[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_164":165,"_166":167},"routes/_app.x.$handle.$name",{"_11":12,"_13":14,"_15":-7,"_61":134,"_16":-7,"_135":-5,"_17":136,"_137":138,"_28":139,"_21":140,"_23":141,"_142":-5,"_143":-5,"_144":-5,"_145":36,"_146":147,"_148":34,"_149":-5,"_150":-5,"_151":152},"routes/_app.x.$handle.$name._index",{"_11":12,"_13":14,"_15":-7,"_16":-7,"_17":18,"_19":20,"_21":22,"_23":24,"_25":-7,"_26":-5,"_27":-5,"_28":29,"_30":-5,"_31":-5,"_32":-5,"_33":34,"_35":36,"_37":34},"handle","niteshgarg","name","yxmd2dataiku","branchName","version","project",{"_39":78,"_13":14,"_50":79,"_80":-5,"_81":36,"_82":83,"_84":83,"_85":86,"_87":-5,"_88":89,"_90":-5,"_91":36,"_92":-5,"_93":-5,"_94":67,"_95":96,"_97":-5,"_98":-5,"_99":-5,"_100":-5,"_101":102,"_103":102,"_104":105,"_106":105,"_107":105,"_108":102,"_109":102,"_110":111,"_112":113,"_114":115,"_116":36},"tags",[],"counts",{"_66":67,"_68":69,"_70":67,"_71":72,"_73":74},"contents",[63],"readme","upstream","pull","files",[38],"preview","ahead","behind","showHomeCustomizationHint",true,"showHTTPPreviewSettingsHint",false,"hasHTTPFiles",{"_39":40,"_13":41,"_42":43,"_44":-5,"_45":46,"_47":48,"_49":34,"_50":51,"_16":52,"_53":54,"_55":56,"_57":36,"_58":14,"_59":60,"_61":62},"id","49fca7b2-aa1b-4b2b-bce0-1b5728855042","main.ts","fileId","759817c8-7ec8-11f1-97ca-1607ee4eb77e","parentId","branchId","de99b434-7511-4e88-ba6c-6d6da2be5e6b","branchVersion",6,"latest","createdAt",["D",1783966855505],5,"fileType","http","content","// yxmd2dataiku — Alteryx .yxmd -> importable Dataiku DSS project export ZIP (+ report).\n// One Alteryx tool -> one Dataiku recipe (visual where the on-disk format is confirmed,\n// exact-logic Python otherwise). Import via \"+ New project > Import project\".\n// Formats mirror real DSS 7/8 exports (manifestVersion 3, claims DSS 8.0.3).\n\nimport { XMLParser } from \"npm:fast-xml-parser\";\nimport { zipSync, strToU8 } from \"npm:fflate\";\n\n// ========\n// small utils\n// ========\nconst arr = (x: any): any[] => (x == null ? [] : Array.isArray(x) ? x : [x]);\nconst NOW = Date.now();\nconst VT = { versionNumber: 0, lastModifiedBy: { login: \"admin\" }, lastModifiedOn: NOW };\n\nfunction sanitizeName(s: string, max = 56): string {\n let n = String(s || \"\").replace(/[^A-Za-z0-9_]+/g, \"_\").replace(/^_+|_+$/g, \"\").slice(0, max);\n if (!n) n = \"x\";\n if (!/^[A-Za-z]/.test(n)) n = \"t_\" + n;\n return n;\n}\n\nfunction csvEscape(v: string): string {\n const s = String(v ?? \"\");\n return /[\",\\n\\r]/.test(s) ? '\"' + s.replace(/\"/g, '\"\"') + '\"' : s;\n}\n\n// Alteryx field type -> DSS storage type\nfunction dkuType(t: string): string {\n const x = String(t || \"\").toLowerCase();\n if (/^(bool)/.test(x)) return \"boolean\";\n if (/^(byte|int16|int32|int64)/.test(x)) return \"bigint\";\n if (/^(float|double|fixeddecimal)/.test(x)) return \"double\";\n if (/^(date$|datetime)/.test(x)) return \"date\";\n return \"string\"; // string, wstring, v_string, v_wstring, time, spatialobj, blob...\n}\n\n// ========\n// Formula translation: Alteryx expression -> Dataiku formula (GREL-like)\n// ========\n// Masking sentinels (never appear in real formulas)\nconst S_LIT = \"\u0001\"; // string literals\nconst S_COL = \"\u0002\"; // [column] refs\nconst S_FN = \"\u0003\"; // guards emitted function names from re-transformation\n\ninterface Xlat { expr: string; warnings: string[] }\n\nfunction colRef(name: string): string {\n return /^[A-Za-z][A-Za-z0-9_]*$/.test(name) ? name : `val(\"${name.replace(/\"/g, '\\\\\"')}\")`;\n}\n\nconst FN_WHITELIST = new Set([\n \"if\", \"switch\", \"and\", \"or\", \"not\", \"val\", \"strval\", \"numval\", \"isblank\", \"isnonblank\",\n \"isnull\", \"length\", \"contains\", \"indexof\", \"lastindexof\", \"startswith\", \"endswith\",\n \"substring\", \"touppercase\", \"tolowercase\", \"totitlecase\", \"trim\", \"strip\", \"replace\",\n \"match\", \"tonumber\", \"tostring\", \"abs\", \"ceil\", \"floor\", \"round\", \"pow\", \"sqrt\", \"exp\",\n \"ln\", \"log\", \"min\", \"max\", \"mod\", \"now\", \"trunc\", \"inc\", \"diff\", \"datepart\",\n \"asdatetimetz\", \"asdatetimenotz\", \"asdateonly\", \"concat\", \"coalesce\", \"rand\", \"split\",\n \"format\", \"asbool\", \"isnumeric\", \"iserror\", \"avg\", \"sum\", \"chomp\", \"slice\", \"get\", \"cos\", \"sin\", \"tan\",\n]);\n\nconst KNOWN_UNSUPPORTED: Record = {\n datetimeformat: \"DateTimeFormat has no formula equivalent — add a 'Format date with custom format' Prepare step (DateFormatter) instead.\",\n padleft: \"PadLeft has no formula equivalent — consider format() printf-style formatting or a Python step.\",\n padright: \"PadRight has no formula equivalent — consider format() printf-style formatting or a Python step.\",\n regex_countmatches: \"REGEX_CountMatches has no formula equivalent — consider a Python step.\",\n datetimefirstofmonth: \"Use trunc(, 'month') manually.\",\n getword: \"GetWord has no direct equivalent — split() returns an array; review manually.\",\n};\n\n// find args of fn call starting at s[open]='(' ; returns [args[], endIdxAfterClose]\nfunction parseArgs(s: string, open: number): [string[], number] | null {\n let depth = 0; const args: string[] = []; let cur = \"\";\n for (let i = open; i < s.length; i++) {\n const ch = s[i];\n if (ch === \"(\") { depth++; if (depth === 1) continue; }\n else if (ch === \")\") { depth--; if (depth === 0) { args.push(cur.trim()); return [args, i + 1]; } }\n else if (ch === \",\" && depth === 1) { args.push(cur.trim()); cur = \"\"; continue; }\n if (depth >= 1) cur += ch;\n }\n return null;\n}\n\nfunction replaceFunc(s: string, name: string, build: (args: string[], warn: (w: string) => void) => string, warn: (w: string) => void): string {\n const re = new RegExp(\"\\\\b\" + name + \"\\\\s*\\\\(\", \"i\");\n let guard = 0;\n while (guard++ < 200) {\n const m = re.exec(s);\n if (!m) break;\n const open = m.index + m[0].length - 1;\n const parsed = parseArgs(s, open);\n if (!parsed) { warn(`Unbalanced parentheses near ${name}( — left untranslated.`); break; }\n const [args, end] = parsed;\n s = s.slice(0, m.index) + build(args, warn) + s.slice(end);\n }\n return s;\n}\n\nconst fx = (n: string) => n + S_FN + \"(\"; // guarded emit\n\nexport function translateFormula(src: string): Xlat {\n const warnings: string[] = [];\n const warn = (w: string) => { if (!warnings.includes(w)) warnings.push(w); };\n if (!src || !String(src).trim()) return { expr: \"\", warnings };\n let s = String(src);\n\n // 1) mask string literals ('...' and \"...\")\n const lits: string[] = [];\n { let out = \"\", i = 0;\n while (i < s.length) {\n const ch = s[i];\n if (ch === \"'\" || ch === '\"') {\n let j = i + 1;\n while (j < s.length && s[j] !== ch) j++;\n lits.push(s.slice(i, Math.min(j + 1, s.length)));\n out += S_LIT + (lits.length - 1) + S_LIT;\n i = j + 1;\n } else { out += ch; i++; }\n }\n s = out;\n }\n const unlit = (a: string): string | null => {\n const m = /^\\s*\u0001(\\d+)\u0001\\s*$/.exec(a);\n if (!m) return null;\n const raw = lits[+m[1]];\n return raw.slice(1, -1);\n };\n const relit = (content: string) => { lits.push(\"'\" + content.replace(/'/g, \"\\\\'\") + \"'\"); return S_LIT + (lits.length - 1) + S_LIT; };\n const remask = (raw: string) => { lits.push(raw); return S_LIT + (lits.length - 1) + S_LIT; };\n\n // 1.5) strip Alteryx comments (// line, /* */ block) — literals already masked\n s = s.replace(/\\/\\*[\\s\\S]*?\\*\\//g, \" \").replace(/\\/\\/[^\\n\\r]*/g, \" \");\n\n // 2) mask [column] refs\n const cols: string[] = [];\n s = s.replace(/\\[([^\\]\\[]+)\\]/g, (_m, c) => { cols.push(String(c)); return S_COL + (cols.length - 1) + S_COL; });\n\n // 3) IIF -> if\n s = s.replace(/\\bIIF\\s*\\(/gi, fx(\"if\"));\n\n // 4) IF..THEN..ELSEIF..ELSE..ENDIF -> nested if(,,) — innermost first\n { let guard = 0;\n while (guard++ < 100) {\n const endm = /\\bENDIF\\b/i.exec(s);\n if (!endm) { if (new RegExp(\"\\\\bIF\\\\b(?!\" + S_FN + \")\", \"i\").test(s)) warn(\"IF without ENDIF — could not fully translate conditional.\"); break; }\n let ifIdx = -1;\n const reIf = new RegExp(\"\\\\bIF\\\\b(?!\" + S_FN + \")\", \"gi\"); let mm: RegExpExecArray | null;\n while ((mm = reIf.exec(s)) && mm.index < endm.index) ifIdx = mm.index;\n if (ifIdx < 0) { warn(\"ENDIF without IF — check conditional syntax.\"); break; }\n const seg = s.slice(ifIdx + 2, endm.index);\n // split seg at depth-0 THEN / ELSEIF / ELSE\n const parts: { kw: string; text: string }[] = [{ kw: \"COND\", text: \"\" }];\n let depth = 0;\n for (let i = 0; i < seg.length; i++) {\n const ch = seg[i];\n if (ch === \"(\") depth++;\n else if (ch === \")\") depth--;\n if (depth === 0) {\n const rest = seg.slice(i);\n const km = /^(THEN|ELSEIF|ELSE)\\b/i.exec(rest);\n if (km && /[^A-Za-z0-9_]/.test(seg[i - 1] || \" \")) {\n parts.push({ kw: km[1].toUpperCase(), text: \"\" });\n i += km[1].length - 1;\n continue;\n }\n }\n parts[parts.length - 1].text += ch;\n }\n // build: COND then VALUE [ELSEIF cond THEN value]* [ELSE value]\n const conds: string[] = []; const vals: string[] = []; let elseVal = \"null\"; let ok = true;\n for (let p = 0; p < parts.length;) {\n const pt = parts[p];\n if (pt.kw === \"COND\" || pt.kw === \"ELSEIF\") {\n if (parts[p + 1]?.kw === \"THEN\") { conds.push(pt.text.trim()); vals.push(parts[p + 1].text.trim()); p += 2; }\n else { ok = false; break; }\n } else if (pt.kw === \"ELSE\") { elseVal = pt.text.trim(); p++; }\n else { ok = false; break; }\n }\n if (!ok || !conds.length) { warn(\"Could not parse IF/THEN/ELSE structure — review manually.