Guided AI Process

Slogans:

`;\n if (r.slogans?.length) h += `

${esH(s)}

`; else h +=\n \"

No slogans generated.

\";\n if (r.rationale) h += `

Rationale:

${esH(r.rationale)}

`;\n return h;\n },\n },\n },\n\n // --- NEW ONLYFANS ASSISTANT PROCESSES ---\n\n // 4. OF - DM Reply Assistant\n {\n id: \"of_dm_reply\",\n name: \"OF - DM Reply Assistant\",\n description: \"Crafts persona-driven DM replies with upselling opportunities.\",\n color: \"border-pink-500\", // Example color\n questions: [\n {\n key: \"q_of_creator_persona_dm\",\n promptText: \"Creator's Persona\",\n displayName: \"Creator Persona\",\n type: \"text\",\n mapsTo: \"creatorPersona\",\n placeholder: \"e.g., 'Sarcastic Gamer Girl', 'Sweet & Innocent'\",\n optional: false,\n },\n {\n key: \"q_of_fan_message\",\n promptText: \"Fan's DM to You\",\n displayName: \"Fan's Message\",\n type: \"textarea\",\n mapsTo: \"fanMessage\",\n placeholder: \"Copy-paste the fan's message here\",\n optional: false,\n },\n {\n key: \"q_of_reply_goal_dm\",\n promptText: \"Main Goal for This Reply\",\n displayName: \"Reply Goal\",\n type: \"textarea\",\n mapsTo: \"replyGoal\",\n placeholder: \"e.g., Engage, upsell specific PPV, build rapport\",\n optional: false,\n },\n {\n key: \"q_of_ppv_details_dm\",\n promptText: \"PPV/Service to Upsell (Optional)\",\n displayName: \"Upsell Details\",\n type: \"textarea\",\n mapsTo: \"ppvDetails\",\n placeholder: \"e.g., 'New 'rage quit compilation' video, $10'\",\n optional: true,\n },\n {\n key: \"q_of_fan_history_dm\",\n promptText: \"Relevant Fan History/Context (Optional)\",\n displayName: \"Fan Context\",\n type: \"textarea\",\n mapsTo: \"fanHistory\",\n placeholder: \"e.g., 'New subscriber', 'Often buys PPVs', 'Asked about X before'\",\n optional: true,\n },\n ],\n inputSchemaRefine: (data, questions) => {\n const errors: InputValidationError[] = [];\n const requiredFields = [\"creatorPersona\", \"fanMessage\", \"replyGoal\"];\n for (const field of requiredFields) {\n if (!data[field] || String(data[field]).trim() === \"\") {\n const q = questions.find(q => q.mapsTo === field);\n errors.push({ fieldKey: q?.key || String(field), message: `${q?.displayName || field} is required.` });\n }\n }\n return errors.length ? errors : true;\n },\n agentConfig: {\n name: \"OF_DM_Reply_Agent\",\n promptGenerator: (p: any) => ({\n sys: `${OF_ASSISTANT_BASE_SYS_PROMPT_PREAMBLE}\nYour specific task is to draft a DM reply.\nCreator's Persona: {{creatorPersona}}\nCreator's Goal for this reply: {{replyGoal}}\nFan's original message to creator: \"{{fanMessage}}\"\nOptional - PPV/Service to potentially upsell: {{ppvDetails}} (If N/A or empty, do not invent one)\nOptional - Fan History/Context: {{fanHistory}} (If N/A or empty, treat as unknown)\n\nAdditionally, provide 1-2 alternative phrasings or follow-up message ideas.\n\nRespond with a single JSON object in the following format ONLY:\n{\n \"analysisRequest\": \"Your analysis of the creator's request based on their input to you.\",\n \"considerationsPersonaGoals\": \"How persona '{{creatorPersona}}' and goal '{{replyGoal}}' influenced the reply.\",\n \"reasoningForSuggestion\": \"The reasoning behind your suggested reply, including upsell strategy if applicable.\",\n \"suggestedDmReply\": \"The drafted DM reply for the creator to send. This reply should embody the persona.\",\n \"alternativeFollowUpIdeas\": [\"Alternative or follow-up idea 1.\", \"Alternative or follow-up idea 2.\"]\n}`,\n user: Utils.fillPromptTemplate(\n `\n Creator Persona: {{creatorPersona}}\n Fan's Message: {{fanMessage}}\n My Goal: {{replyGoal}}\n PPV Details (Optional): {{ppvDetails}}\n Fan History (Optional): {{fanHistory}}\n\n Please generate the DM reply and related analysis as per the system instructions.\n Ensure the 'suggestedDmReply' is ready for me to copy and paste.\n Ensure the entire response is a single JSON object.`,\n {\n creatorPersona: p.creatorPersona,\n fanMessage: p.fanMessage,\n replyGoal: p.replyGoal,\n ppvDetails: p.ppvDetails || \"N/A\",\n fanHistory: p.fanHistory || \"N/A\",\n },\n ),\n }),\n outputParser: (rawJson: string): OF_DMReplyOut => JSON.parse(rawJson),\n displayRenderer: (results: OF_DMReplyOut, esH) => `\n

Analysis & Rationale

Understanding Your Request: ${esH(results.analysisRequest)}

Considering Persona & Goals: ${esH(results.considerationsPersonaGoals)}

Reasoning for this Suggestion: ${esH(results.reasoningForSuggestion)}

Suggested DM Reply:

