j]));\n return s < 1e-10 ? 1 : s;\n });\n const yMean = mean(y);\n const Xs = X.map((row) => row.map((v, j) => (v - xMeans[j]) / xStds[j]));\n const ys = y.map((v) => v - yMean);\n\n const beta = new Array(p).fill(0);\n const residual = [...ys];\n\n for (let iter = 0; iter < maxIter; iter++) {\n let maxChange = 0;\n for (let j = 0; j < p; j++) {\n // Add back current beta_j contribution\n for (let i = 0; i < n; i++) residual[i] += Xs[i][j] * beta[j];\n // Compute rho = X_j^T * residual / n\n let rho = 0;\n for (let i = 0; i < n; i++) rho += Xs[i][j] * residual[i];\n rho /= n;\n // Compute norm\n let xjNorm = 0;\n for (let i = 0; i < n; i++) xjNorm += Xs[i][j] ** 2;\n xjNorm /= n;\n // Soft threshold\n const newBeta = softThreshold(rho, lambda) / xjNorm;\n const change = Math.abs(newBeta - beta[j]);\n if (change > maxChange) maxChange = change;\n beta[j] = newBeta;\n // Update residual\n for (let i = 0; i < n; i++) residual[i] -= Xs[i][j] * beta[j];\n }\n if (maxChange < tol) break;\n }\n\n // Convert back to original scale\n const coefficients = beta.map((b, j) => b / xStds[j]);\n const intercept = yMean - coefficients.reduce((s, c, j) => s + c * xMeans[j], 0);\n\n const predictions = X.map((row) => intercept + row.reduce((s, v, j) => s + v * coefficients[j], 0));\n const ssTot = y.reduce((s, v) => s + (v - yMean) ** 2, 0);\n const ssRes = y.reduce((s, v, i) => s + (v - predictions[i]) ** 2, 0);\n const rSquared = ssTot > 0 ? 1 - ssRes / ssTot : 0;\n const pActive = coefficients.filter((c) => Math.abs(c) > 1e-10).length;\n const adjRSquared = n > pActive + 1 ? 1 - ((1 - rSquared) * (n - 1)) / (n - pActive - 1) : rSquared;\n const mse = ssRes / n;\n const mae = y.reduce((s, v, i) => s + Math.abs(v - predictions[i]), 0) / n;\n\n return { coefficients, intercept, rSquared, adjRSquared, mse, mae, predictions };\n}\n\nfunction softThreshold(x: number, lambda: number): number {\n if (x > lambda) return x - lambda;\n if (x < -lambda) return x + lambda;\n return 0;\n}\n\n// ---------- Elastic Net ----------\nexport function elasticNetRegression(X: Mat, y: Vec, lambda = 1.0, l1Ratio = 0.5, maxIter = 1000, tol = 1e-6): RegressionResult {\n const n = X.length;\n const p = X[0].length;\n const xMeans = Array.from({ length: p }, (_, j) => mean(X.map((row) => row[j])));\n const xStds = Array.from({ length: p }, (_, j) => {\n const s = std(X.map((row) => row[j]));\n return s < 1e-10 ? 1 : s;\n });\n const yMean = mean(y);\n const Xs = X.map((row) => row.map((v, j) => (v - xMeans[j]) / xStds[j]));\n const ys = y.map((v) => v - yMean);\n\n const beta = new Array(p).fill(0);\n const residual = [...ys];\n const l1 = lambda * l1Ratio;\n const l2 = lambda * (1 - l1Ratio);\n\n for (let iter = 0; iter < maxIter; iter++) {\n let maxChange = 0;\n for (let j = 0; j < p; j++) {\n for (let i = 0; i < n; i++) residual[i] += Xs[i][j] * beta[j];\n let rho = 0;\n for (let i = 0; i < n; i++) rho += Xs[i][j] * residual[i];\n rho /= n;\n let xjNorm = 0;\n for (let i = 0; i < n; i++) xjNorm += Xs[i][j] ** 2;\n xjNorm /= n;\n const newBeta = softThreshold(rho, l1) / (xjNorm + l2);\n const change = Math.abs(newBeta - beta[j]);\n if (change > maxChange) maxChange = change;\n beta[j] = newBeta;\n for (let i = 0; i < n; i++) residual[i] -= Xs[i][j] * beta[j];\n }\n if (maxChange < tol) break;\n }\n\n const coefficients = beta.map((b, j) => b / xStds[j]);\n const intercept = yMean - coefficients.reduce((s, c, j) => s + c * xMeans[j], 0);\n const predictions = X.map((row) => intercept + row.reduce((s, v, j) => s + v * coefficients[j], 0));\n const ssTot = y.reduce((s, v) => s + (v - yMean) ** 2, 0);\n const ssRes = y.reduce((s, v, i) => s + (v - predictions[i]) ** 2, 0);\n const rSquared = ssTot > 0 ? 1 - ssRes / ssTot : 0;\n const pActive = coefficients.filter((c) => Math.abs(c) > 1e-10).length;\n const adjRSquared = n > pActive + 1 ? 