\"); s = s.slice(0, ifIdx) + seg + s.slice(endm.index + endm[0].length); continue; }\n let built = elseVal;\n for (let i = conds.length - 1; i >= 0; i--) built = `${fx(\"if\")}${conds[i]}, ${vals[i]}, ${built})`;\n s = s.slice(0, ifIdx) + built + s.slice(endm.index + endm[0].length);\n }\n }\n\n // 5) operators\n s = s.replace(/<>/g, \"!=\");\n s = s.replace(/(?!=])=(?!=)/g, \"==\");\n s = s.replace(/\\bAND\\b/gi, \"&&\").replace(/\\bOR\\b/gi, \"||\");\n // NOT \n { let guard = 0;\n const reNot = new RegExp(\"\\\\bNOT\\\\b(?!\" + S_FN + \")\\\\s*\", \"i\");\n while (guard++ < 100) {\n const m = reNot.exec(s);\n if (!m) break;\n let i = m.index + m[0].length;\n let operand = \"\";\n if (s[i] === \"(\") {\n const p = parseArgs(s, i);\n if (!p) { warn(\"Unbalanced parentheses after NOT.\"); break; }\n operand = s.slice(i, p[1]); i = p[1];\n } else {\n const idm = /^[A-Za-z_][A-Za-z0-9_]*\\s*\\(/.exec(s.slice(i));\n if (idm) {\n const open = i + idm[0].length - 1;\n const p = parseArgs(s, open);\n if (!p) { warn(\"Unbalanced parentheses after NOT.\"); break; }\n operand = s.slice(i, p[1]); i = p[1];\n } else {\n // consume until top-level boundary\n let depth = 0, j = i;\n for (; j < s.length; j++) {\n const ch = s[j];\n if (ch === \"(\") depth++;\n else if (ch === \")\") { if (depth === 0) break; depth--; }\n else if (depth === 0 && (ch === \",\" || (ch === \"&\" && s[j + 1] === \"&\") || (ch === \"|\" && s[j + 1] === \"|\"))) break;\n }\n operand = s.slice(i, j); i = j;\n }\n }\n s = s.slice(0, m.index) + `${fx(\"not\")}${operand.trim()})` + s.slice(i);\n }\n }\n\n // 6) special function transforms (order matters; emitted names are S_FN-guarded)\n s = replaceFunc(s, \"REGEX_Replace\", (a, w) => {\n if (a.length < 2) { w(\"REGEX_Replace needs 2-3 args.\"); return `regex_replace${S_FN}(${a.join(\", \")})`; }\n const pat = unlit(a[1]);\n if (a[2] && /\\$\\d/.test(unlit(a[2]) ?? a[2])) w(\"REGEX_Replace uses $-group references — verify group substitution behaves as expected in Dataiku replace().\");\n if (pat == null) { w(\"REGEX_Replace with non-literal pattern — review manually.\"); return `${fx(\"replace\")}${a[0]}, ${a[1]}, ${a[2] ?? relit(\"\")})`; }\n return `${fx(\"replace\")}${a[0]}, ${remask(\"/\" + pat.replace(/\\//g, \"\\\\/\") + \"/\")}, ${a[2] ?? relit(\"\")})`;\n }, warn);\n s = replaceFunc(s, \"REGEX_Match\", (a, w) => {\n const pat = a[1] != null ? unlit(a[1]) : null;\n if (pat == null) { w(\"REGEX_Match with non-literal pattern — review manually.\"); return `(${fx(\"match\")}${a.join(\", \")}) != null)`; }\n return `(${fx(\"match\")}${a[0]}, ${remask(\"/\" + pat.replace(/\\//g, \"\\\\/\") + \"/\")}) != null)`;\n }, warn);\n s = replaceFunc(s, \"SUBSTRING\", (a, w) => {\n if (a.length === 2) return `${fx(\"substring\")}${a[0]}, ${a[1]})`;\n if (a.length !== 3) { w(\"Substring expects 2-3 args.\"); return `${fx(\"substring\")}${a.join(\", \")})`; }\n return `${fx(\"substring\")}${a[0]}, ${a[1]}, (${a[1]}) + (${a[2]}))`; // Alteryx len -> DSS end index\n }, warn);\n s = replaceFunc(s, \"LEFT\", (a) => `${fx(\"substring\")}${a[0]}, 0, ${a[1]})`, warn);\n s = replaceFunc(s, \"RIGHT\", (a) => `${fx(\"substring\")}${a[0]}, ${fx(\"length\")}${a[0]}) - (${a[1]}), ${fx(\"length\")}${a[0]}))`, warn);\n s = replaceFunc(s, \"ROUND\", (a) => {\n const mult = a[1] != null ? a[1].trim() : \"1\";\n if (mult === \"1\") return `${fx(\"round\")}${a[0]})`;\n return `(${fx(\"round\")}(${a[0]}) / (${mult})) * (${mult}))`;\n }, warn);\n s = replaceFunc(s, \"LOG10\", (a) => `${fx(\"log\")}${a.join(\", \")})`, warn);\n s = replaceFunc(s, \"LOG\", (a) => `${fx(\"ln\")}${a.join(\", \")})`, warn); // Alteryx LOG = natural log\n s = replaceFunc(s, \"DateTimeTrim\", (a, w) => {\n const unit = (unlit(a[1] ?? \"\") || \"\").toLowerCase();\n const map: Record = { firstofmonth: \"month\", month: \"month\", firstofyear: \"year\", year: \"year\", day: \"day\", firstofday: \"day\", hour: \"hour\", minute: \"minute\" };\n if (map[unit]) return `${fx(\"trunc\")}${a[0]}, ${relit(map[unit])})`;\n w(`DateTimeTrim unit '${unit || a[1]}' has no direct trunc() unit — review manually (lastofmonth etc. need inc()+trunc()).`);\n return `${fx(\"trunc\")}${a[0]}, ${a[1]})`;\n }, warn);\n s = replaceFunc(s, \"DateTimeYear\", (a) => `${fx(\"datePart\")}${a[0]}, ${relit(\"year\")})`, warn);\n s = replaceFunc(s, \"DateTimeMonth\", (a) => `${fx(\"datePart\")}${a[0]}, ${relit(\"months\")})`, warn);\n s = replaceFunc(s, \"DateTimeDay\", (a) => `${fx(\"datePart\")}${a[0]}, ${relit(\"days\")})`, warn);\n s = replaceFunc(s, \"DateTimeHour\", (a) => `${fx(\"datePart\")}${a[0]}, ${relit(\"hours\")})`, warn);\n s = replaceFunc(s, \"DateTimeMinutes\", (a) => `${fx(\"datePart\")}${a[0]}, ${relit(\"minutes\")})`, warn);\n s = replaceFunc(s, \"DateTimeSeconds\", (a) => `${fx(\"datePart\")}${a[0]}, ${relit(\"seconds\")})`, warn);\n s = replaceFunc(s, \"DateTimeToday\", () => `${fx(\"trunc\")}${fx(\"now\")}), ${relit(\"day\")})`, warn);\n s = replaceFunc(s, \"DateTimeParse\", (a, w) => {\n const f = a[1] != null ? unlit(a[1]) : null;\n if (f == null) { w(\"DateTimeParse with non-literal format — review manually.\"); return `${fx(\"asDatetimeNoTz\")}${a.join(\", \")})`; }\n const conv: Record = { \"%Y\": \"yyyy\", \"%y\": \"yy\", \"%m\": \"MM\", \"%d\": \"dd\", \"%H\": \"HH\", \"%I\": \"hh\", \"%M\": \"mm\", \"%S\": \"ss\", \"%b\": \"MMM\", \"%B\": \"MMMM\", \"%a\": \"EEE\", \"%A\": \"EEEE\", \"%p\": \"a\", \"%j\": \"DDD\", \"%e\": \"d\", \"%l\": \"h\" };\n let java = \"\", i = 0, bad = false;\n while (i < f.length) {\n if (f[i] === \"%\") { const code = f.slice(i, i + 2); if (conv[code]) { java += conv[code]; i += 2; } else { bad = true; java += code; i += 2; } }\n else { java += f[i]; i++; }\n }\n if (bad) w(`DateTimeParse format '${f}' contains %-codes with no Java-pattern mapping — review.`);\n return `${fx(\"asDatetimeNoTz\")}${a[0]}, ${relit(java)})`;\n }, warn);\n s = replaceFunc(s, \"TrimLeft\", (a) => `${fx(\"replace\")}${a[0]}, /^\\\\s+/, ${relit(\"\")})`, warn);\n s = replaceFunc(s, \"TrimRight\", (a) => `${fx(\"replace\")}${a[0]}, /\\\\s+$/, ${relit(\"\")})`, warn);\n s = replaceFunc(s, \"Null\", () => \"null\", warn);\n\n // 7) simple renames\n const renames: Record = {\n isempty: \"isBlank\", isnull: \"isNull\", length: \"length\", uppercase: \"toUppercase\",\n lowercase: \"toLowercase\", titlecase: \"toTitlecase\", trim: \"trim\", contains: \"contains\",\n startswith: \"startsWith\", endswith: \"endsWith\", findstring: \"indexOf\", replace: \"replace\",\n tonumber: \"toNumber\", tostring: \"toString\", abs: \"abs\", ceil: \"ceil\", floor: \"floor\",\n sqrt: \"sqrt\", pow: \"pow\", exp: \"exp\", min: \"min\", max: \"max\", mod: \"mod\", rand: \"rand\",\n datetimenow: \"now\", datetimeadd: \"inc\", datetimediff: \"diff\", average: \"avg\",\n };\n for (const [from, to] of Object.entries(renames)) {\n s = s.replace(new RegExp(\"\\\\b\" + from + \"\\\\s*\\\\(\", \"gi\"), fx(to));\n }\n\n // 8) strip guards; scan unknown functions and '%' while columns+literals are still masked;\n // collapse whitespace BEFORE restoring literals (so literal contents are never rewritten)\n s = s.replace(new RegExp(S_FN, \"g\"), \"\");\n {\n const re = /\\b([A-Za-z_][A-Za-z0-9_]*)\\s*\\(/g; let m: RegExpExecArray | null;\n while ((m = re.exec(s))) {\n const low = m[1].toLowerCase();\n if (FN_WHITELIST.has(low)) continue;\n if (KNOWN_UNSUPPORTED[low]) warn(KNOWN_UNSUPPORTED[low]);\n else warn(`Function '${m[1]}' has no known Dataiku formula equivalent — MANUAL REVIEW required.`);\n }\n }\n if (/%/.test(s)) warn(\"'%' is modulo in Dataiku formulas (integer division is //) — verify intent.\");\n s = s.replace(/\\s+/g, \" \").trim();\n s = s.replace(new RegExp(S_COL + \"(\\\\d+)\" + S_COL, \"g\"), (_m, i) => colRef(cols[+i]));\n s = s.replace(new RegExp(S_LIT + \"(\\\\d+)\" + S_LIT, \"g\"), (_m, i) => lits[+i]);\n return { expr: s, warnings };\n}\n\n// ========\n// yxmd parsing\n// ========\nexport interface AXNode {\n id: string; plugin: string; short: string; container: string; annotation: string;\n cfg: any; schemas: Record;\n}\nexport interface AXConn { from: string; fromAnchor: string; to: string; toAnchor: string; order: number }\n\nexport function parseYxmd(xml: string): { nodes: AXNode[]; conns: AXConn[]; name: string } {\n const parser = new XMLParser({ ignoreAttributes: false, attributeNamePrefix: \"@\", parseTagValue: false, parseAttributeValue: false });\n const doc = parser.parse(xml);\n const ad = doc.AlteryxDocument;\n if (!ad) throw new Error(\"Not a valid .yxmd (no AlteryxDocument root)\");\n const nodes: AXNode[] = [];\n\n const extractSchemas = (props: any) => {\n const out: Record = {};\n for (const mi of arr(props?.MetaInfo)) {\n const anchor = mi?.[\"@connection\"] || \"Output\";\n const fields = arr(mi?.RecordInfo?.Field).map((f: any) => ({ name: String(f?.[\"@name\"] ?? \"\"), type: dkuType(f?.[\"@type\"]) })).filter((f: any) => f.name);\n if (fields.length) out[anchor] = fields;\n }\n return out;\n };\n\n const seenIds = new Set();\n const walk = (list: any[], container: string) => {\n for (const node of list) {\n const plugin = String(node?.GuiSettings?.[\"@Plugin\"] || \"\");\n if (/ToolContainer/i.test(plugin)) {\n const cap = node?.Properties?.Configuration?.Caption;\n walk(arr(node?.ChildNodes?.Node), String(cap || container || \"Container\"));\n continue;\n }\n let id = String(node?.[\"@ToolID\"] ?? \"?\");\n let dup = 2;\n while (seenIds.has(id)) id = String(node?.[\"@ToolID\"] ?? \"?\") + \"_dup\" + dup++;\n seenIds.add(id);\n const cfg = node?.Properties?.Configuration ?? {};\n const annotation = node?.Properties?.Annotation?.AnnotationText || \"\";\n nodes.push({\n id, plugin: plugin || \"unknown\", short: (plugin.split(\".