요약 실패, Input Token 적은 경우에 대한 대응

package.json

@@ -51,6 +51,7 @@
     "cross-env": "^7.0.3",
     "express": "^4.19.2",
     "iconv-lite": "^0.6.3",
+    "js-tiktoken": "^1.0.14",
     "lodash": "^4.17.21",
     "mongoose": "^8.4.1",
     "openai": "^4.52.2",

src/app.module.ts

@@ -13,12 +13,12 @@ import { DiscordModule } from './infrastructure/discord/discord.module';
 import { AuthModule } from './modules/auth/auth.module';
 import { ClassificationModule } from './modules/classification/classification.module';
 import { FoldersModule } from './modules/folders/folders.module';
+import { LaunchingEventsModule } from './modules/launching-events/launching-events.module';
 import { LinksModule } from './modules/links/links.module';
 import { MetricsModule } from './modules/metrics/metrics.module';
 import { OnboardModule } from './modules/onboard/onboard.module';
 import { PostsModule } from './modules/posts/posts.module';
 import { UsersModule } from './modules/users/users.module';
-import { LaunchingEventsModule } from './modules/launching-events/launching-events.module';
 
 @Module({
   imports: [

src/bootstrap.ts

@@ -1,9 +1,9 @@
 import { NestFactory } from '@nestjs/core';
 import { ExpressAdapter } from '@nestjs/platform-express';
 import express from 'express';
+import { nestAppConfig, nestResponseConfig } from './app.config';
 import { AppModule } from './app.module';
 import { nestSwaggerConfig } from './app.swagger';
-import { nestAppConfig, nestResponseConfig } from './app.config';
 
 export async function bootstrap() {
   const expressInstance: express.Express = express();

src/common/constant/index.ts

@@ -1 +1,2 @@
 export const IS_LOCAL = process.env.NODE_ENV === 'local';
+export const DEFAULT_FOLDER_NAME = '나중에 읽을 링크';

src/common/utils/parser.util.ts

@@ -11,6 +11,7 @@ export async function parseLinkTitleAndContent(url: string): Promise<{
   title: string;
   content: string;
   thumbnail: string;
+  thumbnailDescription: string;
 }> {
   const response = await fetch(url);
   const arrayBuffer = await response.arrayBuffer();
@@ -49,6 +50,10 @@ export async function parseLinkTitleAndContent(url: string): Promise<{
   const title = $('title').text();
   // Page Thumbnail Parsing
   const thumbnail = $('meta[property="og:image"]').attr('content');
+  // Page Thumbnail Description Parsing
+  const thumbnailDescription = $('meta[property="og:description"]').attr(
+    'content',
+  );
   // HTML Body내에 있는 script태그랑 css style태그 제거
   $('body script, body style').remove();
   // HTML Element의 body의 text content
@@ -61,6 +66,7 @@ export async function parseLinkTitleAndContent(url: string): Promise<{
     title: title ?? '',
     content,
     thumbnail: sanitizeThumbnail(thumbnail),
+    thumbnailDescription,
   };
 }
 

src/common/utils/tokenizer.ts

@@ -0,0 +1,41 @@
+// Do not import 'tiktoken'
+
+import { gptVersion } from '@src/infrastructure/ai/ai.constant';
+import { summarizeURLContentFunctionFactory } from '@src/infrastructure/ai/functions';
+import { encodingForModel } from 'js-tiktoken';
+
+const encoder = encodingForModel(gptVersion);
+
+// Reference: https://platform.openai.com/docs/advanced-usage/managing-tokens
+// 주의: 실제 Open AI랑 약간의 오차 존재
+export function promptTokenCalculator(content: string, folderList: string[]) {
+  let tokenCount = 0;
+
+  // Prompt Calculation
+  const messages = [
+    {
+      role: 'system',
+      content: '한글로 답변 부탁해',
+    },
+    {
+      role: 'user',
+      content: `주어진 글에 대해 요약하고 키워드 추출, 분류 부탁해
+
+${content}
+      `,
+    },
+  ];
+  for (const message of messages) {
+    // Message struct Overhead
+    tokenCount += 4;
+    tokenCount += encoder.encode(message.role).length;
+    tokenCount += encoder.encode(message.content).length;
+  }
+  tokenCount += 2;
+
+  // Function Calculation
+  tokenCount += encoder.encode(
+    JSON.stringify(summarizeURLContentFunctionFactory(folderList)),
+  ).length;
+  return tokenCount;
+}

src/infrastructure/ai/ai.service.ts

@@ -1,5 +1,6 @@
 import { Injectable } from '@nestjs/common';
 import { ConfigService } from '@nestjs/config';
+import { promptTokenCalculator } from '@src/common/utils/tokenizer';
 import OpenAI, { OpenAIError, RateLimitError } from 'openai';
 import { DiscordAIWebhookProvider } from '../discord/discord-ai-webhook.provider';
 import { gptVersion } from './ai.constant';
@@ -14,6 +15,7 @@ import { SummarizeURLContent } from './types/types';
 @Injectable()
 export class AiService {
   private openai: OpenAI;
+  private leastTokenThreshold = 300;
 
   constructor(
     private readonly config: ConfigService,
@@ -26,13 +28,23 @@ export class AiService {
 