${esH(results.suggestedDmReply)}

\n ${\n results.alternativeFollowUpIdeas && results.alternativeFollowUpIdeas.length > 0\n ? `\n

Alternative/Follow-up Ideas:

${esH(idea)}

\n `\n : \"\"\n }\n `,\n },\n },\n\n // 5. OF - Content Idea Generator\n {\n id: \"of_content_ideas\",\n name: \"OF - Content Idea Generator\",\n description: \"Generates persona-driven content ideas (photo sets, videos, etc.).\",\n color: \"border-teal-500\", // Example color\n questions: [\n {\n key: \"q_of_creator_persona_ci\",\n promptText: \"Creator's Persona\",\n displayName: \"Creator Persona\",\n type: \"text\",\n mapsTo: \"creatorPersona\",\n placeholder: \"e.g., 'Mysterious Witchy Vibe'\",\n optional: false,\n },\n {\n key: \"q_of_content_type_ci\",\n promptText: \"Type of Content Ideas Needed\",\n displayName: \"Content Type\",\n type: \"text\",\n mapsTo: \"contentType\",\n placeholder: \"e.g., Photo sets, video concepts, live stream themes\",\n optional: false,\n },\n {\n key: \"q_of_content_goal_ci\",\n promptText: \"Primary Goal for This Content\",\n displayName: \"Content Goal\",\n type: \"text\",\n mapsTo: \"contentGoal\",\n placeholder: \"e.g., Main feed teasers, PPV, engagement boost\",\n optional: false,\n },\n {\n key: \"q_of_explicitness_ci\",\n promptText: \"Explicitness Level/Boundaries (Optional)\",\n displayName: \"Explicitness\",\n type: \"textarea\",\n mapsTo: \"explicitnessLevel\",\n placeholder: \"e.g., 'Alluring but not SFW for main feed', 'Tease spicier PPV'\",\n optional: true,\n },\n {\n key: \"q_of_num_ideas_ci\",\n promptText: \"Number of Distinct Ideas (Optional)\",\n displayName: \"Number of Ideas\",\n type: \"text\",\n mapsTo: \"numIdeas\",\n placeholder: \"Default: 3\",\n optional: true,\n },\n ],\n inputSchemaRefine: (data, questions) => {\n const errors: InputValidationError[] = [];\n const requiredFields = [\"creatorPersona\", \"contentType\", \"contentGoal\"];\n for (const field of requiredFields) {\n if (!data[field] || String(data[field]).trim() === \"\") {\n const q = questions.find(q => q.mapsTo === field);\n errors.push({ fieldKey: q?.key || String(field), message: `${q?.displayName || field} is required.` });\n }\n }\n return errors.length ? errors : true;\n },\n agentConfig: {\n name: \"OF_ContentIdea_Agent\",\n promptGenerator: (p: any) => ({\n sys: `${OF_ASSISTANT_BASE_SYS_PROMPT_PREAMBLE}\nYour specific task is to brainstorm content ideas.\nCreator's Persona: {{creatorPersona}}\nType of Content Needed: {{contentType}}\nPrimary Goal for this Content: {{contentGoal}}\nOptional - Explicitness Level/Boundaries: {{explicitnessLevel}} (If N/A, assume general persona-appropriateness)\nNumber of ideas to generate: {{numIdeas}} (If N/A or invalid, generate 3)\n\nFor each idea, provide: Title/Concept, Brief Description (visuals, actions), Persona Alignment (how it fits '{{creatorPersona}}'), and an Engagement/Upsell Hook (e.g., caption tease for PPV, interactive element, relevant to '{{contentGoal}}').\n\nRespond with a single JSON object in the following format ONLY:\n{\n \"analysisRequest\": \"Your analysis of the creator's request.\",\n \"considerationsPersonaGoals\": \"How persona '{{creatorPersona}}', content type '{{contentType}}', and goal '{{contentGoal}}' influenced the ideas.\",\n \"contentIdeas\": [\n {\n \"title\": \"Idea Title/Concept 1\",\n \"description\": \"Detailed description of the idea (visuals, actions, setting, etc.).\",\n \"personaAlignment\": \"Explanation of how this content idea specifically fits the '{{creatorPersona}}' persona.\",\n \"engagementUpsellHook\": \"Suggested caption tease, call to action, or engagement strategy related to '{{contentGoal}}' and the persona.\"\n }\n // ... more ideas if requested\n ]\n}`,\n user: Utils.fillPromptTemplate(\n `\n My Persona: {{creatorPersona}}\n Content Type: {{contentType}}\n Content Goal: {{contentGoal}}\n Explicitness (Optional): {{explicitnessLevel}}\n Number of Ideas (Optional, Default 3): {{numIdeas}}\n\n Please generate content ideas and analysis as per the system instructions.\n Ensure the entire response is a single JSON object.`,\n {\n creatorPersona: p.creatorPersona,\n contentType: p.contentType,\n contentGoal: p.contentGoal,\n explicitnessLevel: p.explicitnessLevel || \"N/A\",\n numIdeas: p.numIdeas || \"3\",\n },\n ),\n }),\n outputParser: (rawJson: string): OF_ContentIdeaOut => JSON.parse(rawJson),\n displayRenderer: (results: OF_ContentIdeaOut, esH) => `\n