\").pop() || \"Unknown\"),\n container, annotation: typeof annotation === \"string\" ? annotation : \"\",\n cfg, schemas: extractSchemas(node?.Properties),\n });\n }\n };\n walk(arr(ad?.Nodes?.Node), \"\");\n\n const conns: AXConn[] = arr(ad?.Connections?.Connection).map((c: any, i: number) => ({\n from: String(c?.Origin?.[\"@ToolID\"] ?? \"?\"), fromAnchor: String(c?.Origin?.[\"@Connection\"] ?? \"Output\"),\n to: String(c?.Destination?.[\"@ToolID\"] ?? \"?\"), toAnchor: String(c?.Destination?.[\"@Connection\"] ?? \"Input\"),\n order: i,\n }));\n let name = ad?.Properties?.MetaInfo?.Name;\n if (typeof name !== \"string\" || !name.trim()) name = \"Converted_Workflow\";\n return { nodes, conns, name };\n}\n\n// ========\n// DSS JSON builders (formats mirror real DSS 7/8 project exports)\n// ========\nfunction managedDatasetJson(projectKey: string, name: string, schema: { name: string; type: string }[] | null, tags: string[]) {\n return {\n type: \"Filesystem\", managed: true, formatType: \"csv\",\n checklists: { checklists: [] }, customMeta: { kv: {} },\n flowOptions: { virtualizable: false, rebuildBehavior: \"NORMAL\", crossProjectBuildBehavior: \"DEFAULT\" },\n readWriteOptions: { preserveOrder: false, writeBuckets: 1, forceSingleOutputFile: false, defaultReadOrdering: { enabled: false, rules: [] } },\n formatParams: {\n style: \"unix\", charset: \"utf8\", separator: \"\\t\", quoteChar: '\"', escapeChar: \"\\\\\",\n dateSerializationFormat: \"ISO\", arrayMapFormat: \"json\",\n hiveSeparators: [\"\u0002\", \"\u0003\", \"\u0004\", \"\u0005\", \"\u0006\", \"\u0007\", \"\\b\"],\n skipRowsBeforeHeader: 0, parseHeaderRow: false, skipRowsAfterHeader: 0,\n probableNumberOfRecords: 0, normalizeBooleans: false, normalizeDoubles: true,\n readAdditionalColumnsBehavior: \"INSERT_IN_DATA_WARNING\", readMissingColumnsBehavior: \"DISCARD_SILENT\",\n readDataTypeMismatchBehavior: \"DISCARD_WARNING\", writeDataTypeMismatchBehavior: \"DISCARD_WARNING\",\n fileReadFailureBehavior: \"FAIL\", compress: \"gz\",\n },\n partitioning: { ignoreNonMatchingFile: false, considerMissingRequestedPartitionsAsEmpty: false, dimensions: [] },\n versionTag: VT, creationTag: VT, tags,\n params: { connection: \"filesystem_managed\", path: \"${projectKey}/\" + name, notReadyIfEmpty: false, filesSelectionRules: { mode: \"ALL\", excludeRules: [], includeRules: [], explicitFiles: [] } },\n schema: { columns: (schema || []).map(c => ({ name: c.name, type: c.type })), userModified: !!(schema && schema.length) },\n customFields: {},\n };\n}\n\nfunction uploadedDatasetJson(name: string, schema: { name: string; type: string }[] | null, tags: string[]) {\n return {\n type: \"UploadedFiles\", managed: false, formatType: \"csv\",\n checklists: { checklists: [] }, customMeta: { kv: {} },\n flowOptions: { virtualizable: false, rebuildBehavior: \"NORMAL\", crossProjectBuildBehavior: \"DEFAULT\" },\n readWriteOptions: { preserveOrder: false, writeBuckets: 1, forceSingleOutputFile: false, defaultReadOrdering: { enabled: false, rules: [] } },\n formatParams: {\n style: \"excel\", charset: \"utf-8\", separator: \",\", quoteChar: '\"', escapeChar: \"\\\\\",\n dateSerializationFormat: \"ISO\", arrayMapFormat: \"json\",\n hiveSeparators: [\"\u0002\", \"\u0003\", \"\u0004\", \"\u0005\", \"\u0006\", \"\u0007\", \"\\b\"],\n skipRowsBeforeHeader: 0, parseHeaderRow: true, skipRowsAfterHeader: 0,\n probableNumberOfRecords: 0, normalizeBooleans: true, normalizeDoubles: true,\n readAdditionalColumnsBehavior: \"INSERT_IN_DATA_WARNING\", readMissingColumnsBehavior: \"DISCARD_SILENT\",\n readDataTypeMismatchBehavior: \"DISCARD_WARNING\", writeDataTypeMismatchBehavior: \"DISCARD_WARNING\",\n fileReadFailureBehavior: \"FAIL\", compress: \"\",\n },\n partitioning: { ignoreNonMatchingFile: false, considerMissingRequestedPartitionsAsEmpty: false, dimensions: [] },\n versionTag: VT, creationTag: VT, tags,\n params: { uploadConnection: \"Default (in DSS data dir.)\", notReadyIfEmpty: false, filesSelectionRules: { mode: \"ALL\", excludeRules: [], includeRules: [], explicitFiles: [] } },\n schema: { columns: (schema || []).map(c => ({ name: c.name, type: c.type })), userModified: !!(schema && schema.length) },\n customFields: {},\n };\n}\n\nfunction recipeJson(type: string, inputs: string[], outputs: string[], tags: string[], paramsExtra?: any) {\n const j: any = {\n type, neverRecomputeExistingPartitions: false, optionalDependencies: false,\n customMeta: { kv: {} }, redispatchPartitioning: false, maxRunningActivities: 0,\n inputs: { main: { items: inputs.map(r => ({ ref: r, deps: [] })) } },\n outputs: { main: { items: outputs.map(r => ({ ref: r, appendMode: false })) } },\n versionTag: VT, creationTag: VT, tags, customFields: {}, checklists: { checklists: [] },\n };\n if (paramsExtra) j.params = paramsExtra;\n return j;\n}\n\nconst PY_RECIPE_PARAMS = { envSelection: { envMode: \"INHERIT\" }, containerSelection: { containerMode: \"INHERIT\" }, skipPrerunValidate: false };\n\nfunction shakerPayload(steps: any[]) {\n return {\n columnsSelection: { mode: \"ALL\" },\n explorationSampling: {\n _refreshTrigger: 0,\n selection: {\n filter: { distinct: false, enabled: false }, latestPartitionsN: 1, maxRecords: 10000,\n ordering: { rules: [], enabled: false }, withinFirstN: -1, partitionSelectionMethod: \"ALL\",\n maxStoredBytes: 104857600, targetRatio: 0.02, maxReadUncompressedBytes: -1, samplingMethod: \"HEAD_SEQUENTIAL\",\n },\n autoRefreshSample: false,\n },\n explorationFilters: [], origin: \"PREPARE_RECIPE\", exploreUIParams: { autoRefresh: true },\n steps,\n maxProcessedMemTableBytes: -1, previewMode: \"ALL_ROWS\",\n vizSampling: { _refreshTrigger: 0, autoRefreshSample: false },\n analysisColumnData: {}, sorting: [], globalSearchQuery: \"\",\n coloring: { scheme: \"MEANING_AND_STATUS\", individualColumns: [], valueColoringMode: \"HASH\" },\n };\n}\n\nconst step = (type: string, params: any, name?: string) => ({ preview: false, metaType: \"PROCESSOR\", disabled: false, type, params, alwaysShowComment: false, ...(name ? { name } : {}) });\n\nfunction joinPayload(joinType: string, lkeys: string[], rkeys: string[], leftSchema: { name: string; type: string }[], rightSchema: { name: string; type: string }[]) {\n const rk = new Set(rkeys);\n const selected = [\n ...leftSchema.map(c => ({ name: c.name, type: c.type, table: 0 })),\n ...rightSchema.filter(c => !rk.has(c.name) || !lkeys.includes(c.name)).map(c => ({ name: c.name, type: c.type, table: 1 })),\n ];\n return {\n joins: [{\n table1: 0, table2: 1, conditionsMode: \"AND\", rightLimit: { decisionColumn: {} },\n type: joinType, outerJoinOnTheLeft: true,\n on: lkeys.map((k, i) => ({\n column1: { name: k, table: 0 }, column2: { name: rkeys[i], table: 1 },\n type: \"EQ\", maxMatches: 1, caseInsensitive: false, maxDistance: 0, normalizeText: false, strict: false, dateDiffUnit: \"DAY\",\n })),\n }],\n selectedColumns: selected,\n virtualInputs: [\n { preFilter: { distinct: false, enabled: false }, autoSelectColumns: false, index: 0, computedColumns: [] },\n { preFilter: { distinct: false, enabled: false }, autoSelectColumns: false, index: 1, computedColumns: [] },\n ],\n computedColumns: [], postFilter: { enabled: false, distinct: false }, enableAutoCastInJoinConditions: false,\n };\n}\n\n// Shared parser for an Alteryx Select-style field list (used by AlteryxSelect AND the\n// Select panel embedded inside a Join). Returns Prepare steps + warnings.\nfunction buildSelectSteps(sfs: any[], inputSchema: { name: string; type: string }[] | null): { steps: any[]; warns: string[]; renames: { from: string; to: string }[] } {\n const arrSfs = arr(sfs);\n const deselected = arrSfs.filter((f: any) => String(f?.[\"@selected\"]).toLowerCase() === \"false\" && f?.[\"@field\"] !== \"*Unknown\").map((f: any) => String(f[\"@field\"]));\n const unknownKept = !arrSfs.some((f: any) => f?.[\"@field\"] === \"*Unknown\" && String(f?.[\"@selected\"]).toLowerCase() === \"false\");\n const renames = arrSfs.filter((f: any) => f?.[\"@rename\"] && String(f?.[\"@selected\"]).toLowerCase() !== \"false\").map((f: any) => ({ from: String(f[\"@field\"]), to: String(f[\"@rename\"]) }));\n const steps: any[] = []; const warns: string[] = [];\n if (!unknownKept && inputSchema) {\n const explicit = new Set(arrSfs.map((f: any) => String(f?.[\"@field\"])));\n const kept = inputSchema.map(c => c.name).filter(c => (!explicit.has(c) ? false : !deselected.includes(c)));\n if (kept.length) steps.push(step(\"ColumnsSelector\", { columns: kept, keep: true, appliesTo: \"COLUMNS\" }));\n } else {\n if (deselected.length) steps.push(step(\"ColumnsSelector\", { columns: deselected, keep: false, appliesTo: \"COLUMNS\" }));\n if (!unknownKept && !inputSchema) warns.push(\"'*Unknown' was deselected but the input schema is unknown — extra columns may pass through; review.