1 - ((1 - rSquared) * (n - 1)) / (n - pActive - 1) : rSquared;\n const mse = ssRes / n;\n const mae = y.reduce((s, v, i) => s + Math.abs(v - predictions[i]), 0) / n;\n\n return { coefficients, intercept, rSquared, adjRSquared, mse, mae, predictions };\n}\n\n// ---------- K-Fold cross validation ----------\nexport function kFoldCV(\n X: Mat,\n y: Vec,\n k: number,\n regressFn: (X: Mat, y: Vec) => RegressionResult,\n): { scores: number[]; meanScore: number } {\n const n = X.length;\n const indices = Array.from({ length: n }, (_, i) => i);\n // Shuffle\n for (let i = n - 1; i > 0; i--) {\n const j = Math.floor(Math.random() * (i + 1));\n [indices[i], indices[j]] = [indices[j], indices[i]];\n }\n const foldSize = Math.floor(n / k);\n const scores: number[] = [];\n\n for (let fold = 0; fold < k; fold++) {\n const testStart = fold * foldSize;\n const testEnd = fold === k - 1 ? n : testStart + foldSize;\n const testIdx = indices.slice(testStart, testEnd);\n const trainIdx = [...indices.slice(0, testStart), ...indices.slice(testEnd)];\n\n const Xtrain = trainIdx.map((i) => X[i]);\n const ytrain = trainIdx.map((i) => y[i]);\n const Xtest = testIdx.map((i) => X[i]);\n const ytest = testIdx.map((i) => y[i]);\n\n try {\n const result = regressFn(Xtrain, ytrain);\n // Compute R² on test set\n const preds = Xtest.map(\n (row) => result.intercept + row.reduce((s, v, j) => s + v * result.coefficients[j], 0),\n );\n const yMean = mean(ytest);\n const ssTot = ytest.reduce((s, v) => s + (v - yMean) ** 2, 0);\n const ssRes = ytest.reduce((s, v, i) => s + (v - preds[i]) ** 2, 0);\n scores.push(ssTot > 0 ? 1 - ssRes / ssTot : 0);\n } catch {\n scores.push(0);\n }\n }\n\n return { scores, meanScore: mean(scores) };\n}\n\n// ---------- Format equation ----------\nexport function formatEquation(\n target: string,\n features: string[],\n coefficients: number[],\n intercept: number,\n): string {\n let eq = `${target} = ${intercept.toFixed(4)}`;\n coefficients.forEach((c, i) => {\n if (Math.abs(c) < 1e-10) return;\n const sign = c >= 0 ? \" + \" : \" - \";\n eq += `${sign}${Math.abs(c).toFixed(4)} × ${features[i]}`;\n });\n return eq;\n}\n",[],[181,264],"76bd199c-56ae-11f1-8e14-ee650bb23af1","db.ts","76bd1b5e-56ae-11f1-8e14-ee650bb23af1",5,["D",1779544274151],"import { sqlite } from \"https://esm.town/v/std/sqlite/main.ts\";\n\nexport { sqlite };\n\nexport async function initDB() {\n await sqlite.batch([\n {\n sql: `CREATE TABLE IF NOT EXISTS datasets (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n name TEXT NOT NULL,\n source TEXT NOT NULL DEFAULT 'upload',\n row_count INTEGER DEFAULT 0,\n col_count INTEGER DEFAULT 0,\n columns_json TEXT,\n created_at TEXT DEFAULT (datetime('now')),\n updated_at TEXT DEFAULT (datetime('now'))\n )`,\n },\n {\n sql: `CREATE TABLE IF NOT EXISTS dataset_rows (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n dataset_id INTEGER NOT NULL,\n row_index INTEGER NOT NULL,\n data_json TEXT NOT NULL,\n FOREIGN KEY (dataset_id) REFERENCES datasets(id) ON DELETE CASCADE\n )`,\n },\n {\n sql: `CREATE TABLE IF NOT EXISTS training_results (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n dataset_id INTEGER NOT NULL,\n target_column TEXT NOT NULL,\n method TEXT NOT NULL DEFAULT 'lasso',\n alpha REAL,\n selected_features TEXT,\n cv_scores TEXT,\n mean_cv_score REAL,\n coefficients TEXT,\n created_at TEXT DEFAULT (datetime('now'))\n )`,\n },\n {\n sql: `CREATE TABLE IF NOT EXISTS regression_models (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n training_id INTEGER,\n dataset_id INTEGER NOT NULL,\n method TEXT NOT NULL,\n equation TEXT,\n coefficients TEXT,\n intercept REAL,\n r_squared REAL,\n adj_r_squared REAL,\n mse REAL,\n mae REAL,\n features TEXT,\n target TEXT,\n is_recommended INTEGER DEFAULT 0,\n created_at TEXT DEFAULT (datetime('now'))\n )`,\n },\n {\n sql: `CREATE TABLE IF NOT EXISTS prediction_pages (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n model_id INTEGER NOT NULL,\n name TEXT NOT NULL,\n description TEXT,\n is_active INTEGER DEFAULT 1,\n created_at TEXT DEFAULT (datetime('now'))\n )`,\n },\n {\n sql: `CREATE TABLE IF NOT EXISTS predictions (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n page_id INTEGER NOT NULL,\n model_id INTEGER NOT NULL,\n input_json TEXT NOT NULL,\n predicted_value REAL NOT