${escapeHtml(code)}

`;\n }\n });\n return { __html: markedContent };\n };\n\n const escapeHtml = (unsafe) => {\n return unsafe\n .replace(/&/g, \"&\")\n .replace(//g, \">\")\n .replace(/\"/g, \""\")\n .replace(/'/g, \"'\");\n };\n\n const handleSubmit = async (e) => {\n e.preventDefault();\n if (!input.trim()) return;\n\n const newMessages = [...messages, { role: \"user\", content: input }];\n setMessages(newMessages);\n setInput(\"\");\n setIsLoading(true);\n\n try {\n const response = await fetch(\"/chat\", {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({ \n messages: newMessages, \n model: model \n })\n });\n\n if (!response.ok) throw new Error(\"Network response was not ok\");\n\n const data = await response.json();\n setMessages(prev => [...prev, { role: \"assistant\", content: data.message }]);\n } catch (error) {\n console.error(\"Error:\", error);\n setMessages(prev => [...prev, { \n role: \"assistant\", \n content: `🚨 Error: ${error.message}\\n\\nSorry, there was an issue processing your request.` \n }]);\n } finally {\n setIsLoading(false);\n }\n };\n\n return (\n

🤖 AI Assistant

\n \n \n

\n {messages.map((msg, index) => (\n

\n \n

\n ))}\n {isLoading && (\n

\n 🤖 Thinking...\n

\n )}\n

\n );\n}\n\nfunction client() {\n createRoot(document.getElementById(\"root\")).render();\n}\nif (typeof document !== \"undefined\") { client(); }\n\nexport default async function server(request: Request): Promise {\n if (request.method === \"POST\" && new URL(request.url).pathname === \"/chat\") {\n const { sqlite } = await import(\"https://esm.town/v/stevekrouse/sqlite\");\n \n const KEY = new URL(import.meta.url).pathname.split(\"/\").at(-1);\n await sqlite.execute(`\n CREATE TABLE IF NOT EXISTS ${KEY}_free_chats (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n role TEXT NOT NULL,\n content TEXT NOT NULL,\n model TEXT NOT NULL,\n timestamp DATETIME DEFAULT CURRENT_TIMESTAMP\n )\n `);\n\n const body = await request.json();\n const model = body.model || \"fallback\";\n const userMessage = body.messages[body.messages.length - 1].content;\n\n // Comprehensive Fallback Responses\n const fallbackResponses = [\n \"I apologize, but I'm experiencing technical difficulties. Could you rephrase your question?\",\n \"Our AI is taking a momentary break. Would you like to try again?\",\n \"Hmm, I'm having trouble understanding that request. Can you be more specific?\",\n \"Technical challenges are preventing a complete response right now.\",\n \"I'm processing your request, but encountering some obstacles.\",\n \"My circuits are a bit tangled at the moment. Let's try a simpler approach.\",\n \"Communication seems challenging right now. Could you simplify your request?\"\n ];\n\n // Advanced Fallback Generation Function\n function generateFallbackResponse(userInput: string): string {\n const keywords = userInput.toLowerCase().split(/\\s+/);\n \n const contextualResponses = {\n 'help': \"I'm here to assist you, but I'm experiencing some technical difficulties.\",\n 'code': \"I'd love to help with coding, but my code generation is temporarily unavailable.\",\n 'question': \"Great question! Unfortunately, I can't provide a detailed answer right now.\",\n 'explain': \"I'm ready to explain, but my explanation module is currently offline.\",\n 'how': \"How-to queries are welcome, though my current response capabilities are limited.\",\n 'why': \"Interesting 'why' question! Sadly, my reasoning circuits are momentarily disrupted.\"\n };\n\n for (const keyword of keywords) {\n if (contextualResponses[keyword]) {\n return contextualResponses[keyword];\n }\n }\n\n return fallbackResponses[Math.floor(Math.random() * fallbackResponses.length)];\n }\n\n try {\n // Deterministic Fallback Response Generation\n const aiMessage = generateFallbackResponse(userMessage);\n\n // Store messages in SQLite with explicit error handling\n try {\n await sqlite.execute(`\n INSERT INTO ${KEY}_free_chats (role, content, model) VALUES \n ('user', ?, ?), \n ('assistant', ?, ?)\n `, [\n userMessage, \n model,\n aiMessage, \n model\n ]);\n } catch (storageError) {\n console.error('Storage Error:', storageError);\n // Non-blocking storage error\n }\n\n return new Response(JSON.stringify({ \n message: aiMessage \n }), {\n headers: { \n \"Content-Type\": \"application/json\",\n \"X-Fallback-Reason\": \"AI Generation Failure\"\n },\n status: 200\n });\n\n } catch (error) {\n console.error('Critical AI Generation Error:', error);\n \n const criticalFallbackMessage = \"I'm currently experiencing significant technical difficulties. Please try again later.\";\n\n return new Response(JSON.stringify({ \n message: criticalFallbackMessage\n }), {\n headers: { \n \"Content-Type\": \"application/json\",\n \"X-Error-Type\": \"Critical Failure\"\n },\n status: 500\n });\n }\n }\n\n return new Response(`\n \n \n Resilient AI Assistant\n \n \n \n \n