\");\n }\n if (renames.length) steps.push(step(\"ColumnRenamer\", { renamings: renames }));\n return { steps, warns, renames };\n}\n\nconst AGG_FLAGS = [\"min\", \"max\", \"count\", \"last\", \"sum\", \"concat\", \"countDistinct\", \"stddev\", \"avg\", \"first\"];\nfunction groupingKey(col: string, type: string) {\n const o: any = { column: col, type };\n for (const f of AGG_FLAGS) o[f] = false;\n return o;\n}\nfunction groupingValue(col: string, type: string, flag: string, idx: number, concatSeparator?: string) {\n const o: any = { column: col, type, $idx: idx };\n for (const f of AGG_FLAGS) o[f] = f === flag;\n if (flag === \"concat\") { o.concatDistinct = false; o.concatSeparator = concatSeparator ?? \",\"; }\n return o;\n}\nfunction groupingPayload(keys: any[], values: any[]) {\n return {\n outputColumnNameOverrides: {}, keys, values, globalCount: false, computedColumns: [],\n preFilter: { enabled: false, distinct: false }, postFilter: { enabled: false, distinct: false },\n };\n}\n\nfunction vstackPayload(inputNames: string[], columns: string[]) {\n return {\n mode: \"UNION\", selectedColumns: columns, originColumnName: \"original_dataset\",\n virtualInputs: inputNames.map((n, i) => ({ preFilter: {}, originLabel: n, index: i })),\n postFilter: { enabled: false, distinct: false }, addOriginColumn: false,\n };\n}\n\nfunction sortPayload(orders: { column: string; desc: boolean }[]) {\n return { orders, rowNumber: false, rank: false, denseRank: false, preFilter: { enabled: false, distinct: false }, postFilter: { enabled: false, distinct: false } };\n}\n\n// ---------------- python recipe code builders ----------------\nfunction pyHeader(node: AXNode, note: string): string[] {\n const L = [\n `# -------------------------------------------------------------------`,\n `# Converted from Alteryx tool ${node.id} (${node.short})${node.container ? ` [container: ${node.container}]` : \"\"}`,\n ];\n if (node.annotation) L.push(`# Annotation: ${node.annotation.replace(/\\r?\\n/g, \" | \")}`);\n if (note) L.push(`# ${note}`);\n L.push(`# -------------------------------------------------------------------`, `import dataiku`, `import pandas as pd`, ``);\n return L;\n}\nconst pyRead = (ds: string, as_: string) => `${as_} = dataiku.Dataset(${JSON.stringify(ds)}).get_dataframe()`;\nconst pyWrite = (ds: string, from_: string) => `dataiku.Dataset(${JSON.stringify(ds)}).write_with_schema(${from_})`;\n\nfunction pyPassthrough(node: AXNode, inDs: string, outDs: string, extra: string): string {\n return [...pyHeader(node, `NO AUTOMATIC MAPPING — passthrough placeholder. TODO: implement the original logic.${extra ? \" \" + extra : \"\"}`),\n pyRead(inDs, \"df\"), ``, `# TODO: replicate the Alteryx '${node.short}' logic here`, ``, pyWrite(outDs, \"df\"), ``].join(\"\\n\");\n}\n\n// ========\n// Conversion core\n// ========\ninterface RecipeOut { name: string; json: any; payloadExt: string; payload: string }\ninterface ReportRow { toolId: string; alteryxTool: string; kind: string; objects: string[]; notes: string[]; warnings: string[] }\n\nexport function convert(xml: string) {\n const { nodes, conns, name } = parseYxmd(xml);\n const projectKey = (sanitizeName(name, 30).toUpperCase() || \"MIGRATED\");\n const todos: string[] = [];\n const report: ReportRow[] = [];\n const datasets: Record = {}; // name -> dataset json\n const recipes: RecipeOut[] = [];\n const uploadsFiles: Record = {}; // zip path -> bytes\n const includedDatasets: any[] = [];\n const usedNames = new Set();\n const usedRecipeNames = new Set();\n\n const uniqueName = (base: string) => {\n let n = sanitizeName(base); let i = 2;\n while (usedNames.has(n)) n = sanitizeName(base, 52) + \"_\" + i++;\n usedNames.add(n); return n;\n };\n const uniqueRecipeName = (base: string) => {\n let n = sanitizeName(base, 60); let i = 2;\n while (usedRecipeNames.has(n)) n = sanitizeName(base, 56) + \"_\" + i++;\n usedRecipeNames.add(n); return n;\n };\n\n const byId: Record = {};\n for (const n of nodes) byId[n.id] = n;\n const inbound = (id: string) => conns.filter(c => c.to === id).sort((a, b) => a.order - b.order);\n const outbound = (id: string) => conns.filter(c => c.from === id);\n const isWired = (id: string, anchor: string) => conns.some(c => c.from === id && c.fromAnchor === anchor);\n\n // ---- pass 1: register output datasets per (tool, anchor) ----\n interface OutReg { ds: string; schema: { name: string; type: string }[] | null }\n const outMap: Record = {}; // `${id}|${anchor}`\n const reg = (node: AXNode, anchor: string, dsName: string) => {\n const schema = node.schemas[anchor] || (anchor === \"Output\" ? node.schemas[\"\"] : null) || null;\n outMap[`${node.id}|${anchor}`] = { ds: dsName, schema };\n return dsName;\n };\n const lookupOut = (id: string, anchor: string): OutReg | null =>\n outMap[`${id}|${anchor}`] || outMap[`${id}|Output`] || Object.entries(outMap).find(([k]) => k.startsWith(id + \"|\"))?.[1] || null;\n\n const OMIT = /BrowseV2|ToolContainer/i;\n const anchorsFor = (node: AXNode): string[] => {\n const p = node.plugin;\n if (/\\.Filter\\./.test(p)) return [\"True\", \"False\"];\n if (/\\.Join\\.Join/.test(p)) return [\"Join\", \"Left\", \"Right\"];\n if (/\\.Unique\\./.test(p)) return [\"Unique\", \"Dupes\"];\n return [\"Output\"];\n };\n\n for (const node of nodes) {\n if (OMIT.test(node.plugin)) continue;\n const short = node.short;\n const anchors = anchorsFor(node);\n for (const a of anchors) {\n const suffix = anchors.length > 1 ? \"_\" + a[0] : \"\";\n let base = `t${node.id}_${short}${suffix}`;\n if (/DbFileInput/i.test(node.plugin) || /LockInInput/i.test(node.plugin)) {\n const f = typeof node.cfg?.File === \"string\" ? node.cfg.File : node.cfg?.File?.[\"#text\"] || \"\";\n const fb = String(f).split(/[\\\\/]/).pop()?.split(\".\")[0] || \"input\";\n base = `t${node.id}_in_${fb}`;\n } else if (/TextInput/i.test(node.plugin)) {\n base = `t${node.id}_TextInput`;\n } else if (/DbFileOutput/i.test(node.plugin)) {\n const f = typeof node.cfg?.File === \"string\" ? node.cfg.File : node.cfg?.File?.[\"#text\"] || \"\";\n const fb = String(f).split(/[\\\\/]/).pop()?.split(\".\")[0] || \"output\";\n base = `t${node.id}_out_${fb}`;\n }\n reg(node, a, uniqueName(base));\n // TextInput: derive schema from its Fields config when no MetaInfo is present\n if (/TextInput/i.test(node.plugin) && !outMap[`${node.id}|${a}`].schema) {\n const fields = arr(node.cfg?.Fields?.Field).map((f: any) => ({ name: String(f?.[\"@name\"] ?? \"\"), type: \"string\" })).filter((f: any) => f.name);\n if (fields.length) outMap[`${node.id}|${a}`].schema = fields;\n }\n }\n }\n\n // resolve the input dataset feeding (toolId, destAnchor)\n const inputFor = (node: AXNode, destAnchor?: string): OutReg | null => {\n const cs = inbound(node.id);\n const c = destAnchor ? cs.find(x => x.toAnchor === destAnchor) || null : cs[0] || null;\n if (!c) return null;\n return lookupOut(c.from, c.fromAnchor);\n };\n\n const addManaged = (dsName: string, schema: { name: string; type: string }[] | null, tags: string[]) => {\n if (!datasets[dsName]) datasets[dsName] = managedDatasetJson(projectKey, dsName, schema, tags);\n };\n const addRecipe = (type: string, baseName: string, inputs: string[], outputs: string[], tags: string[], payloadExt: string, payload: string, paramsExtra?: any) => {\n const rn = uniqueRecipeName(baseName);\n recipes.push({ name: rn, json: recipeJson(type, inputs, outputs, tags, paramsExtra), payloadExt, payload });\n return rn;\n };\n const rrow = (node: AXNode, kind: string, objects: string[], notes: string[] = [], warnings: string[] = []): ReportRow => {\n const r = { toolId: node.id, alteryxTool: node.short, kind, objects, notes, warnings };\n report.push(r);\n for (const w of warnings) todos.push(`Tool ${node.id} (${node.short}): ${w}`);\n return r;\n };\n\n // ---- pass 2: build ----\n for (const node of nodes) {\n if (OMIT.test(node.plugin)) {\n if (/BrowseV2/i.test(node.plugin)) rrow(node, \"omitted\", [], [\"Use the dataset Explore tab instead.\"]);\n continue;\n }\n const tags = node.container ? [node.container] : [];\n const cfg = node.cfg || {};\n const p = node.plugin;\n const primaryOut = lookupOut(node.id, \"Output\")!;\n\n // ---------- INPUTS ----------\n if (/\\.TextInput\\.TextInput/i.test(p)) {\n const fields = arr(cfg?.Fields?.Field).map((f: any) => String(f?.[\"@name\"] ?? \"\"));\n const rows = arr(cfg?.Data?.r).map((r: any) => arr(r?.c).map((c: any) => (c && typeof c === \"object\" ? c[\"#text\"] ?? \"\" : c ?? \"\")));\n const schema = primaryOut.schema || fields.map(f => ({ name: f, type: \"string\" }));\n datasets[primaryOut.ds] = uploadedDatasetJson(primaryOut.ds, schema, tags);\n const csv = [fields.map(csvEscape).join(\",\"), ...rows.map(r => r.map(csvEscape).join(\",\"))].join(\"\\n\") + \"\\n\";\n const bytes = strToU8(csv);\n uploadsFiles[`uploads/${primaryOut.ds}/${primaryOut.ds}.csv`] = bytes;\n includedDatasets.push({ name: primaryOut.ds, type: \"UploadedFiles\", usedTransferMethod: \"UPLOAD\", size: bytes.length });\n rrow(node, \"input dataset (data embedded)\", [primaryOut.ds], [`${rows.length} row(s) of Text Input data embedded in the zip.`]);\n continue;\n }\n if (/\\.DbFileInput\\./i.test(p) || /LockInGui\\.LockInInput/i.test(p)) {\n const f = typeof cfg?.File === \"string\" ? cfg.File : cfg?.File?.[\"#text\"] || \"\";\n datasets[primaryOut.ds] = uploadedDatasetJson(primaryOut.ds, primaryOut.schema, tags);\n const notes = [`Source: ${f || \"unknown\"}. After import, open this dataset and upload the file (Settings > Files).`];\n const warns: string[] = [];\n if (/LockInInput/i.test(p)) warns.push(`Was an In-DB input — recreate as an SQL dataset on your database connection if needed (placeholder UploadedFiles dataset created).`);\n if (!primaryOut.schema) warns.push(`No schema metadata in the yxmd for input '${primaryOut.ds}' — DSS will infer it from the uploaded file.