NULL,\n actual_value REAL,\n error REAL,\n created_at TEXT DEFAULT (datetime('now'))\n )`,\n },\n ]);\n}\n",[],[181,272],"b61de2b0-56ae-11f1-9595-ee650bb23af1","train.ts","b61de396-56ae-11f1-9595-ee650bb23af1","72270dc0-56ae-11f1-8e14-ee650bb23af1",9,["D",1779544380480],"import { Hono } from \"npm:hono\";\nimport { sqlite } from \"../db.ts\";\nimport { lassoRegression, kFoldCV } from \"../math.ts\";\n\nexport const trainRoutes = new Hono();\n\n// Run LASSO training with 10-fold CV\ntrainRoutes.post(\"/lasso\", async (c) => {\n try {\n const { datasetId, targetColumn, alpha = 0.1 } = await c.req.json<{\n datasetId: number;\n targetColumn: string;\n alpha?: number;\n }>();\n\n // Load data\n const ds = await sqlite.execute({ sql: \"SELECT * FROM datasets WHERE id = ?\", args: [datasetId] });\n if (!ds.rows.length) return c.json({ error: \"Dataset not found\" }, 404);\n\n const columns: string[] = JSON.parse((ds.rows[0] as any).columns_json);\n if (!columns.includes(targetColumn)) return c.json({ error: \"Target column not found\" }, 400);\n\n const rowsResult = await sqlite.execute({\n sql: \"SELECT data_json FROM dataset_rows WHERE dataset_id = ? ORDER BY row_index\",\n args: [datasetId],\n });\n const rawData = rowsResult.rows.map((r: any) => JSON.parse(r.data_json));\n\n // Build numeric matrix\n const features = columns.filter((c) => c !== targetColumn);\n const numericFeatures = features.filter((col) => {\n return rawData.every((row: any) => {\n const v = row[col];\n return v !== null && v !== undefined && v !== \"\" && !isNaN(Number(v));\n });\n });\n\n if (numericFeatures.length === 0) return c.json({ error: \"No numeric features found\" }, 400);\n\n const X = rawData.map((row: any) => numericFeatures.map((col) => Number(row[col])));\n const y = rawData.map((row: any) => Number(row[targetColumn]));\n\n // Run LASSO\n const result = lassoRegression(X, y, alpha);\n\n // 10-fold CV\n const cv = kFoldCV(X, y, 10, (Xt, yt) => lassoRegression(Xt, yt, alpha));\n\n // Selected features (non-zero coefficients)\n const selectedFeatures = numericFeatures.filter((_, i) => Math.abs(result.coefficients[i]) > 1e-10);\n const selectedCoeffs = result.coefficients.filter((c) => Math.abs(c) > 1e-10);\n\n // Save training result\n const saveResult = await sqlite.execute({\n sql: `INSERT INTO training_results (dataset_id, target_column, method, alpha, selected_features, cv_scores, mean_cv_score, coefficients)\n VALUES (?, ?, ?, ?, ?, ?, ?, ?)`,\n args: [\n datasetId,\n targetColumn,\n \"lasso\",\n alpha,\n JSON.stringify(selectedFeatures),\n JSON.stringify(cv.scores),\n cv.meanScore,\n JSON.stringify(Object.fromEntries(selectedFeatures.map((f, i) => [f, selectedCoeffs[i]]))),\n ],\n });\n\n return c.json({\n trainingId: saveResult.lastInsertRowid,\n method: \"lasso\",\n alpha,\n allFeatures: numericFeatures,\n selectedFeatures,\n coefficients: Object.fromEntries(selectedFeatures.map((f, i) => [f, selectedCoeffs[i]])),\n intercept: result.intercept,\n rSquared: result.rSquared,\n cvScores: cv.scores,\n meanCvScore: cv.meanScore,\n totalFeatures: numericFeatures.length,\n selectedCount: selectedFeatures.length,\n });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n\n// Get training results\ntrainRoutes.get(\"/results\", async (c) => {\n const result = await sqlite.execute(\"SELECT * FROM training_results ORDER BY created_at DESC\");\n return c.json(result.rows.map((r: any) => ({\n ...r,\n selected_features: JSON.parse(r.selected_features || \"[]\"),\n cv_scores: JSON.parse(r.cv_scores || \"[]\"),\n coefficients: JSON.parse(r.coefficients || \"{}\"),\n })));\n});\n\ntrainRoutes.get(\"/results/:id\", async (c) => {\n const id = c.req.param(\"id\");\n const result = await sqlite.execute({ sql: \"SELECT * FROM training_results WHERE id = ?