   async summarizeLinkContent(
     content: string,
+    baseThumbnailContent: string,
     userFolderList: string[],
     url: string,
     temperature = 0.5,
   ): Promise<SummarizeURLContentDto> {
     try {
       // 사용자 폴더 + 서버에서 임의로 붙여주는 폴더 리스트
       const folderLists = [...userFolderList];
+      // Calculate post content
+      const tokenCount = promptTokenCalculator(content, folderLists);
+      if (tokenCount <= this.leastTokenThreshold) {
+        return new SummarizeURLContentDto({
+          success: false,
+          message: 'Too low input token count',
+          thumbnailContent: baseThumbnailContent,
+        });
+      }
       // AI Summary 호출
       const promptResult = await this.invokeAISummary(
         content,
@@ -63,7 +75,18 @@ export class AiService {
             : err instanceof OpenAIError
               ? err.message
               : '요약에 실패하였습니다.',
+        thumbnailContent: baseThumbnailContent,
       });
+
+      // return new SummarizeURLContentDto({
+      //   success: false,
+      //   message:
+      //     err instanceof RateLimitError
+      //       ? '크레딧을 모두 소진하였습니다.'
+      //       : err instanceof OpenAIError
+      //         ? err.message
+      //         : '요약에 실패하였습니다.',
+      // });
     }
   }
 
@@ -104,7 +127,6 @@ ${content}
     const functionResult: AiClassificationFunctionResult = JSON.parse(
       promptResult.choices[0].message.tool_calls[0].function.arguments,
     );
-
     return functionResult;
   }
 

src/infrastructure/ai/dto/summarizeURL.response.ts

@@ -9,6 +9,7 @@ type SummarizeSuccessType = {
 type SummarizeFailType = {
   success: false;
   message: string;
+  thumbnailContent: string;
 };
 
 type SummarizeResultType = SummarizeSuccessType | SummarizeFailType;
@@ -18,6 +19,7 @@ export class SummarizeURLContentDto {
   isUserCategory?: boolean;
   response?: SummarizeURLContent;
   message?: string;
+  thumbnailContent?: string;
 
   constructor(data: SummarizeResultType) {
     // true를 명시하지 않으면 Discriminate Union이 동작 안함
@@ -27,6 +29,7 @@ export class SummarizeURLContentDto {
       this.response = data.response;
     } else {
       this.message = data.message;
+      this.thumbnailContent = data.thumbnailContent;
     }
   }
 }

src/infrastructure/aws-lambda/type.ts

@@ -1,5 +1,6 @@
 export type AiClassificationPayload = {
   postContent: string;
+  postThumbnailContent: string;
   folderList: { id: string; name: string }[];
   userId: string;
   postId: string;

src/modules/ai-classification/ai-classification.service.ts

@@ -34,6 +34,7 @@ export class AiClassificationService {
       const start = process.hrtime();
       const summarizeUrlContent = await this.aiService.summarizeLinkContent(
         payload.postContent,
+        payload.postThumbnailContent,
         Object.keys(folderMapper),
         payload.url,
       );
@@ -100,15 +101,17 @@ export class AiClassificationService {
       await this.metricsRepository.createMetrics(
         summarizeUrlContent.success,
         timeSecond,
-        post.url,
-        post._id.toString(),
+        payload.url,
+        payload.postId,
       );
 
       await this.postRepository.updatePostClassificationForAIClassification(
         postAiStatus,
         postId,
         classificationId,
-        summarizeUrlContent.response.summary,
+        summarizeUrlContent.success === true
+          ? summarizeUrlContent.response.summary
+          : summarizeUrlContent.thumbnailContent,
       );
       return summarizeUrlContent;
     } catch (error: unknown) {

src/modules/posts/posts.service.ts

@@ -79,10 +79,8 @@ export class PostsService {
     });
 
     // NOTE : URL에서 얻은 정보 가져옴
-    const { title, content, thumbnail } = await parseLinkTitleAndContent(
-      createPostDto.url,
-    );
-
+    const { title, content, thumbnail, thumbnailDescription } =
+      await parseLinkTitleAndContent(createPostDto.url);
     const userFolderList = await this.folderRepository.findByUserId(userId);
     const folderList = userFolderList.map((folder) => {
       return {
@@ -101,6 +99,7 @@ export class PostsService {
     const payload = {
       url: createPostDto.url,
       postContent: content,
+      postThumbnailContent: thumbnailDescription,
       folderList: folderList,
       postId: post._id.toString(),
       userId,

src/modules/users/users.service.ts

@@ -1,4 +1,5 @@
 import { Injectable } from '@nestjs/common';
+import { DEFAULT_FOLDER_NAME } from '@src/common/constant';
 import { FolderType } from '@src/infrastructure/database/types/folder-type.enum';
 import { JwtPayload } from 'src/common/types/type';
 import { AuthService } from '../auth/auth.service';
@@ -23,7 +24,7 @@ export class UsersService {
       user = await this.userRepository.findOrCreate(dto.deviceToken);
       await this.folderRepository.create(
         user._id.toString(),
-        '나중에 읽을 링크',
+        DEFAULT_FOLDER_NAME,
         FolderType.DEFAULT,
       );
     }

tsconfig.json

@@ -6,6 +6,7 @@
     "emitDecoratorMetadata": true,
     "experimentalDecorators": true,
     "allowSyntheticDefaultImports": true,
+    "resolveJsonModule": true,
     "target": "ES2021",
     "sourceMap": true,
     "outDir": "./dist",