\n \n \n \n \n `, {\n headers: { \"Content-Type\": \"text/html\" }\n });\n}\n\nconst css = `\n:root {\n --bg-primary: #f4f4f9;\n --bg-secondary: #ffffff;\n --text-primary: #2c3e50;\n --text-secondary: #ffffff;\n --accent-color: #3498db;\n --border-color: #e0e4e8;\n}\n\nbody {\n font-family: 'Inter', -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, 'Open Sans', 'Helvetica Neue', sans-serif;\n margin: 0;\n padding: 0;\n background-color: var(--bg-primary);\n color: var(--text-primary);\n line-height: 1.6;\n}\n\n.chat-container {\n max-width: 600px;\n margin: 0 auto;\n height: 100vh;\n display: flex;\n flex-direction: column;\n background-color: var(--bg-secondary);\n box-shadow: 0 0 20px rgba(0,0,0,0.1);\n}\n\n.chat-header {\n display: flex;\n justify-content: space-between;\n align-items: center;\n padding: 15px 20px;\n background-color: var(--accent-color);\n color: var(--text-secondary);\n}\n\n.messages {\n flex-grow: 1;\n overflow-y: auto;\n padding: 20px;\n}\n\n.message {\n margin-bottom: 15px;\n}\n\n.message-content {\n background-color: #f1f2f6;\n padding: 10px 15px;\n border-radius: 10px;\n}\n\n.input-area {\n display: flex;\n padding: 15px;\n background-color: var(--bg-secondary);\n border-top: 1px solid var(--border-color);\n}\n\n.input-area textarea {\n flex-grow: 1;\n padding: 10px;\n margin-right: 10px;\n border: 1px solid var(--border-color);\n border-radius: 4px;\n}\n\n.input-area button {\n padding: 10px 20px;\n background-color: var(--accent-color);\n color: var(--text-secondary);\n border: none;\n border-radius: 4px;\n}\n`;","createdAt",["D",1740814559058],"version",9,"currentUserHasLiked","likeCount",1,"forkCount","referenceCount","dependencyCount","commentCount","prCount","mergedPr",{"_130":131,"_132":133},"handle","keval","profileImageUrl","https://img.clerk.com/eyJ0eXBlIjoiZGVmYXVsdCIsImlpZCI6Imluc18yTm10cGhQUklwU0tqZHNiOHNnNGd2ZjJURFUiLCJyaWQiOiJ1c2VyXzJ0ZzJ6ZGdrczFXOEh3Q2JKOUZ2bUdRalUyNyJ9","forkedVal","interval","slug","keval/ecstaticSalmonOrangutan","description","@jsxImportSource https://esm.sh/react@18.2.0","versions",[142,145,149,153,157,161,165,169,172,175],{"_118":119,"_143":-5,"_116":144,"_113":111},"prId",["D",1740814559058],{"_118":146,"_143":-5,"_116":147,"_113":148},8,["D",1740814477977],"a14fad82-f66f-11ef-bfbf-569c3dd06744",{"_118":150,"_143":-5,"_116":151,"_113":152},7,["D",1740803454758],"f6f6ae6c-f655-11ef-964f-569c3dd06744",{"_118":154,"_143":-5,"_116":155,"_113":156},6,["D",1740803361365],"bf4c8482-f655-11ef-84b1-569c3dd06744",{"_118":158,"_143":-5,"_116":159,"_113":160},5,["D",1740760898329],"e15c0ff6-f5f2-11ef-80f6-569c3dd06744",{"_118":162,"_143":-5,"_116":163,"_113":164},4,["D",1740760834731],"bb7794a4-f5f2-11ef-ae2a-569c3dd06744",{"_118":166,"_143":-5,"_116":167,"_113":168},3,["D",1740760770190],"94faa398-f5f2-11ef-9cda-569c3dd06744",{"_118":91,"_143":-5,"_116":170,"_113":171},["D",1740760695448],"686e4528-f5f2-11ef-9284-569c3dd06744",{"_118":122,"_143":-5,"_116":173,"_113":174},["D",1740760628503],"40870fea-f5f2-11ef-97b9-569c3dd06744",{"_118":32,"_143":-5,"_116":176,"_113":177},["D",1740760322541],"8a2a4604-f5f1-11ef-8932-569c3dd06744","recentThread","breadcrumbs","comments",[],"isOwner","routes/_app.v.$handle.$name.dependencies",{"_185":186,"_247":122,"_248":137,"_249":32,"_95":250},"dependenciesList",[187],{"_101":188,"_189":190},"08f95904-3584-11ef-8995-0e10024e3b32","initialData",{"_99":191,"_96":192,"_101":188,"_103":193,"_104":-5,"_105":106,"_107":194,"_108":195,"_110":196,"_112":26,"_113":196,"_114":197,"_116":198,"_118":199,"_120":26,"_121":200,"_124":201,"_125":122,"_126":122,"_123":202,"_127":32,"_128":-5,"_20":203,"_134":206,"_135":-7,"_138":210,"_211":-5,"_136":212,"_140":213},"sqlite","a0bf3b31-15a5-4d5c-880e-4b1e22c9bc18","76a8911e-393e-11ef-8e71-de64eea55b61","# LLM-Safe Fork of @std/sqlite\n\nWe found that LLMs have trouble with the inputs and outputs of our @std/sqlite method.