\", args: [+id] });\n if (!result.rows.length) return c.json({ error: \"Not found\" }, 404);\n const r: any = result.rows[0];\n return c.json({\n ...r,\n selected_features: JSON.parse(r.selected_features || \"[]\"),\n cv_scores: JSON.parse(r.cv_scores || \"[]\"),\n coefficients: JSON.parse(r.coefficients || \"{}\"),\n });\n});\n",[],[181,288,280],"routes","beda32f0-56ae-11f1-b03a-ee650bb23af1","regression.ts","beda3390-56ae-11f1-b03a-ee650bb23af1",10,["D",1779544395133],"import { Hono } from \"npm:hono\";\nimport { sqlite } from \"../db.ts\";\nimport {\n olsRegression,\n ridgeRegression,\n lassoRegression,\n elasticNetRegression,\n formatEquation,\n type RegressionResult,\n} from \"../math.ts\";\n\nexport const regressionRoutes = new Hono();\n\n// Run multiple regression methods on selected features\nregressionRoutes.post(\"/analyze\", async (c) => {\n try {\n const { datasetId, targetColumn, features, trainingId } = await c.req.json<{\n datasetId: number;\n targetColumn: string;\n features: string[];\n trainingId?: number;\n }>();\n\n // Load data\n const rowsResult = await sqlite.execute({\n sql: \"SELECT data_json FROM dataset_rows WHERE dataset_id = ? ORDER BY row_index\",\n args: [datasetId],\n });\n const rawData = rowsResult.rows.map((r: any) => JSON.parse(r.data_json));\n const X = rawData.map((row: any) => features.map((col) => Number(row[col])));\n const y = rawData.map((row: any) => Number(row[targetColumn]));\n\n // Run multiple regression methods\n const methods: { name: string; fn: () => RegressionResult }[] = [\n { name: \"OLS\", fn: () => olsRegression(X, y) },\n { name: \"Ridge (λ=0.1)\", fn: () => ridgeRegression(X, y, 0.1) },\n { name: \"Ridge (λ=1.0)\", fn: () => ridgeRegression(X, y, 1.0) },\n { name: \"LASSO (λ=0.01)\", fn: () => lassoRegression(X, y, 0.01) },\n { name: \"LASSO (λ=0.1)\", fn: () => lassoRegression(X, y, 0.1) },\n { name: \"ElasticNet (λ=0.1, α=0.5)\", fn: () => elasticNetRegression(X, y, 0.1, 0.5) },\n ];\n\n const results = [];\n for (const m of methods) {\n try {\n const result = m.fn();\n const equation = formatEquation(targetColumn, features, result.coefficients, result.intercept);\n results.push({\n method: m.name,\n rSquared: result.rSquared,\n adjRSquared: result.adjRSquared,\n mse: result.mse,\n mae: result.mae,\n coefficients: Object.fromEntries(features.map((f, i) => [f, result.coefficients[i]])),\n intercept: result.intercept,\n equation,\n });\n } catch {\n results.push({ method: m.name, error: \"Failed to compute\" });\n }\n }\n\n // Sort by adj R² and recommend best\n const valid = results.filter((r: any) => !r.error && r.adjRSquared != null);\n valid.sort((a: any, b: any) => b.adjRSquared - a.adjRSquared);\n const recommended = valid[0];\n\n // Save all models\n for (const r of valid) {\n const isRec = r === recommended ? 1 : 0;\n await sqlite.execute({\n sql: `INSERT INTO regression_models (training_id, dataset_id, method, equation, coefficients, intercept, r_squared, adj_r_squared, mse, mae, features, target, is_recommended)\n VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,\n args: [\n trainingId || null,\n datasetId,\n (r as any).method,\n (r as any).equation,\n JSON.stringify((r as any).coefficients),\n (r as any).intercept,\n (r as any).rSquared,\n (r as any).adjRSquared,\n (r as any).mse,\n (r as any).mae,\n JSON.stringify(features),\n targetColumn,\n isRec,\n ],\n });\n }\n\n return c.json({ results, recommended: recommended ? { ...recommended, isRecommended: true } : null });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n\n// Get all models\nregressionRoutes.get(\"/models\", async (c) => {\n const result = await sqlite.execute(\"SELECT * FROM regression_models ORDER BY adj_r_squared DESC\");\n return c.json(\n result.rows.map((r: any) => ({\n ...r,\n coefficients: JSON.parse(r.coefficients || \"{}\"),\n features: JSON.parse(r.features || \"[]\"),\n })),\n );\n});\n\n// Get single model\nregressionRoutes.get(\"/models/:id\", async (c) => {\n const id = c.req.param(\"id\");\n const result = await sqlite.execute({ sql: \"SELECT * FROM regression_models WHERE id = ?