`);\n rrow(node, \"input dataset\", [primaryOut.ds], notes, warns);\n continue;\n }\n\n // ---------- generic input resolution ----------\n const firstIn = inputFor(node);\n\n // helper: emit passthrough python when nothing else fits\n const passthrough = (extraNote: string, warns: string[] = []) => {\n if (!firstIn) {\n datasets[primaryOut.ds] = uploadedDatasetJson(primaryOut.ds, primaryOut.schema, tags);\n rrow(node, \"input dataset (placeholder)\", [primaryOut.ds], [], [`Unmapped source tool '${node.short}' — placeholder dataset created. ${extraNote}`.trim()]);\n return;\n }\n addManaged(primaryOut.ds, primaryOut.schema || firstIn.schema, tags);\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", pyPassthrough(node, firstIn.ds, primaryOut.ds, extraNote), PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (passthrough TODO)\", [rn, primaryOut.ds], [], [`No automatic mapping — passthrough Python recipe keeps the flow runnable. Implement the '${node.short}' logic inside it. ${extraNote}`.trim(), ...warns]);\n };\n\n // ---------- FORMULA ----------\n if (/\\.Formula\\.Formula$/i.test(p)) {\n if (!firstIn) { passthrough(\"Formula tool had no wired input.\"); continue; }\n const steps: any[] = []; const warns: string[] = []; const notes: string[] = [];\n for (const fm of arr(cfg?.FormulaFields?.FormulaField)) {\n const field = String(fm?.[\"@field\"] ?? \"\");\n const t = translateFormula(String(fm?.[\"@expression\"] ?? \"\"));\n steps.push(step(\"CreateColumnWithGREL\", { column: field, expression: t.expr }));\n notes.push(`${field} = ${t.expr}`);\n warns.push(...t.warnings.map(w => `formula '${field}': ${w}`));\n }\n addManaged(primaryOut.ds, primaryOut.schema || firstIn.schema, tags);\n const rn = addRecipe(\"shaker\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"shaker\", JSON.stringify(shakerPayload(steps), null, 1));\n rrow(node, \"Prepare recipe (formulas)\", [rn, primaryOut.ds], notes, warns);\n continue;\n }\n\n // ---------- SELECT ----------\n if (/\\.AlteryxSelect\\.AlteryxSelect$/i.test(p)) {\n if (!firstIn) { passthrough(\"Select tool had no wired input.\"); continue; }\n const sfs = arr(cfg?.SelectFields?.SelectField);\n const { steps, warns } = buildSelectSteps(sfs, firstIn.schema);\n const typeChanges = sfs.filter((f: any) => f?.[\"@type\"]).length;\n if (typeChanges) warns.push(`Select changes ${typeChanges} column type(s) — storage types are set on the output dataset schema in DSS; verify after first build.`);\n addManaged(primaryOut.ds, primaryOut.schema, tags);\n const rn = addRecipe(\"shaker\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"shaker\", JSON.stringify(shakerPayload(steps), null, 1));\n rrow(node, \"Prepare recipe (select/rename)\", [rn, primaryOut.ds], [], warns);\n continue;\n }\n\n // ---------- FILTER ----------\n if (/\\.Filter\\.Filter$/i.test(p)) {\n if (!firstIn) { passthrough(\"Filter tool had no wired input.\"); continue; }\n let exprSrc = typeof cfg?.Expression === \"string\" ? cfg.Expression : \"\";\n const warns: string[] = []; const notes: string[] = [];\n if (!exprSrc && cfg?.Simple) {\n const sm = cfg.Simple;\n const field = String(sm?.Field ?? \"\");\n const op = String(sm?.Operator ?? \"=\");\n const operands = arr(sm?.Operands?.Operand).map((o: any) => (o && typeof o === \"object\" ? o[\"#text\"] ?? \"\" : o ?? \"\"));\n const v0 = String(operands[0] ?? \"\");\n const lit = /^-?\\d+(\\.\\d+)?$/.test(v0) ? v0 : `'${v0.replace(/'/g, \"\\\\'\")}'`;\n const f = colRef(field);\n const m: Record = {\n \"=\": `${f} == ${lit}`, \"!=\": `${f} != ${lit}`, \">\": `${f} > ${lit}`, \">=\": `${f} >= ${lit}`,\n \"<\": `${f} < ${lit}`, \"<=\": `${f} <= ${lit}`, \"Contains\": `contains(${f}, ${lit})`,\n \"DoesNotContain\": `not(contains(${f}, ${lit}))`, \"IsNull\": `isNull(${f})`, \"Null\": `isNull(${f})`,\n \"IsNotNull\": `not(isNull(${f}))`, \"NotNull\": `not(isNull(${f}))`, \"IsEmpty\": `isBlank(${f})`, \"IsNotEmpty\": `isNonBlank(${f})`,\n };\n let builtExpr: string;\n if (m[op]) { notes.push(`basic mode: ${field} ${op} ${v0}`); builtExpr = m[op]; }\n else { warns.push(`Basic-mode filter operator '${op}' not translated — the step keeps ALL rows (expression 'true'); set the filter formula manually.`); builtExpr = \"true\"; }\n const outs: [string, string][] = [];\n if (isWired(node.id, \"True\")) outs.push([\"True\", \"KEEP_ROW\"]);\n if (isWired(node.id, \"False\")) outs.push([\"False\", \"REMOVE_ROW\"]);\n if (!outs.length) outs.push([\"True\", \"KEEP_ROW\"]);\n const objs: string[] = [];\n for (const [anchor, action] of outs) {\n const o = lookupOut(node.id, anchor)!;\n addManaged(o.ds, o.schema || firstIn.schema, tags);\n const rn = addRecipe(\"shaker\", `compute_${o.ds}`, [firstIn.ds], [o.ds], tags, \"shaker\",\n JSON.stringify(shakerPayload([step(\"FilterOnFormula\", { expression: builtExpr, action })]), null, 1));\n objs.push(rn, o.ds);\n }\n rrow(node, \"Prepare recipe(s) (filter)\", objs, notes, warns);\n continue;\n }\n const t = translateFormula(exprSrc || \"\");\n if (!exprSrc) warns.push(\"Filter expression missing/unparsed — FilterOnFormula step left with placeholder 'true'; set it manually.\");\n warns.push(...t.warnings);\n const expr = exprSrc ? t.expr : \"true\";\n notes.push(`filter: ${expr}`);\n const outs: [string, string][] = [];\n if (isWired(node.id, \"True\")) outs.push([\"True\", \"KEEP_ROW\"]);\n if (isWired(node.id, \"False\")) outs.push([\"False\", \"REMOVE_ROW\"]);\n if (!outs.length) outs.push([\"True\", \"KEEP_ROW\"]);\n const objs: string[] = [];\n for (const [anchor, action] of outs) {\n const o = lookupOut(node.id, anchor)!;\n addManaged(o.ds, o.schema || firstIn.schema, tags);\n const rn = addRecipe(\"shaker\", `compute_${o.ds}`, [firstIn.ds], [o.ds], tags, \"shaker\",\n JSON.stringify(shakerPayload([step(\"FilterOnFormula\", { expression: expr, action })]), null, 1));\n objs.push(rn, o.ds);\n }\n rrow(node, \"Prepare recipe(s) (filter)\", objs, notes, warns);\n continue;\n }\n\n // ---------- SORT ----------\n if (/\\.Sort\\.Sort$/i.test(p)) {\n if (!firstIn) { passthrough(\"Sort had no wired input.\"); continue; }\n const orders = arr(cfg?.SortInfo?.Field).map((f: any) => ({ column: String(f?.[\"@field\"] ?? \"\"), desc: /desc/i.test(String(f?.[\"@order\"] ?? \"\")) }));\n addManaged(primaryOut.ds, primaryOut.schema || firstIn.schema, tags);\n const rn = addRecipe(\"sort\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"sort\", JSON.stringify(sortPayload(orders), null, 1));\n rrow(node, \"Sort recipe\", [rn, primaryOut.ds], orders.map(o => `${o.column} ${o.desc ? \"desc\" : \"asc\"}`));\n continue;\n }\n\n // ---------- JOIN ----------\n if (/\\.Join\\.Join$/i.test(p)) {\n const leftIn = inputFor(node, \"Left\"); const rightIn = inputFor(node, \"Right\");\n if (!leftIn || !rightIn) { passthrough(\"Join is missing a wired Left/Right input.\"); continue; }\n const lkeys: string[] = []; const rkeys: string[] = [];\n for (const ji of arr(cfg?.JoinInfo)) {\n const side = String(ji?.[\"@connection\"] ?? \"\");\n const fields = arr(ji?.Field).map((f: any) => String(f?.[\"@field\"] ?? \"\"));\n if (/left/i.test(side)) lkeys.push(...fields); else if (/right/i.test(side)) rkeys.push(...fields);\n }\n const warns: string[] = []; const objs: string[] = [];\n const keysOk = !!lkeys.length && lkeys.length === rkeys.length;\n if (!keysOk) warns.push(\"Could not fully parse join keys — verify the join conditions.\");\n\n // Embedded Select panel inside the Join (drops/renames on an output anchor).\n const selectConfigFor = (anchor: string): any[] => {\n for (const c of arr(cfg?.SelectConfiguration?.Configuration)) {\n if (String(c?.[\"@outputConnection\"] ?? \"Join\").toLowerCase() === anchor.toLowerCase()) return arr(c?.SelectFields?.SelectField);\n }\n return [];\n };\n\n // J anchor\n if (isWired(node.id, \"Join\") || (!isWired(node.id, \"Left\") && !isWired(node.id, \"Right\"))) {\n const oJ = lookupOut(node.id, \"Join\")!;\n\n // Effective input schemas: if one side is missing but the join's own output schema\n // is known, derive the other side from it (safe only when there is no name clash,\n // which Alteryx marks with a 'Right_' prefix in the output).\n let Lsch = leftIn.schema, Rsch = rightIn.schema;\n const outSchema = oJ.schema;\n if (keysOk && outSchema && (!Lsch || !Rsch) && (Lsch || Rsch) && !outSchema.some(c => /^Right_/.test(c.name))) {\n if (Lsch && !Rsch) { const ln = new Set(Lsch.map(c => c.name)); Rsch = outSchema.filter(c => !ln.has(c.name)).map(c => ({ name: c.name, type: c.type })); }\n else if (Rsch && !Lsch) { const rn = new Set(Rsch.map(c => c.name)); Lsch = outSchema.filter(c => !rn.has(c.name)).map(c => ({ name: c.name, type: c.type })); }\n if (leftIn.schema !== Lsch || rightIn.schema !== Rsch) warns.push(\"One input schema was missing — reconstructed it from the join's output metadata to keep a visual Join recipe (verify columns after first build).\");\n }\n const clash = (Lsch && Rsch)\n ? Rsch.map(c => c.name).filter(cn => Lsch!.some(lc => lc.name === cn) && !(rkeys.includes(cn) && lkeys[rkeys.indexOf(cn)] === cn))\n : null;\n\n const embSel = selectConfigFor(\"Join\");\n const hasEmbSel = embSel.some((f: any) => String(f?.[\"@selected\"]).toLowerCase() === \"false\" || f?.[\"@rename\"]);\n const joinOut = hasEmbSel ? uniqueName(`t${node.id}_joined`) : oJ.ds;\n if (hasEmbSel) addManaged(joinOut, null, tags);\n addManaged(oJ.ds, oJ.schema, tags);\n\n if (Lsch && Rsch && clash && clash.length === 0 && keysOk) {\n const rn = addRecipe(\"join\", `compute_${joinOut}`, [leftIn.ds, rightIn.ds], [joinOut], tags, \"join\",\n JSON.stringify(joinPayload(\"INNER\", lkeys, rkeys, Lsch, Rsch), null, 1));\n objs.push(rn, joinOut);\n } else {\n const reason = !Lsch || !Rsch ? \"input schemas unknown and not derivable from output metadata\" : (clash && clash.length ? `overlapping non-key columns (${clash.slice(0, 5).join(\", \")}) get Alteryx's Right_ prefix` : \"join keys unparsed\");\n const code = [...pyHeader(node, `Inner join (J output). Python used because: ${reason}.