perf: 加快拆分QA和生成向量；余额不足提醒

fbbc3236 · archer · dc329041 · fbbc3236 · fbbc3236 · fbbc3236
Commit fbbc3236 authored Apr 05, 2023 by archer
8 changed files
--- a/src/pages/api/timer/initDataItemTime.ts
+++ b/src/pages/api/timer/initDataItemTime.ts
 // Next.js API route support: https://nextjs.org/docs/api-routes/introduction
 import type { NextApiRequest, NextApiResponse } from 'next';
 import { jsonRes } from '@/service/response';
-import { connectToDatabase, DataItem, Data } from '@/service/mongo';
+import { connectToDatabase, SplitData } from '@/service/mongo';
 export default async function handler(req: NextApiRequest, res: NextApiResponse) {
  try {
@@ -10,20 +10,18 @@ export default async function handler(req: NextApiRequest, res: NextApiResponse)
    }
    await connectToDatabase();
-    // await DataItem.updateMany(
+    const data = await SplitData.aggregate([
-    //   {},
+      { $match: { textList: { $exists: true, $ne: [] } } },
-    //   {
+      { $sample: { size: 1 } }
-    //     type: 'QA'
+    ]);
-    //     // times: 2
-    //   }
-    // );
-    await Data.updateMany(
+    const dataItem: any = data[0];
-      {},
+    const textList: string[] = dataItem.textList.slice(-5);
-      {
+    console.log(textList);
-        type: 'QA'
+    console.log(dataItem.textList.slice(0, -5));
-      }
+    await SplitData.findByIdAndUpdate(dataItem._id, {
-    );
+      textList: dataItem.textList.slice(0, -5)
+    });
    jsonRes(res, {
      data: {}

--- a/src/service/errorCode.ts
+++ b/src/service/errorCode.ts
@@ -6,6 +6,9 @@ export const openaiError: Record<string, string> = {
  'Too Many Requests': '请求次数太多了，请慢点~',
  'Bad Gateway': '网关异常，请重试'
 };
+export const openaiError2: Record<string, string> = {
+  insufficient_quota: 'API 余额不足'
+};
 export const proxyError: Record<string, boolean> = {
  ECONNABORTED: true,
  ECONNRESET: true

--- a/src/service/events/generateQA.ts
+++ b/src/service/events/generateQA.ts
@@ -9,39 +9,35 @@ import { generateVector } from './generateVector';
 import { connectRedis } from '../redis';
 import { VecModelDataPrefix } from '@/constants/redis';
 import { customAlphabet } from 'nanoid';
+import { ModelSplitDataSchema } from '@/types/mongoSchema';
 const nanoid = customAlphabet('abcdefghijklmnopqrstuvwxyz1234567890', 12);
-export async function generateQA(): Promise<any> {
+export async function generateQA(next = false): Promise<any> {
-  // 最多 5 个进程
+  if (global.generatingQA === true && !next) return;
-  if (global.generatingQA >= 5) {
+  global.generatingQA = true;
-    console.log('QA 最多5个进程');
-    return;
-  }
-  global.generatingQA++;
  let dataId = null;
  try {
    const redis = await connectRedis();
    // 找出一个需要生成的 dataItem
-    const dataItem = await SplitData.findOne({
+    const data = await SplitData.aggregate([
-      textList: { $exists: true, $ne: [] }
+      { $match: { textList: { $exists: true, $ne: [] } } },
-    });
+      { $sample: { size: 1 } }
+    ]);
+    const dataItem: ModelSplitDataSchema = data[0];
    if (!dataItem) {
      console.log('没有需要生成 QA 的数据');
-      global.generatingQA = 0;
+      global.generatingQA = false;
      return;
    }
    dataId = dataItem._id;
-    // 源文本
+    // 获取 5 个源文本
-    const text = dataItem.textList[dataItem.textList.length - 1];
+    const textList: string[] = dataItem.textList.slice(-5);
-    if (!text) {
-      await SplitData.findByIdAndUpdate(dataItem._id, { $pop: { textList: 1 } }); // 弹出无效文本
-      throw new Error('无文本');
-    }
    // 获取 openapi Key
    let userApiKey, systemKey;
@@ -62,7 +58,7 @@ export async function generateQA(): Promise<any> {
      throw new Error('获取 openai key 失败');
    }
-    console.log('正在生成一组QA, ID:', dataItem._id);
+    console.log(`正在生成一组QA, 包含 ${textList.length} 组文本。ID: ${dataItem._id}`);
    const startTime = Date.now();
@@ -76,33 +72,50 @@ export async function generateQA(): Promise<any> {
    };
    // 请求 chatgpt 获取回答
-    const response = await chatAPI
+    const response = await Promise.allSettled(
-      .createChatCompletion(
+      textList.map((text) =>
-        {
+        chatAPI
-          model: ChatModelNameEnum.GPT35,
+          .createChatCompletion(
-          temperature: 0.8,
-          n: 1,
-          messages: [
-            systemPrompt,
            {
-              role: 'user',
+              model: ChatModelNameEnum.GPT35,
-              content: text
+              temperature: 0.8,
+              n: 1,
+              messages: [
+                systemPrompt,
+                {
+                  role: 'user',
+                  content: text
+                }
+              ]
+            },
+            {
+              timeout: 180000,
+              httpsAgent
            }
-          ]
+          )
-        },
+          .then((res) => ({
-        {
+            rawContent: res?.data.choices[0].message?.content || '', // chatgpt原本的回复
-          timeout: 180000,
+            result: splitText(res?.data.choices[0].message?.content || '') // 格式化后的QA对
-          httpsAgent
+          }))
-        }
      )
-      .then((res) => ({
+    );
-        rawContent: res?.data.choices[0].message?.content || '', // chatgpt原本的回复
-        result: splitText(res?.data.choices[0].message?.content || '') // 格式化后的QA对
+    // 获取成功的回答
-      }));
+    const successResponse: {
+      rawContent: string;
+      result: {
+        q: string;
+        a: string;
+      }[];
+    }[] = response.filter((item) => item.status === 'fulfilled').map((item: any) => item.value);
+    const resultList = successResponse.map((item) => item.result).flat();
    await Promise.allSettled([
-      SplitData.findByIdAndUpdate(dataItem._id, { $pop: { textList: 1 } }), // 弹出已经拆分的文本
+      SplitData.findByIdAndUpdate(dataItem._id, {
-      ...response.result.map((item) => {
+        textList: dataItem.textList.slice(0, -5)
+      }), // 删掉后5个数据
+      ...resultList.map((item) => {
        // 插入 redis
        return redis.sendCommand([
          'HMSET',
@@ -125,20 +138,21 @@ export async function generateQA(): Promise<any> {
      '生成QA成功，time:',
      `${(Date.now() - startTime) / 1000}s`,
      'QA数量：',
-      response.result.length
+      resultList.length
    );
    // 计费
    pushSplitDataBill({
-      isPay: !userApiKey && response.result.length > 0,
+      isPay: !userApiKey && resultList.length > 0,
      userId: dataItem.userId,
      type: 'QA',
-      text: systemPrompt.content + text + response.rawContent
+      text:
+        systemPrompt.content +
+        textList.join('') +
+        successResponse.map((item) => item.rawContent).join('')
    });
-    global.generatingQA--;
+    generateQA(true);
-    generateQA();
    generateVector();
  } catch (error: any) {
    // log
@@ -157,14 +171,13 @@ export async function generateQA(): Promise<any> {
        errorText: 'api 余额不足'
      });
-      generateQA();
+      generateQA(true);
      return;
    }
    setTimeout(() => {
-      global.generatingQA--;
+      generateQA(true);
-      generateQA();
+    }, 4000);
-    }, 5000);
  }
 }