\", args: [+id] });\n if (!result.rows.length) return c.json({ error: \"Not found\" }, 404);\n const r: any = result.rows[0];\n return c.json({\n ...r,\n coefficients: JSON.parse(r.coefficients || \"{}\"),\n features: JSON.parse(r.features || \"[]\"),\n });\n});\n",[],[181,288,290],"ac479efc-56ae-11f1-8228-ee650bb23af1","preprocess.ts","ac479fc4-56ae-11f1-8228-ee650bb23af1",8,["D",1779544363980],"import { Hono } from \"npm:hono\";\nimport { sqlite } from \"../db.ts\";\nimport { mean, median, mode } from \"../math.ts\";\n\nexport const preprocessRoutes = new Hono();\n\n// Get preprocess preview\npreprocessRoutes.get(\"/preview/:datasetId\", async (c) => {\n const datasetId = +c.req.param(\"datasetId\");\n const ds = await sqlite.execute({ sql: \"SELECT * FROM datasets WHERE id = ?\", args: [datasetId] });\n if (!ds.rows.length) return c.json({ error: \"Dataset not found\" }, 404);\n\n const rowsResult = await sqlite.execute({\n sql: \"SELECT data_json FROM dataset_rows WHERE dataset_id = ? ORDER BY row_index\",\n args: [datasetId],\n });\n const data = rowsResult.rows.map((r: any) => JSON.parse(r.data_json));\n const columns = JSON.parse((ds.rows[0] as any).columns_json);\n\n // Analyze each column\n const analysis = columns.map((col: string) => {\n const values = data.map((row: any) => row[col]);\n const nonNull = values.filter((v: any) => v !== null && v !== undefined && v !== \"\");\n const nullCount = values.length - nonNull.length;\n const unique = new Set(nonNull).size;\n const isNumeric = nonNull.length > 0 && nonNull.every((v: any) => !isNaN(Number(v)));\n\n let stats: any = { total: values.length, missing: nullCount, unique, isNumeric };\n if (isNumeric && nonNull.length > 0) {\n const nums = nonNull.map(Number);\n stats.min = Math.min(...nums);\n stats.max = Math.max(...nums);\n stats.mean = mean(nums);\n stats.median = median(nums);\n }\n return { column: col, ...stats };\n });\n\n return c.json({ datasetId, columns, rowCount: data.length, analysis });\n});\n\n// Apply preprocessing\npreprocessRoutes.post(\"/apply\", async (c) => {\n try {\n const body = await c.req.json<{\n datasetId: number;\n operations: {\n deduplicate?: boolean;\n fillMissing?: Record;\n dropColumns?: string[];\n groupBy?: { column: string; agg: Record };\n };\n }>();\n\n const { datasetId, operations } = body;\n const ds = await sqlite.execute({ sql: \"SELECT * FROM datasets WHERE id = ?\", args: [datasetId] });\n if (!ds.rows.length) return c.json({ error: \"Dataset not found\" }, 404);\n\n const rowsResult = await sqlite.execute({\n sql: \"SELECT data_json FROM dataset_rows WHERE dataset_id = ? ORDER BY row_index\",\n args: [datasetId],\n });\n let data: Record[] = rowsResult.rows.map((r: any) => JSON.parse(r.data_json));\n let columns: string[] = JSON.parse((ds.rows[0] as any).columns_json);\n\n // 1) Drop columns\n if (operations.dropColumns?.length) {\n columns = columns.filter((c) => !operations.dropColumns!.includes(c));\n data = data.map((row) => {\n const newRow: any = {};\n columns.forEach((col) => (newRow[col] = row[col]));\n return newRow;\n });\n }\n\n // 2) Deduplicate\n if (operations.deduplicate) {\n const seen = new Set();\n data = data.filter((row) => {\n const key = JSON.stringify(columns.map((c) => row[c]));\n if (seen.has(key)) return false;\n seen.add(key);\n return true;\n });\n }\n\n // 3) Fill missing values\n if (operations.fillMissing) {\n for (const [col, method] of Object.entries(operations.fillMissing)) {\n const nonNull = data\n .map((row) => row[col])\n .filter((v) => v !== null && v !== undefined && v !== \"\");\n\n if (method === \"drop\") {\n data = data.filter((row) => row[col] !== null && row[col] !== undefined && row[col] !== \"\");\n continue;\n }\n\n const isNumeric = nonNull.every((v) => !isNaN(Number(v)));\n if (!isNumeric) continue;\n\n const nums = nonNull.map(Number);\n let fillValue: number;\n if (method === \"mean\") fillValue = mean(nums);\n else if (method === \"median\") fillValue = median(nums);\n else fillValue = mode(nums);\n\n data = data.map((row) => ({\n ...row,\n [col]: row[col] === null || row[col] === undefined || row[col] === \"\" ? fillValue : row[col],\n }));\n }\n }\n\n // 4) Group by\n if (operations.groupBy) {\n const { column: groupCol, agg } = operations.groupBy;\n const groups: Record[]> = {};\n data.forEach((row) => {\n const key = String(row[groupCol] ?? \"null\");\n if (!groups[key]) groups[key] = [];\n groups[key].push(row);\n });\n\n data = Object.entries(groups).map(([key, rows]) => {\n const result: any = { [groupCol]: key };\n for (const [col, fn] of Object.entries(agg)) {\n const vals = rows.map((r) => Number(r[col])).filter((v) => !isNaN(v));\n if (fn === \"sum\") result[col] = vals.reduce((a, b) => a + b, 0);\n else if (fn === \"mean\") result[col] = vals.length ? mean(vals) : 0;\n else if (fn === \"count\") result[col] = vals.length;\n else if (fn === \"min\") result[col] = vals.length ? Math.min(...vals) : 0;\n else if (fn === \"max\") result[col] = vals.length ? Math.max(...vals) : 0;\n }\n return result;\n });\n columns = Object.keys(data[0] || {});\n }\n\n // Save processed data as new dataset\n const newName = (ds.rows[0] as any).name + \"_processed\";\n const newDs = await sqlite.execute({\n sql: \"INSERT INTO datasets (name, source, row_count, col_count, columns_json) VALUES (?, ?, ?, ?, ?)\",\n args: [newName, \"preprocessed\", data.length, columns.length, JSON.stringify(columns)],\n });\n const newId = newDs.lastInsertRowid;\n\n const batchSize = 50;\n for (let i = 0; i < data.length; i += batchSize) {\n const batch = data.slice(i, i + batchSize);\n await sqlite.batch(\n batch.map((row, idx) => ({\n sql: \"INSERT INTO dataset_rows (dataset_id, row_index, data_json) VALUES (?, ?, ?)\",\n args: [newId, i + idx, JSON.stringify(row)],\n })),\n );\n }\n\n return c.json({ id: newId, name: newName, rows: data.length, columns });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n",[],[181,288,298],"cd8b3cd6-56ae-11f1-8274-ee650bb23af1","predict.ts","cd8b3d76-56ae-11f1-8274-ee650bb23af1",11,["D",1779544419786],"import { Hono } from \"npm:hono\";\nimport { sqlite } from \"../db.ts\";\n\nexport const predictRoutes = new Hono();\n\n// Create a prediction page from a model\npredictRoutes.post(\"/pages\", async (c) => {\n try {\n const { modelId, name, description } = await c.req.json<{\n modelId: number;\n name: string;\n description?: string;\n }>();\n const model = await sqlite.execute({ sql: \"SELECT * FROM regression_models WHERE id = ?\", args: [modelId] });\n if (!model.rows.length) return c.json({ error: \"Model not found\" }, 404);\n\n const result = await sqlite.execute({\n sql: \"INSERT INTO prediction_pages (model_id, name, description) VALUES (?, ?, ?)\",\n args: [modelId, name, description || \"\"],\n });\n return c.json({ id: result.lastInsertRowid, modelId, name });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n\n// List prediction pages\npredictRoutes.get(\"/pages\", async (c) => {\n const result = await sqlite.execute(`\n SELECT p.*, m.equation, m.method, m.features, m.target, m.coefficients, m.intercept, m.r_squared\n FROM prediction_pages p\n JOIN regression_models m ON p.model_id = m.id\n WHERE p.is_active = 1\n ORDER BY p.created_at DESC\n `);\n return c.json(\n result.rows.map((r: any) => ({\n ...r,\n features: JSON.parse(r.features || \"[]\"),\n coefficients: JSON.parse(r.coefficients || \"{}\"),\n })),\n );\n});\n\n// Get single page\npredictRoutes.get(\"/pages/:id\", async (c) => {\n const id = c.req.param(\"id\");\n const result = await sqlite.execute({\n sql: `SELECT p.*, m.equation, m.method, m.features, m.target, m.coefficients, m.intercept, m.r_squared, m.adj_r_squared\n FROM prediction_pages p\n JOIN regression_models m ON p.model_id = m.id\n WHERE p.id = ?`,\n args: [+id],\n });\n if (!result.rows.length) return c.json({ error: \"Not found\" }, 404);\n const r: any = result.rows[0];\n return c.json({\n ...r,\n features: JSON.parse(r.features || \"[]\"),\n coefficients: JSON.parse(r.coefficients || \"{}\"),\n });\n});\n\n// Make a prediction\npredictRoutes.post(\"/run\", async (c) => {\n try {\n const { pageId, inputs } = await c.req.json<{\n pageId: number;\n inputs: Record;\n }>();\n\n // Get page & model\n const page = await sqlite.execute({\n sql: `SELECT p.*, m.coefficients, m.intercept, m.features, m.target\n FROM prediction_pages p JOIN regression_models m ON p.model_id = m.id WHERE p.id = ?