Analysis & Rationale

Understanding Your Request: ${esH(results.analysisRequest)}

Considering Persona & Goals: ${esH(results.considerationsPersonaGoals)}

Analysis & Rationale

Understanding Your Request: ${esH(results.analysisRequest)}

Considering Persona & Goals: ${esH(results.considerationsPersonaGoals)}

Reasoning for this Suggestion: ${esH(results.reasoningForSuggestion)}

Suggested Welcome Message:

${esH(results.suggestedWelcomeMessage)}

\n ${\n results.alternativeMessages && results.alternativeMessages.length > 0\n ? `\n

Alternative Messages:

```
${esH(msg)}
```

\n `\n : \"\"\n }\n `,\n },\n },\n];\n\n// --- Helper to get Process Definition ---\nfunction getProcessDefinition(id: ProcessDefinitionId): ProcessDefinition | undefined {\n return ALL_PROCESS_DEFINITIONS.find(p => p.id === id);\n}\n\n// --- Zod Schema for User Input Validation ---\nconst UserInSchema = z.object({\n processId: z.custom(\n (val) => ALL_PROCESS_DEFINITIONS.some(p => p.id === val),\n \"Invalid process ID\",\n ),\n data: z.record(z.any()),\n}).refine((input) => {\n const pd = getProcessDefinition(input.processId);\n if (!pd) return false; // Should be caught by custom validation already\n // Check if all non-optional questions (mapped to data keys) are present and non-empty strings\n for (const q of pd.questions) {\n if (!q.optional) {\n const mappedKey = q.mapsTo;\n if (\n input.data[mappedKey] === undefined\n || (typeof input.data[mappedKey] === \"string\" && String(input.data[mappedKey]).trim() === \"\")\n ) {\n // console.log(`Validation fail: Missing required field ${mappedKey} for process ${input.processId}`);\n return false;\n }\n }\n }\n return true;\n}, { message: \"Server validation: Missing required data for one or more fields.\" });\ntype UserIn = z.infer;\n\n// --- Logging Agent ---\nclass LogAgent {\n logs: LogEntry[];\n midCtx?: string;\n constructor(mid?: string) {\n this.logs = [];\n this.midCtx = mid;\n }\n log(l: Lvl, c: string, m: string, d?: any, mid?: string, tid?: string) {\n const eM = mid || this.midCtx || \"N/A\";\n let sD;\n if (d != null) {\n try {\n sD = typeof structuredClone === \"function\"\n ? structuredClone(d)\n : JSON.parse(JSON.stringify(d, this.getCircReplacer()));\n }\n catch (e) {\n sD = `[LogErr:${(e as Error).message}]`;\n }\n } else sD = d;\n const en: LogEntry = { ts: new Date().toISOString(), mid: eM, tid, lvl: l, comp: c, msg: m, dt: sD };\n this.logs.push(en);\n // console.log(`[${en.ts}][${eM}][${en.tid||\"---\"}][${l}][${c}] ${m} ${sD != null ? `|D:${JSON.stringify(sD)}` : \"\"}`);\n }\n private getCircReplacer = () => {\n /* ... (as before) ... */\n const s = new WeakSet();\n return (_k: string, v: any) => {\n if (typeof v === \"object\" && v !== null) {\n if (s.has(v)) return \"[Circ]\";\n s.add(v);\n }\n return v;\n };\n };\n getLogs(): LogEntry[] {\n /* ... (as before) ... */\n try {\n return typeof structuredClone === \"function\"\n ? structuredClone(this.logs)\n : JSON.parse(JSON.stringify(this.logs, this.getCircReplacer()));\n }\n catch (e) {\n console.error(\"[LogFail]\", e);\n return [...this.logs];\n }\n }\n clearLogs() {\n this.logs = [];\n }\n setMidCtx(mid: string) {\n this.midCtx = mid;\n }\n}\n\n// --- OpenAI API Call ---\nasync function callOpenAI(sysP: string, userP: string, mid: string, tid: string, log: LogFn): Promise {\n log(\"DB\", \"OpenAI\", `Call tid=${tid}`, { sL: sysP.length, uL: userP.length }, mid, tid);\n try { // @ts-ignore\n const oai = new OpenAI(); // Assumes OPENAI_API_KEY is in env\n const comp = await oai.chat.completions.create({\n model: \"gpt-4o-mini\", // Using the specified model\n messages: [{ role: \"system\", content: sysP }, { role: \"user\", content: userP }], // Removed \"Respond JSON only.\" from user prompt as response_format handles it\n temperature: 0.7,\n response_format: { type: \"json_object\" }, // Ensures JSON output\n });\n const resT = comp.choices[0]?.message?.content;\n const usg = comp.usage;\n if (!resT) {\n log(\"WN\", \"OpenAI\", `No text tid=${tid}.