`),\n pyRead(leftIn.ds, \"left\"), pyRead(rightIn.ds, \"right\"),\n `lk = ${JSON.stringify(lkeys)}`, `rk = ${JSON.stringify(rkeys)}`,\n `assert lk and len(lk) == len(rk), \"TODO: fix join keys (could not be fully parsed from the yxmd)\"`,\n `same = [k for i, k in enumerate(rk) if i < len(lk) and lk[i] == k]`,\n `ren = {c: \"Right_\" + c for c in right.columns if c in set(left.columns)}`,\n `r2 = right.rename(columns=ren)`,\n `rk2 = [ren.get(k, k) for k in rk]`,\n `out = left.merge(r2, left_on=lk, right_on=rk2, how=\"inner\")`,\n `out = out.drop(columns=[ren[k] for k in same if k in ren]) # Alteryx dedupes equal-named join keys`,\n ``, pyWrite(joinOut, \"out\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${joinOut}`, [leftIn.ds, rightIn.ds], [joinOut], tags, \"py\", code, PY_RECIPE_PARAMS);\n warns.push(`J output implemented as a Python recipe (${reason}).`);\n objs.push(rn, joinOut);\n }\n\n // Replicate the embedded Select panel as a Prepare recipe after the join.\n if (hasEmbSel) {\n const sel = buildSelectSteps(embSel, oJ.schema);\n const rn2 = addRecipe(\"shaker\", `compute_${oJ.ds}`, [joinOut], [oJ.ds], tags, \"shaker\",\n JSON.stringify(shakerPayload(sel.steps.length ? sel.steps : []), null, 1));\n warns.push(...sel.warns.map(w => `join column selection: ${w}`));\n objs.push(rn2, oJ.ds);\n }\n }\n // L / R anchors (unmatched rows) as python anti-joins\n for (const [anchor, sideA, sideB, aKeys, bKeys] of [[\"Left\", \"left\", \"right\", lkeys, rkeys], [\"Right\", \"right\", \"left\", rkeys, lkeys]] as any[]) {\n if (!isWired(node.id, anchor)) continue;\n const o = lookupOut(node.id, anchor)!;\n const aIn = anchor === \"Left\" ? leftIn : rightIn; const bIn = anchor === \"Left\" ? rightIn : leftIn;\n addManaged(o.ds, o.schema || aIn.schema, tags);\n const code = [...pyHeader(node, `${anchor} output = rows of the ${sideA} input with no match (anti-join).`),\n pyRead(aIn.ds, \"a\"), pyRead(bIn.ds, \"b\"),\n `ak = ${JSON.stringify(aKeys)}`, `bk = ${JSON.stringify(bKeys)}`,\n `bkeys = b[bk].drop_duplicates()`,\n `m = a[ak].merge(bkeys, left_on=ak, right_on=bk, how=\"left\", indicator=True)[\"_merge\"]`,\n `out = a[m.eq(\"left_only\").values]`,\n ``, pyWrite(o.ds, \"out\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${o.ds}`, [aIn.ds, bIn.ds], [o.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n objs.push(rn, o.ds);\n }\n rrow(node, \"Join\", objs, [`keys: ${lkeys.map((k, i) => `${k}=${rkeys[i] ?? \"?\"}`).join(\", \") || \"?\"}`], warns);\n continue;\n }\n\n // ---------- UNION ----------\n if (/\\.Union\\.Union$/i.test(p)) {\n const ins = inbound(node.id).map(c => lookupOut(c.from, c.fromAnchor)).filter(Boolean) as OutReg[];\n if (!ins.length) { passthrough(\"Union had no wired inputs.\"); continue; }\n const mode = String(cfg?.Mode ?? \"ByName\");\n addManaged(primaryOut.ds, primaryOut.schema || ins[0].schema, tags);\n const known = primaryOut.schema || (ins.every(i => i.schema) ? (() => { const seen: string[] = []; for (const i of ins) for (const c of i.schema!) if (!seen.includes(c.name)) seen.push(c.name); return seen.map(n => ({ name: n, type: \"string\" })); })() : null);\n if (/byname/i.test(mode) && known) {\n const rn = addRecipe(\"vstack\", `compute_${primaryOut.ds}`, ins.map(i => i.ds), [primaryOut.ds], tags, \"vstack\",\n JSON.stringify(vstackPayload(ins.map(i => i.ds), known.map(c => c.name)), null, 1));\n rrow(node, \"Stack recipe (union by name)\", [rn, primaryOut.ds], [`${ins.length} inputs`]);\n } else {\n const code = [...pyHeader(node, `Union (${mode}).`),\n `frames = [dataiku.Dataset(n).get_dataframe() for n in ${JSON.stringify(ins.map(i => i.ds))}]`,\n ...(/bypos/i.test(mode) ? [\n `base = list(frames[0].columns)`,\n `frames = [f.set_axis(base[:len(f.columns)] + list(f.columns[len(base):]), axis=1) if len(f.columns) >= len(base) else f.set_axis(base[:len(f.columns)], axis=1) for f in frames]`,\n ] : []),\n `out = pd.concat(frames, ignore_index=True, sort=False)`,\n ``, pyWrite(primaryOut.ds, \"out\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, ins.map(i => i.ds), [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (union)\", [rn, primaryOut.ds], [`${ins.length} inputs, mode ${mode}`], known ? [] : [\"Column set unknown — Python concat used instead of a visual Stack recipe.\"]);\n }\n continue;\n }\n\n // ---------- SUMMARIZE ----------\n if (/\\.Summarize\\.Summarize$/i.test(p)) {\n if (!firstIn) { passthrough(\"Summarize had no wired input.\"); continue; }\n const sfs = arr(cfg?.SummarizeFields?.SummarizeField).map((f: any) => ({\n field: String(f?.[\"@field\"] ?? \"\"), action: String(f?.[\"@action\"] ?? \"\"), rename: String(f?.[\"@rename\"] ?? \"\") || null,\n separator: f?.[\"@separator\"] != null ? String(f[\"@separator\"]) : undefined,\n }));\n const ACTION_MAP: Record = { sum: \"sum\", min: \"min\", max: \"max\", avg: \"avg\", count: \"count\", countdistinct: \"countDistinct\", first: \"first\", last: \"last\", concat: \"concat\", stddev: \"stddev\", countnonnull: \"count\" };\n const unsupported = sfs.filter(f => !/^groupby$/i.test(f.action) && !ACTION_MAP[f.action.toLowerCase()]);\n const typeOf = (c: string) => firstIn.schema?.find(x => x.name === c)?.type || \"string\";\n if (!unsupported.length) {\n const keys = sfs.filter(f => /^groupby$/i.test(f.action)).map(f => groupingKey(f.field, typeOf(f.field)));\n const values = sfs.filter(f => !/^groupby$/i.test(f.action)).map((f, i) => groupingValue(f.field, typeOf(f.field), ACTION_MAP[f.action.toLowerCase()], i, f.separator));\n const grpDs = uniqueName(`t${node.id}_grouped`);\n addManaged(grpDs, null, tags);\n const rn1 = addRecipe(\"grouping\", `compute_${grpDs}`, [firstIn.ds], [grpDs], tags, \"grouping\", JSON.stringify(groupingPayload(keys, values), null, 1));\n // rename DSS default names (_) to Alteryx names\n const renames: { from: string; to: string }[] = [];\n for (const f of sfs) {\n if (/^groupby$/i.test(f.action)) { if (f.rename && f.rename !== f.field) renames.push({ from: f.field, to: f.rename }); continue; }\n const flag = ACTION_MAP[f.action.toLowerCase()];\n const dssName = `${f.field}_${flag}`;\n const axName = f.rename || `${f.action}_${f.field}`;\n if (dssName !== axName) renames.push({ from: dssName, to: axName });\n }\n const warns: string[] = [];\n const risky = sfs.some(f => [\"countdistinct\", \"stddev\", \"concat\", \"first\", \"last\", \"countnonnull\"].includes(f.action.toLowerCase()));\n if (risky) warns.push(\"Verify DSS auto output names for non-basic aggregations (assumed _) — adjust the rename Prepare step if they differ.\");\n if (sfs.some(f => /^countnonnull$/i.test(f.action))) warns.push(\"CountNonNull approximated with count — verify null handling.\");\n if (renames.length) {\n addManaged(primaryOut.ds, primaryOut.schema, tags);\n const rn2 = addRecipe(\"shaker\", `compute_${primaryOut.ds}`, [grpDs], [primaryOut.ds], tags, \"shaker\",\n JSON.stringify(shakerPayload([step(\"ColumnRenamer\", { renamings: renames }, \"Restore Alteryx column names\")]), null, 1));\n rrow(node, \"Group recipe + rename\", [rn1, grpDs, rn2, primaryOut.ds], renames.map(r => `${r.from} -> ${r.to}`), warns);\n } else {\n // grouped output IS the final output: re-point registration\n // no renames needed: identity Prepare keeps pass-1 dataset registrations valid\n addManaged(primaryOut.ds, primaryOut.schema, tags);\n const rn2 = addRecipe(\"shaker\", `compute_${primaryOut.ds}`, [grpDs], [primaryOut.ds], tags, \"shaker\",\n JSON.stringify(shakerPayload([]), null, 1));\n rrow(node, \"Group recipe\", [rn1, grpDs, rn2, primaryOut.ds], [], warns);\n }\n } else {\n const aggLines: string[] = [];\n const gb = sfs.filter(f => /^groupby$/i.test(f.action)).map(f => f.field);\n aggLines.push(`gb = ${JSON.stringify(gb)}`);\n aggLines.push(`g = df.groupby(gb, dropna=False, sort=False) if gb else df.assign(__k=0).groupby(\"__k\")`);\n aggLines.push(`parts = []`);\n for (const f of sfs.filter(f => !/^groupby$/i.test(f.action))) {\n const outName = f.rename || `${f.action}_${f.field}`;\n const c = JSON.stringify(f.field); const o = JSON.stringify(outName);\n const a = f.action.toLowerCase();\n const expr =\n a === \"sum\" ? `g[${c}].sum()` : a === \"min\" ? `g[${c}].min()` : a === \"max\" ? `g[${c}].max()` :\n a === \"avg\" ? `g[${c}].mean()` : a === \"count\" ? `g.size()` : a === \"countnonnull\" ? `g[${c}].count()` :\n a === \"countdistinct\" ? `g[${c}].nunique()` : a === \"first\" ? `g[${c}].first()` : a === \"last\" ? `g[${c}].last()` :\n a === \"median\" ? `g[${c}].median()` : a === \"stddev\" ? `g[${c}].std()` :\n a === \"mode\" ? `g[${c}].agg(lambda s: s.mode().iloc[0] if len(s.mode()) else None)` :\n a === \"concat\" ? `g[${c}].agg(lambda s: ${JSON.stringify(f.separator ?? \",\")}.join(s.astype(str)))` :\n `g[${c}].agg(lambda s: None) # TODO: unsupported action '${f.action}'`;\n aggLines.push(`parts.append((${expr}).rename(${o}))`);\n }\n aggLines.push(`out = pd.concat(parts, axis=1).reset_index()`);\n aggLines.push(`if \"__k\" in out.columns: out = out.drop(columns=[\"__k\"])`);\n const code = [...pyHeader(node, `Summarize with actions not supported by the visual Group recipe (${unsupported.map(u => u.action).join(\", \")}).`),\n pyRead(firstIn.ds, \"df\"), ...aggLines, ``, pyWrite(primaryOut.ds, \"out\"), ``].join(\"\\n\");\n addManaged(primaryOut.ds, primaryOut.schema, tags);\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n const PANDAS_OK = /^(sum|min|max|avg|count|countnonnull|countdistinct|first|last|median|stddev|mode|concat)$/i;\n const pyWarns = unsupported.filter(u => !PANDAS_OK.test(u.action)).map(u => `Summarize action '${u.action}' is not auto-translated — its output column is None; implement it in the Python recipe.`);\n rrow(node, \"python recipe (summarize)\", [rn, primaryOut.ds], [], pyWarns);\n }\n continue;\n }\n\n // ---------- UNIQUE ----------\n if (/\\.Unique\\.Unique$/i.test(p)) {\n if (!firstIn) { passthrough(\"Unique had no wired input.\"); continue; }\n const ucols = arr(cfg?.UniqueFields?.Field).map((f: any) => String(f?.[\"@field\"] ?? \"\"));\n const wantU = isWired(node.id, \"Unique\") || !isWired(node.id, \"Dupes\");\n const wantD = isWired(node.id, \"Dupes\");\n const outs: string[] = []; const dsOuts: string[] = [];\n const lines = [...pyHeader(node, `Unique on ${ucols.join(\", \") || \"all columns\"} — first occurrence is Unique, the rest are Dupes.`),\n pyRead(firstIn.ds, \"df\"),\n `cols = ${JSON.stringify(ucols)} or list(df.columns)`,\n `dup = df.duplicated(subset=cols, keep=\"first\")`, ``];\n if (wantU) { const o = lookupOut(node.id, \"Unique\")!; addManaged(o.ds, o.schema || firstIn.schema, tags); lines.push(pyWrite(o.ds, \"df[~dup]\")); dsOuts.push(o.ds); }\n if (wantD) { const o = lookupOut(node.id, \"Dupes\")!; addManaged(o.ds, o.schema || firstIn.schema, tags); lines.push(pyWrite(o.ds, \"df[dup]\")); dsOuts.push(o.ds); }\n const rn = addRecipe(\"python\", `compute_${dsOuts[0]}`, [firstIn.ds], dsOuts, tags, \"py\", lines.join(\"\\n\") + \"\\n\", PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (unique)\", [rn, ...dsOuts], [`keys: ${ucols.join(\", \") || \"all columns\"}`]);\n continue;\n }\n\n // ---------- RECORD ID ----------\n if (/\\.RecordID\\.RecordID$/i.test(p)) {\n if (!firstIn) { passthrough(\"RecordID had no wired input.\"); continue; }\n const fieldName = String(cfg?.FieldName ?? \"RecordID\") || \"RecordID\";\n const startN = Number.isFinite(parseInt(String(cfg?.StartValue ?? \"1\"))) ? parseInt(String(cfg?.StartValue ?? \"1\")) : 1;\n const start = String(startN);\n addManaged(primaryOut.ds, primaryOut.schema || firstIn.schema, tags);\n const code = [...pyHeader(node, `Record ID starting at ${start}.`), pyRead(firstIn.ds, \"df\"),\n `df.insert(0, ${JSON.stringify(fieldName)}, range(${startN}, ${startN} + len(df)))`,\n ``, pyWrite(primaryOut.ds, \"df\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (record id)\", [rn, primaryOut.ds]);\n continue;\n }\n\n // ---------- SAMPLE ----------\n if (/\\.Sample\\.Sample$/i.test(p)) {\n if (!firstIn) { passthrough(\"Sample had no wired input.\"); continue; }\n const mode = String(cfg?.Mode ?? \"First\");\n const nParsed = parseInt(String(cfg?.N ?? \"100\"));\n const n = Number.isFinite(nParsed) ? nParsed : 100;\n const gb = arr(cfg?.GroupFields?.Field).map((f: any) => String(f?.[\"@field\"] ?? \"\"));\n addManaged(primaryOut.ds, primaryOut.schema || firstIn.schema, tags);\n let body: string[]; const warns: string[] = [];\n const grouped = gb.length ? `df.groupby(${JSON.stringify(gb)}, sort=False, group_keys=False)` : null;\n if (/^first$/i.test(mode)) body = [grouped ? `out = ${grouped}.head(${n})` : `out = df.head(${n})`];\n else if (/^last$/i.test(mode)) body = [grouped ? `out = ${grouped}.tail(${n})` : `out = df.tail(${n})`];\n else { body = [`out = df.head(${n}) # TODO: Alteryx sample mode '${mode}' not auto-translated`]; warns.push(`Sample mode '${mode}' not auto-translated — placeholder takes first ${n} rows.`); }\n const code = [...pyHeader(node, `Sample (${mode} N=${n}${gb.length ? \", grouped\" : \"\"}).`), pyRead(firstIn.ds, \"df\"), ...body, ``, pyWrite(primaryOut.ds, \"out\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (sample)\", [rn, primaryOut.ds], [`${mode} ${n}`], warns);\n continue;\n }\n\n // ---------- TRANSPOSE ----------\n if (/\\.Transpose\\.Transpose$/i.test(p)) {\n if (!firstIn) { passthrough(\"Transpose had no wired input.\"); continue; }\n const keyCols = arr(cfg?.KeyFields?.Field).map((f: any) => String(f?.[\"@field\"] ?? \"\"));\n const dataCols = arr(cfg?.DataFields?.Field).filter((f: any) => String(f?.[\"@selected\"]).toLowerCase() !== \"false\" && f?.[\"@field\"] !== \"*Unknown\").map((f: any) => String(f?.[\"@field\"] ?? \"\"));\n addManaged(primaryOut.ds, primaryOut.schema, tags);\n const code = [...pyHeader(node, \"Transpose -> pandas melt (Name/Value columns, exact Alteryx semantics).\"), pyRead(firstIn.ds, \"df\"),\n `id_vars = ${JSON.stringify(keyCols)}`,\n `value_vars = ${JSON.stringify(dataCols)} or [c for c in df.columns if c not in id_vars]`,\n `out = df.melt(id_vars=id_vars, value_vars=value_vars, var_name=\"Name\", value_name=\"Value\")`,\n ``, pyWrite(primaryOut.ds, \"out\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (transpose)\", [rn, primaryOut.ds]);\n continue;\n }\n\n // ---------- CROSSTAB ----------\n if (/\\.CrossTab\\.CrossTab$/i.test(p)) {\n if (!firstIn) { passthrough(\"CrossTab had no wired input.\"); continue; }\n const gb = arr(cfg?.GroupFields?.Field).map((f: any) => String(f?.[\"@field\"] ?? \"\"));\n const header = String(cfg?.HeaderField?.[\"@field\"] ?? cfg?.HeaderField ?? \"\");\n const dataF = String(cfg?.DataField?.[\"@field\"] ?? cfg?.DataField ?? \"\");\n const methods = arr(cfg?.Methods?.Method).map((m: any) => String(m?.[\"@method\"] ?? m ?? \"Sum\"));\n const aggf = /first/i.test(methods[0] || \"\") ? \"first\" : /count/i.test(methods[0] || \"\") ? \"count\" : /avg|mean/i.test(methods[0] || \"\") ? \"mean\" : /concat/i.test(methods[0] || \"\") ? \"CONCAT\" : \"sum\";\n addManaged(primaryOut.ds, primaryOut.schema, tags);\n const code = [...pyHeader(node, `CrossTab: headers from '${header}', values from '${dataF}' (${methods.join(\",\") || \"Sum\"}).`),\n `import re`, pyRead(firstIn.ds, \"df\"),\n aggf === \"CONCAT\"\n ? `out = df.pivot_table(index=${JSON.stringify(gb)}, columns=${JSON.stringify(header)}, values=${JSON.stringify(dataF)}, aggfunc=lambda s: \",\".join(s.astype(str))).reset_index()`\n : `out = df.pivot_table(index=${JSON.stringify(gb)}, columns=${JSON.stringify(header)}, values=${JSON.stringify(dataF)}, aggfunc=${JSON.stringify(aggf)}).reset_index()`,\n `out.columns = [re.sub(r\"[^0-9a-zA-Z_]\", \"_\", str(c)) for c in out.columns] # Alteryx header sanitization`,\n ``, pyWrite(primaryOut.ds, \"out\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (crosstab)\", [rn, primaryOut.ds], [], [\"Verify CrossTab column naming against the original workflow output.\"]);\n continue;\n }\n\n // ---------- APPEND FIELDS ----------\n if (/\\.AppendFields\\.AppendFields$/i.test(p)) {\n const tIn = inputFor(node, \"Targets\") || firstIn; const sIn = inputFor(node, \"Sources\") || inbound(node.id).map(c => lookupOut(c.from, c.fromAnchor))[1] as OutReg | null;\n if (!tIn || !sIn) { passthrough(\"AppendFields missing inputs.\"); continue; }\n addManaged(primaryOut.ds, primaryOut.schema, tags);\n const code = [...pyHeader(node, \"Append Fields = cartesian product (cross join).\"),\n pyRead(tIn.ds, \"targets\"), pyRead(sIn.ds, \"sources\"),\n `out = targets.merge(sources, how=\"cross\")`,\n ``, pyWrite(primaryOut.ds, \"out\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [tIn.ds, sIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (append fields)\", [rn, primaryOut.ds]);\n continue;\n }\n\n // ---------- OUTPUT ----------\n if (/\\.DbFileOutput\\./i.test(p) || /LockInGui\\.LockInStreamOut/i.test(p)) {\n if (!firstIn) { rrow(node, \"omitted\", [], [], [\"Output tool had no wired input.\"]); continue; }\n const f = typeof cfg?.File === \"string\" ? cfg.File : cfg?.File?.[\"#text\"] || \"\";\n addManaged(primaryOut.ds, primaryOut.schema || firstIn.schema, tags);\n const code = [...pyHeader(node, `Final output (was: ${f || \"unknown\"}). Export from this dataset (Export button) or add an export recipe.`),\n pyRead(firstIn.ds, \"df\"), pyWrite(primaryOut.ds, \"df\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"output dataset\", [rn, primaryOut.ds], [`Original target: ${f || \"unknown\"}`]);\n continue;\n }\n\n // ---------- BLOCK UNTIL DONE (identity copy so downstream wiring survives) ----------\n if (/BlockUntilDone/i.test(p)) {\n if (!firstIn) { rrow(node, \"omitted\", [], [], [\"Block Until Done had no wired input.\"]); continue; }\n addManaged(primaryOut.ds, primaryOut.schema || firstIn.schema, tags);\n const code = [...pyHeader(node, \"Block Until Done — plain copy; DSS's flow build order makes this tool unnecessary. Safe to remove and rewire.\"),\n pyRead(firstIn.ds, \"df\"), pyWrite(primaryOut.ds, \"df\"), ``].join(\"\\n\");\n const rn = addRecipe(\"python\", `compute_${primaryOut.ds}`, [firstIn.ds], [primaryOut.ds], tags, \"py\", code, PY_RECIPE_PARAMS);\n rrow(node, \"python recipe (identity)\", [rn, primaryOut.ds], [\"DSS flow build order replaces Block Until Done — recipe is a plain copy.\"]);\n continue;\n }\n\n // ---------- LockIn (In-DB) intermediates ----------\n if (/^LockInGui\\./i.test(p)) { passthrough(\"Was an Alteryx In-DB tool — consider a SQL recipe on your database connection instead.\", []); continue; }\n\n // ---------- everything else ----------\n passthrough(\"\");\n }\n\n // ---- safety sweep: every dataset referenced by a recipe must exist in the zip ----\n for (const rec of recipes) {\n for (const it of [...(rec.json.inputs?.main?.items ?? []), ...