\n\n1. It expects the input to be `(sql: string, args: any[])` but it's `{ sql: string, args: any[] }`\n2. It expects the output to be objects, but they are arrays\n\nInstead of struggling to teach it, we built this val to be a wrapper around @std/sqlite that adheres to what the LLMs expect. \n\nThis val is also backwards-compatible with @std/sqlite, so we're considering merging it in.","script","5631f2d0-dcb9-11ef-80ec-e6cdfca9ef9f","import { API_URL } from \"https://esm.town/v/std/API_URL\";\nimport { LibsqlError, type ResultSet, type Row, type TransactionMode } from \"npm:@libsql/client\";\nimport { z } from \"npm:zod\";\n\n/**\n * Every Val Town account comes with its own private\n * [SQLite database](https://www.sqlite.org/) that\n * is accessible from any of your vals.\n * ([Docs ↗](https://docs.val.town/std/sqlite))\n */\nexport const sqlite = {\n /**\n * Executes a SQLite statement.\n *\n * @param {InStatement} statement - The SQLite statement to execute.\n * @example String query:\n * `sqlite.execute(\"SELECT 1;\")`\n * @example Query with arguments:\n * `sqlite.execute({sql: \"SELECT * from books WHERE year > ?;\", args: [2020]})`\n */\n execute,\n\n /**\n * Executes a batch of SQLite statements.\n *\n * @param {InStatement[]} statements - An array of SQLite statements to execute.\n * @param {TransactionMode} [mode] - The transaction mode for the batch execution.\n */\n batch,\n};\n\n// ------------\n// Functions\n// ------------\n\nasync function execute(statement: InStatement, args?: InArgs): Promise {\n const res = await fetch(`${API_URL}/v1/sqlite/execute`, {\n method: \"POST\",\n headers: {\n Authorization: `Bearer ${Deno.env.get(\"valtown\")}`,\n },\n body: JSON.stringify({ statement: normalizeStatement(statement, args) }),\n });\n if (!res.ok) {\n throw createResError(await res.text());\n }\n const resultSet = await res.json();\n return upgradeResultSet(resultSet);\n}\n\nasync function batch(statements: InStatement[], mode?: TransactionMode): Promise {\n if (mode && ![\"write\", \"read\", \"deferred\"].includes(mode)) {\n throw new Error(\"Invalid transaction mode provided to sqlite.batch. Must be 'write', 'read', or 'deferred.\");\n }\n const res = await fetch(`${API_URL}/v1/sqlite/batch`, {\n method: \"POST\",\n headers: {\n Authorization: `Bearer ${Deno.env.get(\"valtown\")}`,\n },\n body: JSON.stringify({ statements: statements.map(s => normalizeStatement(s)), mode }),\n });\n if (!res.ok) {\n throw createResError(await res.text());\n }\n const resultSets = await res.json();\n return resultSets.map(upgradeResultSet);\n}\n\nfunction createResError(body: string) {\n try {\n const e = zLibsqlError.parse(JSON.parse(body));\n // e.message already contains the code, and LibsqlError adds the\n // code to the beginning, so we remove it here\n const msg = e.message.replace(e.code, \"\").replace(/^:\\s+/, \"\");\n return new LibsqlError(msg, e.code, e.rawCode);\n } catch (_) {\n // Failed to parse libsql error\n }\n return new Error(body);\n}\n\n/* Val Town's Turso Proxy returns rows as an array of values\n * Yet the LibSQL API has a Row type which behave as an array or object,\n * ie you can access it via numerical index or string\n */\ninterface ImpoverishedResultSet extends ResultSet {\n rows: any[];\n}\n\nfunction normalizeStatement(statement: InStatement, args?: InArgs) {\n if (Array.isArray(statement)) {\n // for the case of an array of arrays\n return normalizeStatement(statement[0], statement[1]);\n }\n if (typeof statement === \"string\") {\n if (args) {\n return { sql: statement, args };\n } else return statement;\n }\n else if (typeof statement === \"object\") {\n if (statement.args) {\n return statement;\n }\n if (args) {\n return { sql: statement.sql, args };\n }\n return statement.sql;\n }\n else throw new Error(\"Error normalizing sql input\");\n}\n\nfunction upgradeResultSet(results: ImpoverishedResultSet): ResultSet {\n return {\n ...results,\n rows: results.rows.map(row => rowFromSql(row, results.columns)),\n };\n}\n\n// adapted from\n// https://github.com/tursodatabase/libsql-client-ts/blob/17dd996b840c950dd22b871adfe4ba0eb4a5ead3/packages/libsql-client/src/sqlite3.ts#L314C1-L337C2\nfunction rowFromSql(\n sqlRow: Array,\n columns: Array,\n): Row {\n const row = {};\n // make sure that the \"length\" property is not enumerable\n Object.defineProperty(row, \"length\", { value: sqlRow.length });\n for (let i = 0; i < sqlRow.length; ++i) {\n const value = sqlRow[i];\n Object.defineProperty(row, i, { value });\n\n const column = columns[i];\n if (!Object.hasOwn(row, column)) {\n Object.defineProperty(row, column, {\n value,\n enumerable: true,\n configurable: true,\n writable: true,\n });\n }\n }\n return row as Row;\n}\n\n// ------------\n// Helpers\n// ------------\n\nconst zLibsqlError = z.object({\n message: z.string(),\n code: z.string(),\n rawCode: z.number().optional(),\n});\n\n// We patch these types to only support JSON values\nexport type InValue = null | string | number | boolean;\nexport type InArgs = Array | Record;\nexport type InStatement = {\n /**\n * The SQL statement to execute.\n */\n sql: string;\n\n /**\n * The arguments to bind to the SQL statement.\n */\n args: InArgs;\n} | string;",["D",1737987404484],13,11,16191,21,{"_130":204,"_132":205},"stevekrouse","https://images.clerk.dev/uploaded/img_2PqHa2Gsy93xQrjh2w78Xu0cChW.jpeg",{"_113":207,"_130":208,"_99":191,"_118":154,"_209":-5},"56c849ee-f1f8-11ee-9b87-cab58f78796a","std","prDisabledHint","LLM-Safe Fork of @std/sqlite We found that LLMs have trouble with the inputs and outputs of our @std/sqlite method. It expects the input to be (sql: string, args: any[]) but it's { sql: string, args: any[] } It expects the output to be objects, but they are arrays Instead of struggling to teach it, we built this val to be a wrapper around @std/sqlite that adheres to what the LLMs expect. This val is also backwards-compatible with @std/sqlite, so we're considering merging it in.","latestEvaluation","stevekrouse/sqlite",[214,218,222,225,229,232,235,238,241,244],{"_116":215,"_118":199,"_216":217},["D",1737987404484],"n","1",{"_116":219,"_118":220,"_216":221},["D",1724087412304],12,"2",{"_116":223,"_118":200,"_216":224},["D",1724087404512],"3",{"_116":226,"_118":227,"_216":228},["D",1723494410142],10,"4",{"_116":230,"_118":119,"_216":231},["D",1723494389876],"5",{"_116":233,"_118":146,"_216":234},["D",1723494345363],"6",{"_116":236,"_118":150,"_216":237},["D",1723494323065],"7",{"_116":239,"_118":154,"_216":240},["D",1719604370206],"8",{"_116":242,"_118":158,"_216":243},["D",1719604260994],"9",{"_116":245,"_118":162,"_216":246},["D",1719604144103],"10","dependenciesCount","valname","page",{"_113":111,"_99":100,"_130":131,"_11":97,"_118":119,"_101":102,"_114":115,"_105":106,"_112":26,"_136":137},"routes/_app","actionData","errors"]