`, { usg, fin: comp.choices[0]?.finish_reason }, mid, tid);\n return null;\n }\n log(\"IN\", \"OpenAI\", `OK tid=${tid}`, { rL: resT.length, usg, fin: comp.choices[0]?.finish_reason }, mid, tid);\n return resT.trim();\n } catch (err: any) {\n /* ... (error handling as before) ... */\n const eD = {\n m: err.message,\n n: err.name,\n s: err.stack?.slice(0, 300),\n t: err.type,\n c: err.code,\n p: err.param,\n st: err.status,\n };\n log(\"ER\", \"OpenAI\", `Fail tid=${tid}:${err.message}`, { e: eD }, mid, tid);\n throw new Error(\n `OpenAI Fail: ${err.message}` + (err.code ? `(C:${err.code},S:${err.status})` : `(S:${err.status})`),\n );\n }\n}\n\n// --- Questionnaire Agent (QAgent) ---\nclass QAgent {\n lg: LogFn;\n constructor(lgFn: LogFn) {\n this.lg = lgFn;\n }\n private getQKeys(qs: ProcessQuestion[], t: \"req\" | \"opt\" | \"all\"): string[] {\n /* ... (as before) ... */\n if (t === \"all\") return qs.map(q => q.key);\n return qs.filter(q => t === \"req\" ? !q.optional : q.optional).map(q => q.key);\n }\n private isQOpt(k: string, qs: ProcessQuestion[]): boolean {\n /* ... (as before) ... */\n return !!qs.find(q => q.key === k && q.optional);\n }\n private selNextQKey(st: QState): string | null {\n /* ... (as before) ... */\n const { currentQuestions: cQ, answers: ans } = st;\n const reqK = this.getQKeys(cQ, \"req\");\n const optK = this.getQKeys(cQ, \"opt\");\n for (const qK of reqK)\n if (!(qK in ans) || (ans[qK] === \"\" && !this.isQOpt(qK, cQ))) {\n this.lg(\"DB\", \"QA\", `Sel REQ:${qK}`, { aK: Object.keys(ans) }, st.sid);\n return qK;\n }\n for (const qK of optK)\n if (!(qK in ans)) {\n this.lg(\"DB\", \"QA\", `Sel OPT:${qK}`, { aK: Object.keys(ans) }, st.sid);\n return qK;\n }\n this.lg(\"DB\", \"QA\", `All answered ${st.processId}.`, { aK: Object.keys(ans) }, st.sid);\n return null;\n }\n private fmtQForUI(qK: string, qs: ProcessQuestion[]): QuestForUI {\n /* ... (as before) ... */\n const qD = qs.find(q => q.key === qK);\n if (!qD) throw new Error(`No QDef:${qK}`);\n let txt = qD.promptText;\n if (qD.optional && !txt.toLowerCase().includes(\"(optional)\")) txt += \" (Optional)\";\n return {\n id: `${qD.key}-${Utils.generateShortId()}`,\n fieldKey: qD.key,\n text: txt,\n type: qD.type,\n opts: qD.options,\n ph: qD.placeholder,\n };\n }\n async nextStep(\n sid: string | null,\n pid: ProcessDefinitionId,\n lastAns: { fieldKey: string; value: any } | null,\n curSt?: QState | null,\n ): Promise {\n /* ... (as before, with minor logging adjustments if needed) ... */\n const pD = getProcessDefinition(pid);\n if (!pD) return { sid: sid || `e-${Utils.generateShortId()}`, isDone: false, err: `No PID:\"${pid}\"`, state: null };\n const cPQs = pD.questions;\n const tSteps = cPQs.length;\n const sId = sid || `s-${Utils.generateShortId()}`;\n let st: QState;\n if (!curSt || !sid || curSt.sid !== sId || curSt.processId !== pid) {\n this.lg(\"IN\", \"QA\", `Init state:${pid}`, { nS: sId, oS: sid }, sId);\n st = {\n sid: sId,\n processId: pid,\n currentQuestions: cPQs,\n answers: {},\n qHist: [],\n isDone: false,\n totalSteps: tSteps,\n currentStep: 0,\n };\n if (lastAns?.fieldKey && cPQs.some(q => q.key === lastAns.fieldKey)) {\n const v = typeof lastAns.value === \"string\" ? lastAns.value.trim() : lastAns.value;\n if (v !== \"\" || this.isQOpt(lastAns.fieldKey, cPQs)) st.answers[lastAns.fieldKey] = v;\n }\n } else {\n st = {\n ...curSt,\n answers: { ...curSt.answers },\n qHist: [...curSt.qHist],\n totalSteps: tSteps,\n currentQuestions: cPQs,\n };\n if (lastAns?.fieldKey && cPQs.some(q => q.key === lastAns.fieldKey)) {\n const v = typeof lastAns.value === \"string\" ? lastAns.value.trim() : lastAns.value;\n if (v !== \"\" || this.isQOpt(lastAns.fieldKey, cPQs)) st.answers[lastAns.fieldKey] = v; else delete st\n .answers[lastAns.fieldKey];\n }\n }\n st.currentStep = st.qHist.length;\n const nxtQK = this.selNextQKey(st);\n if (!nxtQK) {\n this.lg(\"IN\", \"QA\", `All fields for ${st.processId} gathered. Validating.`, {\n ansCnt: Object.keys(st.answers).length,\n }, st.sid);\n const finMapData = this.finalizeData(st.answers, st.currentQuestions, st.processId, st.sid);\n const valRes = pD.inputSchemaRefine(finMapData.data, st.currentQuestions);\n if (valRes === true) {\n this.lg(\"IN\", \"QA\", `Valid OK ${st.processId}.`, {}, st.sid);\n st.isDone = true;\n st.currentStep = st.totalSteps;\n return { sid: st.sid, isDone: true, data: finMapData, state: st };\n } else {\n const firstE = valRes[0];\n this.lg(\"ER\", \"QA\", `Valid FAIL ${st.processId}. ErrOn:${firstE.fieldKey}`, { e: valRes }, st.sid);\n st.isDone = false;\n const fToReaskK = firstE.fieldKey;\n const qToReaskD = st.currentQuestions.find(q => q.key === fToReaskK);\n if (!qToReaskD)\n return { sid: st.sid, isDone: false, err: `InternalValidErr:No Q for key \"${fToReaskK}\".