(rec.json.outputs?.main?.items ?? [])]) {\n if (!datasets[it.ref]) {\n const known = Object.values(outMap).find(o => o.ds === it.ref);\n datasets[it.ref] = managedDatasetJson(projectKey, it.ref, known?.schema ?? null, []);\n }\n }\n }\n\n // ---- project files ----\n const paramsJson = {\n dashboardAuthorizations: { allAuthorized: false, authorizations: [] },\n exposedObjects: { objects: [] },\n projectType: \"REGULAR\", owner: \"admin\", permissions: [],\n additionalDashboardUsers: { users: [] }, tutorialProject: false,\n settings: {\n flowAnchorSourcesAndSinks: true, gitCommitMode: \"AUTO\", disableAutomaticTriggers: false,\n integrations: { integrations: [] },\n flowBuildSettings: { mergeSparkPipelines: false, pruneBeforeSparkPipelines: true, mergeSqlPipelines: false, pruneBeforeSqlPipelines: true },\n limitsSettings: { memSampleBytes: { soft: -1, hard: -1 }, shakerMemTableBytes: { soft: -1, hard: -1 }, attachmentBytes: { soft: -1, hard: -1 } },\n relocationSettings: { sqlLikeRecipesInitializationMode: \"RESOLVED_TABLE_REFERENCES\" },\n recipeEnginesPreferences: { forbiddenEngines: [], enginesPreferenceOrder: [], forbiddenByRecipeType: {}, preferenceByRecipeType: {} },\n codeEnvs: { python: { useBuiltinEnv: true, preventOverride: false }, r: { useBuiltinEnv: true, preventOverride: false } },\n container: { containerMode: \"NONE\" },\n datasetsCreationSettings: { useGlobal: true, preferedStorageFormats: \"CSV_ESCAPING_NOGZIP_FORHIVE,CSV_EXCEL_GZIP,PARQUET_HIVE\", virtualizable: false, allowUploadsWithoutConnection: true },\n hiveEngineCreationSettings: { useGlobal: true, executionEngine: \"HIVECLI_LOCAL\" },\n sparkEngineCreationSettings: { useGlobal: true, executionEngine: \"SPARK_SUBMIT\", useGlobalMetastore: false },\n impalaEngineCreationSettings: { useGlobal: true, streamMode: true },\n wikiSettings: { promoted: false },\n cluster: { clusterMode: \"INHERIT\" }, k8sCluster: { clusterMode: \"INHERIT\" },\n },\n name: String(name),\n imgPattern: 3, showInitials: false,\n metrics: { probes: [], engineConfig: { padRunsWithMetrics: false, hive: { active: true, extraConf: [] }, basic: {}, dss: { active: true, selection: { useMemTable: false, filter: { distinct: false, enabled: false }, partitionSelectionMethod: \"ALL\", latestPartitionsN: 1, ordering: { enabled: false, rules: [] }, samplingMethod: \"FULL\", maxRecords: -1, targetRatio: 0.02, withinFirstN: -1, maxReadUncompressedBytes: -1 } }, sql: { active: true }, impala: { active: true }, spark: { active: true, extraConf: [] }, python: {} }, displayedState: { columns: [], metrics: [] } },\n metricsChecks: { runOnBuild: false, checks: [], displayedState: { checks: [] } },\n versionTag: VT, creationTag: VT,\n shortDesc: `Converted from Alteryx workflow '${name}' by yxmd2dataiku`,\n tags: [\"alteryx-migration\"], customFields: {},\n checklists: { checklists: [] },\n };\n\n const manifest = {\n exportedWithOptions: {\n exportUploads: true, exportManagedFS: false, exportAnalysisModels: false, exportSavedModels: false,\n exportManagedFolders: false, exportAllInputDatasets: false, exportAllDatasets: false,\n exportAllInputManagedFolders: false, exportGitRepository: false, exportInsightsData: false,\n useManualPluginsInfo: false, includedDatasetsData: [], includedSavedModels: [], includedManagedFolders: [], shareAllInputDatasets: false,\n },\n actualContent: { includedDatasets, includedSavedModels: [], includedManagedFolders: [], allAnalysisData: false, insightsData: false },\n originalProjectKey: projectKey,\n manifestVersion: 3,\n generatedWithDSSVersion: \"8.0.3\",\n generatedWithDSSConfVersion: \"8020\",\n exportedWithPlugins: [],\n requiredConnections: { filesystem_managed: { name: \"filesystem_managed\", type: \"Filesystem\" } },\n usedCodeEnvRefs: [], hasPartitionedDataset: false,\n };\n\n // ---- zip ----\n const files: Record = {};\n const put = (path: string, content: string | Uint8Array) => { files[path] = typeof content === \"string\" ? strToU8(content) : content; };\n put(\"export-manifest.json\", JSON.stringify(manifest, null, 1));\n put(\"project_config/params.json\", JSON.stringify(paramsJson, null, 1));\n for (const [dn, dj] of Object.entries(datasets)) put(`project_config/datasets/${dn}.json`, JSON.stringify(dj, null, 1));\n for (const r of recipes) {\n put(`project_config/recipes/${r.name}.json`, JSON.stringify(r.json, null, 1));\n put(`project_config/recipes/${r.name}.${r.payloadExt}`, r.payload);\n }\n for (const [pth, bytes] of Object.entries(uploadsFiles)) put(pth, bytes);\n const zip = zipSync(files, { level: 6 });\n\n const kinds = report.reduce((m: Record, r) => { m[r.kind] = (m[r.kind] || 0) + 1; return m; }, {});\n return {\n workflowName: String(name), projectKey,\n toolCount: nodes.length, connectionCount: conns.length,\n datasetCount: Object.keys(datasets).length, recipeCount: recipes.length,\n kinds, report, todos, zip,\n manualSteps: [\n `In Dataiku: + New project > Import project > choose the downloaded zip (project key suggestion: ${projectKey}).`,\n ...(Object.entries(datasets).some(([dn, d]: [string, any]) => d.type === \"UploadedFiles\" && !includedDatasets.some((i: any) => i.name === dn))\n ? [`Upload the original input file(s) into the UploadedFiles input dataset(s) (open dataset > Settings > add file).`] : []),\n `Open the Flow and run 'Build' on the final dataset(s); fix anything listed in the review TODOs below.`,\n ],\n };\n}\n\n// ========\n// HTTP UI\n// ========\nfunction b64(bytes: Uint8Array): string {\n let bin = \"\";\n const CH = 0x8000;\n for (let i = 0; i < bytes.length; i += CH) bin += String.fromCharCode(...bytes.subarray(i, i + CH));\n return btoa(bin);\n}\n\nconst HTML = `\nyxmd → Dataiku Converter\n\n\n

Alteryx .yxmd → Dataiku project converter

Upload or paste a .yxmd. Output: a Dataiku project export ZIP you import via + New project → Import project.\nEach Alteryx tool becomes a Dataiku recipe (visual where safe, exact-logic Python otherwise); Text Input data travels inside the zip.\nTypically the only manual step left is uploading the original input file(s).

\n or paste XML:\n
\n\n

\n\n

View source

\n`;\n\nexport default async function (req: Request): Promise {\n const url = new URL(req.url);\n if (url.pathname === \"/source\") {\n return new Response(null, { status: 302, headers: { Location: \"https://www.val.town/x/niteshgarg/yxmd2dataiku\" } });\n }\n if (url.pathname === \"/convert\" && req.method === \"POST\") {\n try {\n const xml = await req.text();\n const r = convert(xml);\n const { zip, ...rest } = r;\n return Response.json({ ...rest, zipBase64: b64(zip) });\n } catch (e) {\n return Response.json({ error: String((e as any)?.message || e) }, { status: 400 });\n }\n }\n if (url.pathname === \"/convert.zip\" && req.method === \"POST\") {\n try {\n const xml = await req.text();\n const r = convert(xml);\n return new Response(r.zip, { headers: { \"content-type\": \"application/zip\", \"content-disposition\": `attachment; filename=\"${r.projectKey}_dataiku_project.zip\"` } });\n } catch (e) {\n return Response.json({ error: String((e as any)?.message || e) }, { status: 400 });\n }\n }\n return new Response(HTML, { headers: { \"content-type\": \"text/html; charset=utf-8\" } });\n}","deleted","label","children",[],"path",[41],{"_39":40,"_13":41,"_42":43,"_44":-5,"_45":46,"_47":48,"_49":34,"_50":64,"_16":52,"_53":54,"_55":56,"_57":36,"_58":14,"_61":65},["D",1783966855505],[41],"pulls",0,"branches",1,"forks","envVars",{"_75":67,"_77":67},"connections",{"_75":67,"_76":67},"count","unfilledCount","emptyCount","92ecf4b1-1e34-438d-9d71-5b080ff03c78",["D",1783953485606],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","Converts Alteryx .yxmd workflows to Dataiku build scripts","imageUrl","userId","9fb27c8d-12ff-4687-8d23-4c57f7106bcc","pinnedFileId","pinPreview","pinMigrated","screenshotVersion","likeCount","orgId","019f568a-526e-749f-bfce-9d7116ddbe2c","tursoName","tursoHostname","tursoMeasuredAt","pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_11":12,"_133":129},"org",{"_11":12,"_126":127,"_128":129,"_88":89,"_130":131,"_132":36},"branch",{"_16":48,"_13":117,"_39":46,"_118":78,"_119":-5,"_120":-5,"_121":-5,"_122":-5,"_123":-5,"_124":14,"_125":12},"isOwned","main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","displayName","Nitesh Garg","avatar","https://lh3.googleusercontent.com/a/ACg8ocI3s1ahWykesQS62y9kulHCxNb4oEZxdLUspwqUjPgDKzd0QA=s96-c","tier","free","canRestrictVals","profileImageUrl","","adoptOrgId",{"_39":78,"_13":14,"_50":160,"_80":-5,"_81":36,"_82":83,"_84":83,"_85":86,"_87":-5,"_88":89,"_90":-5,"_91":36,"_92":-5,"_93":-5,"_94":67,"_95":96,"_97":-5,"_98":-5,"_99":-5,"_100":-5,"_101":102,"_103":102,"_104":105,"_106":105,"_107":105,"_108":102,"_109":102,"_110":161,"_112":162,"_114":163,"_116":36},"tree",[157],[157],{"_66":67,"_68":69,"_70":67,"_71":155,"_73":156},[153],"file","upstreamProject","upstreamBranch","emptyEnvVars","flags",[],"hasHttpFiles","grants","orgs","customDomains",{},{"_39":40,"_13":41,"_42":43,"_44":-5,"_45":46,"_47":48,"_49":34,"_50":64,"_16":52,"_53":54,"_55":56,"_57":36,"_58":14,"_61":154},[41],{"_75":67,"_77":67},{"_75":67,"_76":67},{"_39":40,"_13":41,"_42":43,"_44":-5,"_45":46,"_47":48,"_49":34,"_50":51,"_16":52,"_53":54,"_55":56,"_57":36,"_58":14,"_59":158,"_61":159},[],[41],["D",1783953485606],{"_11":12,"_133":129},{"_11":12,"_126":127,"_128":129,"_88":89,"_130":131,"_132":36},{"_16":48,"_13":117,"_39":46,"_118":78,"_119":-5,"_120":-5,"_121":-5,"_122":-5,"_123":-5,"_124":14,"_125":12},"ctx",{"_110":-5,"_150":197,"_198":199,"_200":201,"_202":36,"_203":36,"_204":36,"_205":34,"_206":-5,"_207":-7,"_208":-7,"_209":67,"_210":36,"_211":36,"_212":-5,"_213":36,"_214":36,"_215":36,"_216":-5},"ENV",{"_168":169,"_170":171,"_172":117,"_173":174,"_175":176,"_177":178,"_179":180,"_181":182,"_183":184,"_185":186,"_187":188,"_189":190,"_191":192,"_193":194,"_195":196},"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"]