--- a/src/service/events/generateVector.ts
+++ b/src/service/events/generateVector.ts
@@ -75,9 +75,7 @@ export async function generateVector(next = false): Promise<any> {
    console.log(`生成向量成功: ${dataItem.id}`);
-    setTimeout(() => {
+    generateVector(true);
-      generateVector(true);
-    }, 4000);
  } catch (error: any) {
    // log
    if (error?.response) {
@@ -88,7 +86,7 @@ export async function generateVector(next = false): Promise<any> {
    }
    if (dataId && error?.response?.data?.error?.type === 'insufficient_quota') {
-      console.log('api 余额不足');
+      console.log('api 余额不足,删除 redis 模型数据');
      const redis = await connectRedis();
      redis.del(dataId);
      generateVector(true);

--- a/src/service/mongo.ts
+++ b/src/service/mongo.ts
@@ -27,7 +27,6 @@ export async function connectToDatabase(): Promise<void> {
    global.mongodb = null;
  }
-  global.generatingQA = 0;
  generateQA();
  // generateAbstract();
  generateVector(true);

--- a/src/service/preChatStore.ts
+++ b/src/service/preChatStore.ts
-import { ChatItemType } from '../types/chat';
-export const chatWindows = new Map<string, ChatItemType[]>();
-/**
- * 获取聊天窗口信息
- */
-export const getWindowMessages = (id: string) => {
-  return chatWindows.get(id) || [];
-};
-export const pushWindowMessage = (id: string, prompt: ChatItemType) => {
-  const messages = chatWindows.get(id) || [];
-  messages.push(prompt);
-  chatWindows.set(id, messages);
-  return messages;
-};
-export const deleteWindow = (id: string) => {
-  chatWindows.delete(id);
-};
--- a/src/service/response.ts
+++ b/src/service/response.ts
 import { NextApiResponse } from 'next';
-import { openaiError, proxyError } from './errorCode';
+import { openaiError, openaiError2, proxyError } from './errorCode';
 export interface ResponseType<T = any> {
  code: number;
@@ -25,13 +25,19 @@ export const jsonRes = <T = any>(
      msg = error;
    } else if (proxyError[error?.code]) {
      msg = '服务器代理出错';
+    } else if (openaiError2[error?.response?.data?.error?.type]) {
+      msg = openaiError2[error?.response?.data?.error?.type];
    } else if (openaiError[error?.response?.statusText]) {
      msg = openaiError[error.response.statusText];
    }
    console.log('error->');
    console.log('code:', error.code);
-    console.log('statusText:', error?.response?.statusText);
    console.log('msg:', msg);
+    // request 时候报错
+    if (error?.response) {
+      console.log('statusText:', error?.response?.statusText);
+      console.log('type:', error?.response?.data?.error?.type);
+    }
  }
  res.json({

--- a/src/types/index.d.ts
+++ b/src/types/index.d.ts
@@ -4,7 +4,7 @@ import type { RedisClientType } from 'redis';
 declare global {
  var mongodb: Mongoose | string | null;
  var redisClient: RedisClientType | null;
-  var generatingQA: number;
+  var generatingQA: boolean;
  var generatingAbstract: boolean;
  var generatingVector: boolean;
  var QRCode: any;