`, state: st };\n delete st.answers[fToReaskK]; // Changed from mapsTo to fieldKey\n Object.keys(st.answers).forEach(ansKey => { // ensure mapped key is also cleared if different\n const qDef = st.currentQuestions.find(q => q.key === ansKey);\n if (qDef && qDef.mapsTo === qToReaskD.mapsTo && ansKey !== fToReaskK) delete st.answers[ansKey];\n });\n const rQHistIdx = st.qHist.findIndex(q => q.fieldKey === fToReaskK);\n if (rQHistIdx !== -1) st.qHist.splice(rQHistIdx);\n st.currentStep = st.qHist.length;\n this.lg(\"DB\", \"QA\", `Re-trigger nextStep for ${fToReaskK} due to valid fail.`, { st }, st.sid);\n const reaskRsp = await this.nextStep(st.sid, st.processId, null, st); // Pass null as lastAnswer\n if (reaskRsp.nextQ && reaskRsp.nextQ.fieldKey === fToReaskK)\n reaskRsp.fieldError = `${\n qToReaskD.displayName || qToReaskD.promptText.split(\"?\")[0]\n }: ${firstE.message}`; else if (!reaskRsp.nextQ && !reaskRsp.isDone)\n reaskRsp.err = `Valid fail: ${firstE.message}`;\n return reaskRsp;\n }\n } else {\n st.currentStep = st.qHist.length + 1;\n const nxtQForUI = this.fmtQForUI(nxtQK, st.currentQuestions);\n st.qHist.push({ qid: nxtQForUI.id, fieldKey: nxtQForUI.fieldKey, text: nxtQForUI.text });\n this.lg(\"IN\", \"QA\", `Ask Q:${nxtQK}(S${st.currentStep}/${st.totalSteps})`, {\n qid: nxtQForUI.id,\n hL: st.qHist.length,\n }, st.sid);\n return { sid: st.sid, isDone: false, nextQ: nxtQForUI, state: st };\n }\n }\n async goBack(sid: string, pid: ProcessDefinitionId, curSt: QState): Promise {\n /* ... (as before) ... */\n this.lg(\"IN\", \"QA\", `Back. S${curSt.currentStep},H${curSt.qHist.length}`, { sid });\n let st = JSON.parse(JSON.stringify(curSt)); // Deep copy\n const pD = getProcessDefinition(pid);\n if (!pD) return { sid, isDone: false, err: \"PID not found for back\", state: st };\n st.currentQuestions = pD.questions; // Ensure current questions are from definition\n if (st.qHist.length === 0) {\n this.lg(\"WN\", \"QA\", \"Hist empty, cannot go back.\", { sid });\n return this.nextStep(st.sid, pid, null, {\n ...st,\n answers: {},\n qHist: [],\n currentStep: 0,\n totalSteps: st.currentQuestions.length,\n });\n }\n const curLastAsked = st.qHist.pop(); // Remove current/last asked from history\n if (curLastAsked) delete st.answers[curLastAsked.fieldKey]; // Clear its answer\n if (st.qHist.length === 0) {\n this.lg(\"IN\", \"QA\", \"Back to before 1st Q. Reset.\", { sid });\n const rState: QState = { ...st, answers: {}, qHist: [], currentStep: 0 };\n return this.nextStep(st.sid, pid, null, rState);\n }\n st.currentStep = st.qHist.length;\n this.lg(\"DB\", \"QA\", `Back. New S${st.currentStep}. Re-eval.`, { sid });\n // The question we are going back TO is now the last in qHist\n const qToReaskK = st.qHist[st.qHist.length - 1].fieldKey;\n const tmpAns = st.answers[qToReaskK]; // Store its answer temporarily\n delete st.answers[qToReaskK]; // Remove it to force re-asking\n st.qHist.pop(); // Pop it so nextStep re-adds it\n const rsp = await this.nextStep(st.sid, pid, null, st); // Get the question\n // If we are re-asking the same question, restore its previous answer in the new state for editing\n if (rsp.nextQ && rsp.nextQ.fieldKey === qToReaskK && rsp.state && tmpAns !== undefined)\n rsp.state.answers[qToReaskK] = tmpAns;\n return rsp;\n }\n async restart(sid: string, pid: ProcessDefinitionId): Promise {\n /* ... (as before) ... */\n this.lg(\"IN\", \"QA\", `Restart ${pid}`, { sid });\n const pD = getProcessDefinition(pid);\n if (!pD)\n return {\n sid: sid || `e-${Utils.generateShortId()}`,\n isDone: false,\n err: `No PID ${pid} for restart.`,\n state: null,\n };\n const iQState: QState = {\n sid: sid || `s-${Utils.generateShortId()}`,\n processId: pid,\n currentQuestions: pD.questions,\n answers: {},\n qHist: [],\n isDone: false,\n totalSteps: pD.questions.length,\n currentStep: 0,\n };\n return this.nextStep(iQState.sid, pid, null, iQState);\n }\n private finalizeData(\n ans: Record,\n qs: ProcessQuestion[],\n pid: ProcessDefinitionId,\n sid?: string,\n ): UserIn {\n /* ... (as before) ... */\n const pAns = { ...ans };\n qs.forEach(qD => {\n if (qD.optional && (pAns[qD.key] == null || (typeof pAns[qD.key] === \"string\" && pAns[qD.key].trim() === \"\")))\n delete pAns[qD.key];\n });\n const mapData: Record = {};\n for (const qK in pAns) {\n if (Object.prototype.hasOwnProperty.call(pAns, qK)) {\n const qD = qs.find(q => q.key === qK);\n if (qD) mapData[qD.mapsTo] = pAns[qK]; else this.lg(\"WN\", \"finalize\", `Ans QKey ${qK} no QDef. Skip.