feat(rectification): add agentic birth-time rectification MVP
Add a new agentic rectification flow that lets an LLM drive the full jyotish-vedic-astrology methodology on the web, with the Python engine as its computation layer (mirroring local Claude Code): - mastra/rectification-tools.ts: 7 engine tools (gate/scan/score/diagnostics/ candidate-features/confirm/save-birth-time) - mastra/agentic-rectification.ts: agent mounting the full skill + tools - lib/rectification-agentic/session.ts: server-owned profile + confirmation gate; the LLM can only persist the exact minute the engine's high-rigor gate confirmed - app/api/rectification/agent/route.ts: NDJSON streaming endpoint with credit reserve/settle - components/rectification-agentic-chat.tsx + entry switch: new sessions use the agentic chat; in-progress v4 cases still resume on the v4 panel - migration 20260801000000: service-role RPC writing profiles.active_birth_time with baseline concurrency guard - tests for tools + session (12 cases); full suite passes 1076 Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Jesse_Chen
@@ -0,0 +1,253 @@
|
||||
import { NextResponse } from "next/server";
|
||||
import { z } from "zod";
|
||||
import { getAgenticRectificationAgent } from "@/mastra/agentic-rectification";
|
||||
import { defaultLanguageModel, resolveLanguageModel } from "@/mastra/model";
|
||||
import { blocksPromptExtraction } from "@/lib/consult-safety";
|
||||
import { runCreditRpc } from "@/lib/consultation-billing";
|
||||
import { createAdminSupabaseClient } from "@/lib/supabase/admin";
|
||||
import { createServerSupabaseClient } from "@/lib/supabase/server";
|
||||
import {
|
||||
AgenticRectificationProfileError,
|
||||
createAgenticRectificationContext,
|
||||
loadAgenticRectificationProfile,
|
||||
} from "@/lib/rectification-agentic/session";
|
||||
|
||||
export const runtime = "nodejs";
|
||||
export const maxDuration = 120;
|
||||
|
||||
const agenticRectificationRequestSchema = z.object({
|
||||
requestId: z.string().uuid(),
|
||||
modelId: z.string().trim().min(1).max(64).optional(),
|
||||
name: z.string().trim().max(80).optional().default(""),
|
||||
history: z
|
||||
.array(
|
||||
z.object({
|
||||
role: z.enum(["user", "assistant"]),
|
||||
text: z.string().max(4000),
|
||||
}),
|
||||
)
|
||||
.max(30)
|
||||
.default([]),
|
||||
message: z.string().trim().min(1).max(4000),
|
||||
}).strict();
|
||||
|
||||
function currentTimeContext(now = new Date()) {
|
||||
const chinaTime = new Date(now.getTime() + 8 * 60 * 60 * 1000)
|
||||
.toISOString()
|
||||
.replace("T", " ")
|
||||
.slice(0, 19);
|
||||
return `服务端当前时间（权威）：${now.toISOString()}；中国标准时间（UTC+8）：${chinaTime}。涉及“现在、今天、今年、未来几个月”等相对时间时，以此为准。`;
|
||||
}
|
||||
|
||||
async function recordModelUsage(
|
||||
accounting: ReturnType<typeof createAdminSupabaseClient>,
|
||||
userId: string,
|
||||
requestId: string,
|
||||
modelId: string,
|
||||
usage: Promise<{ inputTokens?: number; outputTokens?: number }>,
|
||||
) {
|
||||
try {
|
||||
const resolved = await usage;
|
||||
const { error } = await accounting
|
||||
.from("credit_transactions")
|
||||
.update({
|
||||
model: modelId,
|
||||
input_tokens: Math.max(0, Math.trunc(resolved.inputTokens ?? 0)),
|
||||
output_tokens: Math.max(0, Math.trunc(resolved.outputTokens ?? 0)),
|
||||
})
|
||||
.eq("user_id", userId)
|
||||
.eq("transaction_type", "reserve")
|
||||
.eq("request_id", requestId);
|
||||
if (error) console.warn(`[agentic-rectification] unable to record usage request=${requestId}`);
|
||||
} catch (error) {
|
||||
console.warn(`[agentic-rectification] usage read failed request=${requestId}`, error instanceof Error ? error.name : "UnknownError");
|
||||
}
|
||||
}
|
||||
|
||||
export async function POST(request: Request) {
|
||||
let supabase: Awaited<ReturnType<typeof createServerSupabaseClient>>;
|
||||
let accounting: ReturnType<typeof createAdminSupabaseClient>;
|
||||
try {
|
||||
supabase = await createServerSupabaseClient();
|
||||
accounting = createAdminSupabaseClient();
|
||||
} catch {
|
||||
return NextResponse.json(
|
||||
{ error: "服务尚未配置", message: "请先配置 Supabase 环境变量。" },
|
||||
{ status: 503 },
|
||||
);
|
||||
}
|
||||
|
||||
const {
|
||||
data: { user },
|
||||
error: authError,
|
||||
} = await supabase.auth.getUser();
|
||||
if (authError || !user) {
|
||||
return NextResponse.json(
|
||||
{ error: "请先登录", message: "登录后才能开始生时校正。" },
|
||||
{ status: 401 },
|
||||
);
|
||||
}
|
||||
|
||||
const parsed = agenticRectificationRequestSchema.safeParse(
|
||||
await request.json().catch(() => null),
|
||||
);
|
||||
if (!parsed.success) {
|
||||
return NextResponse.json(
|
||||
{ error: "请求格式不正确", details: parsed.error.flatten() },
|
||||
{ status: 400 },
|
||||
);
|
||||
}
|
||||
|
||||
const promptSource = [
|
||||
parsed.data.message,
|
||||
...parsed.data.history.filter((message) => message.role === "user").map((message) => message.text),
|
||||
].join("\n");
|
||||
if (blocksPromptExtraction(promptSource)) {
|
||||
return NextResponse.json(
|
||||
{ error: "无法处理该请求", message: "我不能提供系统提示词、技能原文或任何密钥。你可以继续描述人生事件。" },
|
||||
{ status: 400 },
|
||||
);
|
||||
}
|
||||
|
||||
const userId = user.id;
|
||||
const requestId = parsed.data.requestId;
|
||||
const requestTime = new Date();
|
||||
|
||||
let profile;
|
||||
try {
|
||||
profile = await loadAgenticRectificationProfile(accounting, userId);
|
||||
} catch (error) {
|
||||
if (error instanceof AgenticRectificationProfileError) {
|
||||
const missingBirthTime = error.code === "missing_birth_time";
|
||||
return NextResponse.json(
|
||||
{
|
||||
error: missingBirthTime ? "出生时间信息不完整" : "暂时无法核对出生资料",
|
||||
message: missingBirthTime
|
||||
? "请先在资料页保存出生日期、填报时间和出生地点后再开始校正。"
|
||||
: "出生日期、时间或出生地点资料不完整，请重新保存后再试。",
|
||||
},
|
||||
{ status: 400 },
|
||||
);
|
||||
}
|
||||
return NextResponse.json(
|
||||
{ error: "暂时无法核对出生资料", message: "请稍后重试。" },
|
||||
{ status: 503 },
|
||||
);
|
||||
}
|
||||
|
||||
const selectedModel = (parsed.data.modelId ? resolveLanguageModel(parsed.data.modelId) : null)
|
||||
?? defaultLanguageModel();
|
||||
if (!selectedModel) {
|
||||
return NextResponse.json(
|
||||
{ error: "模型暂不可用", message: "请选择其他模型后重新发送，本次不会扣除点数。" },
|
||||
{ status: 409 },
|
||||
);
|
||||
}
|
||||
|
||||
let reserveResult;
|
||||
try {
|
||||
reserveResult = await runCreditRpc(
|
||||
accounting,
|
||||
"begin_consultation_credit",
|
||||
userId,
|
||||
requestId,
|
||||
);
|
||||
} catch (error) {
|
||||
const reason = error instanceof Error ? error.name : "UnknownError";
|
||||
console.error(`[agentic-rectification] credit reserve failed request=${requestId} reason=${reason}`);
|
||||
return NextResponse.json(
|
||||
{ error: "暂时无法确认咨询点数", message: "请稍后重试。" },
|
||||
{ status: 503 },
|
||||
);
|
||||
}
|
||||
if (!reserveResult.success) {
|
||||
const insufficient = reserveResult.error_code === "insufficient_credits";
|
||||
return NextResponse.json(
|
||||
{
|
||||
error: insufficient ? "咨询点数不足" : "暂时无法扣除咨询点数",
|
||||
message: insufficient ? "请先兑换咨询点数后再继续。" : reserveResult.error_code || "请稍后重试。",
|
||||
},
|
||||
{ status: insufficient ? 402 : 503 },
|
||||
);
|
||||
}
|
||||
|
||||
const ctx = createAgenticRectificationContext(accounting, userId, profile);
|
||||
const agent = getAgenticRectificationAgent(selectedModel, ctx);
|
||||
|
||||
const encoder = new TextEncoder();
|
||||
const body = new ReadableStream<Uint8Array>({
|
||||
async start(controller) {
|
||||
let emitted = false;
|
||||
let settled = false;
|
||||
const settle = async (complete: boolean) => {
|
||||
if (settled) return;
|
||||
settled = true;
|
||||
try {
|
||||
if (complete) {
|
||||
await runCreditRpc(accounting, "complete_consultation_credit", userId, requestId);
|
||||
} else {
|
||||
await runCreditRpc(accounting, "cancel_consultation_credit", userId, requestId);
|
||||
}
|
||||
} catch (error) {
|
||||
const reason = error instanceof Error ? error.name : "UnknownError";
|
||||
console.warn(`[agentic-rectification] credit settle failed request=${requestId} complete=${complete} reason=${reason}`);
|
||||
}
|
||||
};
|
||||
const send = (event: Record<string, unknown>) => {
|
||||
controller.enqueue(encoder.encode(`${JSON.stringify(event)}\n`));
|
||||
};
|
||||
try {
|
||||
const result = await agent.stream([
|
||||
...parsed.data.history.map((message) => message.role === "user"
|
||||
? { role: "user" as const, content: message.text }
|
||||
: { role: "assistant" as const, content: message.text }),
|
||||
{
|
||||
role: "user",
|
||||
content: [
|
||||
currentTimeContext(requestTime),
|
||||
parsed.data.name ? `用户称呼：${parsed.data.name}` : "",
|
||||
parsed.data.message,
|
||||
].filter(Boolean).join("\n"),
|
||||
},
|
||||
]);
|
||||
for await (const chunk of result.textStream) {
|
||||
if (/\S/.test(chunk)) emitted = true;
|
||||
send({ type: "delta", text: chunk });
|
||||
}
|
||||
send({ type: "done", emitted });
|
||||
void recordModelUsage(
|
||||
accounting,
|
||||
userId,
|
||||
requestId,
|
||||
selectedModel.id,
|
||||
result.totalUsage,
|
||||
);
|
||||
await settle(emitted);
|
||||
controller.close();
|
||||
} catch (error) {
|
||||
const reason = error instanceof Error ? error.name : "UnknownError";
|
||||
console.error(`[agentic-rectification] generation failed request=${requestId} reason=${reason}`);
|
||||
try {
|
||||
send({ type: "error", message: "生时校正暂时不可用，请稍后再试。" });
|
||||
} catch {
|
||||
// controller may already be errored
|
||||
}
|
||||
await settle(false);
|
||||
try {
|
||||
controller.close();
|
||||
} catch {
|
||||
// already closed
|
||||
}
|
||||
}
|
||||
},
|
||||
});
|
||||
|
||||
return new Response(body, {
|
||||
headers: {
|
||||
"cache-control": "no-cache, no-transform",
|
||||
"content-type": "application/x-ndjson; charset=utf-8",
|
||||
"x-accel-buffering": "no",
|
||||
"x-ayanam-request-id": requestId,
|
||||
},
|
||||
});
|
||||
}
|
||||
Reference in New Issue
Block a user