`,\n args: [pageId],\n });\n if (!page.rows.length) return c.json({ error: \"Page not found\" }, 404);\n\n const r: any = page.rows[0];\n const coefficients: Record = JSON.parse(r.coefficients);\n const features: string[] = JSON.parse(r.features);\n const intercept: number = r.intercept;\n\n // Compute prediction\n let predicted = intercept;\n for (const f of features) {\n if (inputs[f] === undefined) return c.json({ error: `Missing input: ${f}` }, 400);\n predicted += coefficients[f] * inputs[f];\n }\n\n // Save prediction\n const saveResult = await sqlite.execute({\n sql: \"INSERT INTO predictions (page_id, model_id, input_json, predicted_value) VALUES (?, ?, ?, ?)\",\n args: [pageId, r.model_id, JSON.stringify(inputs), predicted],\n });\n\n return c.json({\n predictionId: saveResult.lastInsertRowid,\n predicted,\n inputs,\n target: r.target,\n });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n\n// Update actual value for a prediction\npredictRoutes.put(\"/actual/:predictionId\", async (c) => {\n try {\n const predictionId = +c.req.param(\"predictionId\");\n const { actualValue } = await c.req.json<{ actualValue: number }>();\n\n const pred = await sqlite.execute({\n sql: \"SELECT predicted_value FROM predictions WHERE id = ?\",\n args: [predictionId],\n });\n if (!pred.rows.length) return c.json({ error: \"Prediction not found\" }, 404);\n\n const predicted = (pred.rows[0] as any).predicted_value;\n const error = actualValue - predicted;\n\n await sqlite.execute({\n sql: \"UPDATE predictions SET actual_value = ?, error = ? WHERE id = ?\",\n args: [actualValue, error, predictionId],\n });\n\n return c.json({ predictionId, predicted, actual: actualValue, error });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n\n// Get predictions history for a page\npredictRoutes.get(\"/history/:pageId\", async (c) => {\n const pageId = c.req.param(\"pageId\");\n const result = await sqlite.execute({\n sql: `SELECT * FROM predictions WHERE page_id = ? ORDER BY created_at DESC LIMIT 200`,\n args: [+pageId],\n });\n return c.json(\n result.rows.map((r: any) => ({\n ...r,\n input_json: JSON.parse(r.input_json || \"{}\"),\n })),\n );\n});\n\n// Get accuracy stats for a page\npredictRoutes.get(\"/accuracy/:pageId\", async (c) => {\n const pageId = c.req.param(\"pageId\");\n const result = await sqlite.execute({\n sql: \"SELECT predicted_value, actual_value, error FROM predictions WHERE page_id = ? AND actual_value IS NOT NULL\",\n args: [+pageId],\n });\n\n const rows: any[] = result.rows;\n if (!rows.length) return c.json({ message: \"No actual values recorded yet\", count: 0 });\n\n const errors = rows.map((r) => r.error);\n const absErrors = errors.map(Math.abs);\n const mape = rows.reduce((s: number, r: any) => {\n return s + (r.actual_value !== 0 ? Math.abs(r.error / r.actual_value) : 0);\n }, 0) / rows.length * 100;\n\n const meanErr = errors.reduce((a: number, b: number) => a + b, 0) / errors.length;\n const mae = absErrors.reduce((a: number, b: number) => a + b, 0) / absErrors.length;\n const mse = errors.reduce((s: number, e: number) => s + e * e, 0) / errors.length;\n const rmse = Math.sqrt(mse);\n\n return c.json({\n count: rows.length,\n meanError: meanErr,\n mae,\n mse,\n rmse,\n mape: mape,\n });\n});\n",[],[181,288,306],"9f8e094e-56ae-11f1-8228-ee650bb23af1","data.ts","9f8e0a2a-56ae-11f1-8228-ee650bb23af1",7,["D",1779544342625],"import { Hono } from \"npm:hono\";\nimport { sqlite } from \"../db.ts\";\nimport * as XLSX from \"npm:xlsx@0.18.5\";\n\nexport const dataRoutes = new Hono();\n\n// List all datasets\ndataRoutes.get(\"/datasets\", async (c) => {\n const result = await sqlite.execute(\"SELECT * FROM datasets ORDER BY created_at DESC\");\n return c.json(result.rows);\n});\n\n// Get dataset detail with rows\ndataRoutes.get(\"/datasets/:id\", async (c) => {\n const id = c.req.param(\"id\");\n const ds = await sqlite.execute({ sql: \"SELECT * FROM datasets WHERE id = ?\", args: [+id] });\n if (!ds.rows.length) return c.json({ error: \"Not found\" }, 404);\n const rows = await sqlite.execute({\n sql: \"SELECT row_index, data_json FROM dataset_rows WHERE dataset_id = ? ORDER BY row_index LIMIT 500\",\n args: [+id],\n });\n return c.json({\n ...ds.rows[0],\n columns: JSON.parse((ds.rows[0] as any).columns_json || \"[]\"),\n rows: rows.rows.map((r: any) => ({ index: r.row_index, data: JSON.parse(r.data_json) })),\n });\n});\n\n// Upload Excel file\ndataRoutes.post(\"/upload\", async (c) => {\n try {\n const formData = await c.req.formData();\n const file = formData.get(\"file\") as File;\n if (!file) return c.json({ error: \"No file provided\" }, 400);\n\n const buffer = await file.arrayBuffer();\n const workbook = XLSX.read(new Uint8Array(buffer), { type: \"array\" });\n const sheetName = workbook.SheetNames[0];\n const sheet = workbook.Sheets[sheetName];\n const jsonData: Record[] = XLSX.utils.sheet_to_json(sheet);\n\n if (!jsonData.length) return c.json({ error: \"Empty file\" }, 400);\n\n const columns = Object.keys(jsonData[0]);\n const name = file.name.replace(/\\.[^.]+$/, \"\");\n\n // Insert dataset metadata\n const dsResult = await sqlite.execute({\n sql: \"INSERT INTO datasets (name, source, row_count, col_count, columns_json) VALUES (?, ?, ?, ?, ?)\",\n args: [name, \"upload\", jsonData.length, columns.length, JSON.stringify(columns)],\n });\n const datasetId = dsResult.lastInsertRowid;\n\n // Insert rows in batches\n const batchSize = 50;\n for (let i = 0; i < jsonData.length; i += batchSize) {\n const batch = jsonData.slice(i, i + batchSize);\n await sqlite.batch(\n batch.map((row, idx) => ({\n sql: \"INSERT INTO dataset_rows (dataset_id, row_index, data_json) VALUES (?, ?, ?)\",\n args: [datasetId, i + idx, JSON.stringify(row)],\n })),\n );\n }\n\n return c.json({ id: datasetId, name, rows: jsonData.length, columns });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n\n// Manual data input / SQL-like import\ndataRoutes.post(\"/import-json\", async (c) => {\n try {\n const { name, data } = await c.req.json<{ name: string; data: Record[] }>();\n if (!data?.length) return c.json({ error: \"No data\" }, 400);\n\n const columns = Object.keys(data[0]);\n const dsResult = await sqlite.execute({\n sql: \"INSERT INTO datasets (name, source, row_count, col_count, columns_json) VALUES (?, ?, ?, ?, ?)\",\n args: [name || \"imported\", \"json\", data.length, columns.length, JSON.stringify(columns)],\n });\n const datasetId = dsResult.lastInsertRowid;\n\n const batchSize = 50;\n for (let i = 0; i < data.length; i += batchSize) {\n const batch = data.slice(i, i + batchSize);\n await sqlite.batch(\n batch.map((row, idx) => ({\n sql: \"INSERT INTO dataset_rows (dataset_id, row_index, data_json) VALUES (?, ?, ?)\",\n args: [datasetId, i + idx, JSON.stringify(row)],\n })),\n );\n }\n\n return c.json({ id: datasetId, name, rows: data.length, columns });\n } catch (err: any) {\n return c.json({ error: err.message }, 500);\n }\n});\n\n// Delete dataset\ndataRoutes.delete(\"/datasets/:id\", async (c) => {\n const id = c.req.param(\"id\");\n await sqlite.batch([\n { sql: \"DELETE FROM dataset_rows WHERE dataset_id = ?\", args: [+id] },\n { sql: \"DELETE FROM datasets WHERE id = ?\", args: [+id] },\n ]);\n return c.json({ ok: true });\n});\n",[],[181,288,314],"72270d2a-56ae-11f1-8e14-ee650bb23af1",4,["D",1779544266458],[198,199,200,201,202],[181,288],["D",1779544263876],[197,203,204],[181],["D",1779544249028],["D",1789492855745],{"_110":109,"_117":113},{"_110":109,"_111":-5,"_112":113,"_19":61,"_114":115,"_116":31},{"_13":95,"_38":96,"_49":48,"_97":50,"_98":99,"_100":101,"_102":56,"_103":104,"_105":106,"_107":51,"_108":109},"ctx",{"_88":-5,"_139":367,"_368":369,"_370":371,"_372":31,"_373":31,"_374":31,"_375":65,"_376":-5,"_377":-7,"_378":-7,"_379":69,"_380":31,"_381":31,"_382":-5,"_383":31,"_384":31,"_385":31,"_386":-5},"ENV",{"_338":339,"_340":341,"_342":96,"_343":344,"_345":346,"_347":348,"_349":350,"_351":352,"_353":354,"_355":356,"_357":358,"_359":360,"_361":362,"_363":364,"_365":366},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","e87eb04e707cf64e7e532c8e8ec395f209e17f15","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]

Diff · v12