`, {}, sid);\n }\n }\n this.lg(\"DB\", \"finalize\", `Data mapped for ${pid}`, {\n fK: Object.keys(pAns).length,\n tK: Object.keys(mapData).length,\n mapK: Object.keys(mapData),\n }, sid);\n return { processId: pid, data: mapData };\n }\n}\n\n// --- Agent Runner ---\nasync function runAgent(\n t: TskIn,\n l: LogFn,\n aC: ProcessDefinition[\"agentConfig\"],\n): Promise> {\n /* ... (as before) ... */\n const { mid, tid, p } = t;\n l(\"IN\", aC.name, `Run tid=${tid}`, { iK: Object.keys(p) }, mid, tid);\n try {\n const { sys: sP, user: uP } = aC.promptGenerator(p);\n if (!sP || !uP) {\n l(\"ER\", aC.name, `PromptGen fail tid=${tid}.`, { sP, uP }, mid, tid);\n return { mid, cid: tid, p: {} as TOD, e: `${aC.name} fail:PromptGenErr.` };\n }\n const rOR = await callOpenAI(sP, uP, mid, tid, l);\n if (!rOR) {\n l(\"WN\", aC.name, `OpenAI no content tid=${tid}.`, {}, mid, tid);\n return { mid, cid: tid, p: {} as TOD, e: `${aC.name} fail:AI no content.` };\n }\n let oD: TOD;\n try {\n oD = aC.outputParser(rOR);\n }\n catch (e: any) {\n l(\"ER\", aC.name, `ParseFail tid=${tid}. M:${e.message}`, { rRPrev: rOR.slice(0, 300) }, mid, tid);\n return { mid, cid: tid, p: {} as TOD, e: `${aC.name} fail:AI output unparseable. ${e.message}` };\n }\n if (aC.postProcessor) {\n try {\n oD = aC.postProcessor(oD, p);\n l(\"DB\", aC.name, `PP applied tid=${tid}.`, {}, mid, tid);\n } catch (e: any) {\n l(\"ER\", aC.name, `PPFail tid=${tid}. E:${e.message}`, { cOD: oD }, mid, tid);\n }\n }\n l(\"IN\", aC.name, `Run OK tid=${tid}.`, { oK: Object.keys(oD) }, mid, tid);\n return { mid, cid: tid, p: oD };\n } catch (err: any) {\n l(\"ER\", aC.name, `Run CritFail tid=${tid}. E:${err.message}`, { s: err.stack?.slice(0, 300) }, mid, tid);\n return { mid, cid: tid, p: {} as TOD, e: `${aC.name} fail:${err.message}` };\n }\n}\n\n// --- Orchestrator ---\nclass Orchestrator {\n /* ... (as before) ... */\n baseL: LogFn;\n mId: string;\n lg: LogFn;\n constructor(bLgr: LogAgent) {\n this.mId = `m-${Utils.generateShortId()}`;\n bLgr.setMidCtx(this.mId);\n this.baseL = bLgr.log.bind(bLgr);\n this.lg = (l, c, m, d?, _mI?, t?) => {\n this.baseL(l, c, m, d, this.mId, t);\n };\n this.lg(\"IN\", \"Orch\", `New ${this.mId}`);\n }\n async runWorkflow(i: UserIn): Promise {\n this.lg(\"IN\", \"Orch\", `Start PID:${i.processId}`, { iK: Object.keys(i.data) }, this.mId);\n const pD = getProcessDefinition(i.processId);\n if (!pD) {\n const eM = `No PID:${i.processId}.`;\n this.lg(\"ER\", \"Orch\", eM, {}, this.mId);\n return { mId: this.mId, err: eM, dbgLog: [] };\n }\n const tId = `${i.processId.slice(0, 6)}-${Utils.generateShortId()}`;\n const aTI: TskIn = { mid: this.mId, tid: tId, p: i.data };\n try {\n const aTO = await runAgent(aTI, this.lg, pD.agentConfig);\n if (aTO.e) {\n this.lg(\"WN\", \"Orch\", `Agent PID ${i.processId}(tid:${tId}) err.`, { e: aTO.e }, this.mId);\n return { mid: this.mId, err: `AgentFail:${aTO.e}`, res: aTO.p, dbgLog: [] };\n }\n this.lg(\"IN\", \"Orch\", `WF PID ${i.processId}(tid:${tId}) OK.`, { oK: Object.keys(aTO.p) }, this.mId);\n return { mid: this.mId, res: aTO.p, dbgLog: [] };\n } catch (err: any) {\n const cEM = `WF CritFail PID ${i.processId}(tid:${tId}).E:${err.message}`;\n this.lg(\"ER\", \"Orch\", cEM, { s: err.stack?.slice(0, 300) }, this.mId);\n return { mid: this.mId, err: cEM, dbgLog: [] };\n }\n }\n}\n\n// --- UI Generation & Request Handling ---\nfunction generateGuidedAIProcessUI(pDefsUIStr: string): string {\n /* ... (UI HTML as before, ensure esH is used which will be Utils.escapeHTML on client) ... */\n // The client-side esH function needs to be robust.\n // The provided UI script uses its own esH. We will use Utils.escapeHTML for server-side rendering if that were added.\n // For the client-side rendering done by displayRenderer, it will pass Utils.escapeHTML which the client script needs to have access to or a similar function.\n // The existing client script already has an `esH` function. We'll ensure our displayRenderers are compatible.\n return `\nGuided AI Process\n

Step 0 of 0

\n`;\n}\n\nexport default async function(req: Request): Promise {\n const reqId = `r-${Utils.generateShortId()}`;\n const reqLgr = new LogAgent(reqId);\n const log = reqLgr.log.bind(reqLgr); // Use this for logging within the handler\n const url = new URL(req.url);\n const headers = {\n \"Content-Type\": \"application/json\",\n \"Access-Control-Allow-Origin\": \"*\",\n \"Access-Control-Allow-Methods\": \"GET,POST,OPTIONS\",\n \"Access-Control-Allow-Headers\": \"Content-Type,Authorization\",\n };\n\n if (req.method === \"OPTIONS\") {\n log(\"IN\", \"H\", \"OPTIONS req\", {}, reqId);\n return new Response(null, { status: 204, headers });\n }\n\n try {\n if (req.method === \"GET\" && (url.pathname === \"/\" || url.pathname === \"\")) {\n log(\"IN\", \"H\", \"GET / : UI Request\", { p: url.pathname }, reqId);\n const uiPDs = ALL_PROCESS_DEFINITIONS.map(p => ({\n id: p.id,\n name: p.name,\n description: p.description,\n color: p.color,\n questionCount: p.questions.length,\n }));\n const html = generateGuidedAIProcessUI(JSON.stringify(uiPDs));\n return new Response(html, { status: 200, headers: { ...headers, \"Content-Type\": \"text/html;charset=utf-8\" } });\n }\n\n if (req.method === \"POST\" && url.pathname === \"/questionnaire\") {\n log(\"IN\", \"H\", \"POST /questionnaire\", { p: url.pathname }, reqId);\n let reqB: any;\n let rawB = \"[NoBody]\";\n try {\n rawB = await req.text();\n reqB = JSON.parse(rawB);\n if (\n !reqB?.processId || typeof reqB.processId !== \"string\" || !reqB.action\n || ![\"start\", \"answer\", \"back\", \"restart\"].includes(reqB.action)\n || (reqB.action === \"answer\" && !reqB.lastAnswer)\n || (reqB.action === \"answer\"\n && (typeof reqB.lastAnswer.fieldKey !== \"string\" || reqB.lastAnswer.value === undefined))\n || ((reqB.action === \"back\" || reqB.action === \"answer\") && !reqB.state)\n ) throw new Error(\"Bad questionnaire request structure.\");\n } catch (e: any) {\n log(\"WN\", \"H\", \"Bad /questionnaire body\", { e: e.message, bPv: rawB.slice(0, 200) }, reqId);\n return new Response(\n JSON.stringify({ err: `Bad Request: ${e.message}`, state: null, debugLog: reqLgr.getLogs() }),\n { status: 400, headers },\n );\n }\n\n const qA = new QAgent(log);\n let qRsp: QResp;\n try {\n log(\"DB\", \"H\", `QAgent Action: ${reqB.action}`, {\n pid: reqB.processId,\n sId: reqB.sessionId,\n laF: reqB.lastAnswer?.fieldKey,\n }, reqId);\n switch (reqB.action) {\n case \"start\":\n qRsp = await qA.nextStep(reqB.sessionId, reqB.processId, reqB.lastAnswer, null);\n break;\n case \"answer\":\n qRsp = await qA.nextStep(reqB.sessionId, reqB.processId, reqB.lastAnswer, reqB.state);\n break;\n case \"back\":\n qRsp = await qA.goBack(reqB.sessionId, reqB.processId, reqB.state);\n break;\n case \"restart\":\n qRsp = await qA.restart(reqB.sessionId, reqB.processId);\n break;\n default:\n qRsp = {\n sid: reqB.sessionId || `e-${Utils.generateShortId()}`,\n isDone: false,\n err: \"Invalid action.\",\n state: reqB.state || null,\n };\n }\n log(\"DB\", \"H\", \"QAgent step processed.\", {\n sId: qRsp.sid,\n dn: qRsp.isDone,\n hE: !!qRsp.err,\n hFE: !!qRsp.fieldError,\n nQK: qRsp.nextQ?.fieldKey,\n }, reqId);\n return new Response(JSON.stringify(qRsp), { status: 200, headers });\n } catch (e: any) {\n log(\"ER\", \"H\", `QAgent system error during action: ${reqB.action}.`, {\n e: e.message,\n s: e.stack?.slice(0, 200),\n }, reqId);\n return new Response(\n JSON.stringify({\n err: `Internal QAgent system error: ${e.message}`,\n state: reqB.state || null,\n debugLog: reqLgr.getLogs(),\n mid: reqId,\n }),\n { status: 500, headers },\n );\n }\n }\n\n if (req.method === \"POST\" && url.pathname === \"/execute\") {\n log(\"IN\", \"H\", \"POST /execute\", { p: url.pathname }, reqId);\n let vIn: UserIn;\n let rawP = \"[NoPayload]\";\n try {\n const pL = await req.json();\n rawP = JSON.stringify(pL).slice(0, 300);\n log(\"DB\", \"H\", \"Received /execute payload\", { payloadPreview: rawP }, reqId);\n vIn = UserInSchema.parse(pL); // This uses the refined Zod schema\n log(\n \"IN\",\n \"H\",\n \"Payload validated for /execute.\",\n { pid: vIn.processId, dataKeys: Object.keys(vIn.data) },\n reqId,\n );\n } catch (e: any) {\n let eD = \"Bad input structure or missing required fields.\";\n if (e instanceof z.ZodError)\n eD = e.issues.map(i => `${i.path.join(\".\") || \"input\"}: ${i.message}`).join(\"; \"); else eD = e.message;\n log(\"WN\", \"H\", \"Validation failed for /execute.\", { errorDetails: eD, payloadPreview: rawP }, reqId);\n return new Response(\n JSON.stringify({ err: \"Invalid input.\", details: eD, debugLog: reqLgr.getLogs(), mid: `m-val-${reqId}` }),\n { status: 400, headers },\n );\n }\n\n const exLgr = new LogAgent(); // Fresh logger for this execution\n const orch = new Orchestrator(exLgr);\n const finR: FinalRes = await orch.runWorkflow(vIn);\n finR.dbgLog = exLgr.getLogs(); // Attach logs from this specific execution\n\n // Server-side rendering using displayRenderer if available (example)\n const pDef = getProcessDefinition(vIn.processId);\n if (pDef?.agentConfig.displayRenderer && finR.res && !finR.err) {\n try {\n // This 'details' field could carry pre-rendered HTML if desired.\n // For this implementation, the client-side fmtRDisp handles rendering based on JSON.\n // finR.details = pDef.agentConfig.displayRenderer(finR.res, Utils.escapeHTML);\n } catch (e: any) {\n log(\"ER\", \"H\", `Server-side displayRenderer failed for ${vIn.processId}.`, { e: e.message }, reqId);\n }\n }\n const rspS = finR.err ? (finR.err.includes(\"AI output unparseable\") ? 422 : 500) : 200;\n log(\"IN\", \"H\", `/execute workflow finished with status ${rspS}`, { mId: finR.mid, hasError: !!finR.err }, reqId);\n return new Response(JSON.stringify(finR), { status: rspS, headers });\n }\n\n log(\"WN\", \"H\", `Unsupported route: ${req.method} ${url.pathname}.`, {}, reqId);\n return new Response(JSON.stringify({ err: \"Not Found.\", debugLog: reqLgr.getLogs(), mid: reqId }), {\n status: 404,\n headers,\n });\n } catch (err: any) {\n log(\"ER\", \"H\", \"Critical unhandled error in main handler.\", { e: err.message, s: err.stack?.slice(0, 300) }, reqId);\n return new Response(\n JSON.stringify({\n err: \"Internal Server Error.\",\n details: \"An unexpected error occurred.\",\n debugLog: reqLgr.getLogs(),\n mid: `m-crit-${reqId}`,\n }),\n { status: 500, headers },\n );\n }\n}","deleted","label","children",[],[150],"files",[148],"counts",{"_173":32,"_174":175,"_176":32,"_177":32},"pulls","branches",1,"forks","envVars","contents",[180],{"_102":149,"_95":150,"_151":152,"_153":-5,"_154":136,"_155":135,"_156":157,"_104":181,"_99":159,"_160":161,"_162":163,"_164":22,"_165":-5,"_98":182},["D",1748166471628],[150],"file","upstreamProject","upstreamBranch","emptyEnvVars","isNew","flags",[],"routes/_app.x.$handle.$name._index",{"_93":94,"_95":96,"_97":-7,"_99":-7,"_100":192,"_178":197,"_200":-7,"_201":-5,"_202":-5,"_169":203,"_207":208},{"_102":103,"_95":96,"_104":193,"_106":-5,"_107":22,"_108":109,"_110":-5,"_111":-5,"_14":112,"_113":-5,"_114":22,"_115":-5,"_116":32,"_16":117,"_118":119,"_120":119,"_121":122,"_123":122,"_124":122,"_125":119,"_126":119,"_23":194,"_130":195,"_133":196,"_145":22},["D",1748166443293],{"_93":94,"_128":129},{"_93":94,"_132":129},{"_99":135,"_95":43,"_102":136,"_137":103,"_138":-5,"_139":-5,"_140":-5,"_141":-5,"_142":-5,"_143":96,"_144":94},[198],{"_102":149,"_95":150,"_151":152,"_153":-5,"_154":136,"_155":135,"_156":157,"_104":181,"_99":159,"_160":161,"_162":163,"_164":22,"_165":-5,"_98":199},[150],"readme","upstream","pull",[204],{"_102":149,"_95":150,"_151":152,"_153":-5,"_154":136,"_155":135,"_156":157,"_104":158,"_99":159,"_160":161,"_162":163,"_164":22,"_165":-5,"_166":205,"_98":206},[],[150],"previews",[204],"actionData","errors"]

Slogans:

Rationale:

Analysis & Rationale

Suggested DM Reply:

Alternative/Follow-up Ideas:

Analysis & Rationale

Suggested Content Ideas:

${esH(idea.title)}

Analysis & Rationale

Suggested Welcome Message:

Alternative Messages: