690530:1121 ADR-030-231-ocr-sandbox-two-step-flow #01

backend/src/modules/ai/ai-queue.service.ts

@@ -103,7 +103,11 @@ export class AiQueueService {
    * @idempotency `jobId = payload.idempotencyKey`
    */
   async enqueueSandboxJob(
-    jobType: 'sandbox-rag' | 'sandbox-extract',
+    jobType:
+      | 'sandbox-rag'
+      | 'sandbox-extract'
+      | 'sandbox-ocr-only'
+      | 'sandbox-ai-extract',
     payload: {
       idempotencyKey: string;
       projectPublicId?: string;
@@ -111,6 +115,7 @@ export class AiQueueService {
       userPublicId?: string;
       filePublicId?: string;
       pdfPath?: string;
+      extraPayload?: Record<string, unknown>;
     }
   ): Promise<string> {
     const job = await this.batchQueue.add(
@@ -124,6 +129,7 @@ export class AiQueueService {
           userPublicId: payload.userPublicId,
           filePublicId: payload.filePublicId,
           pdfPath: payload.pdfPath,
+          ...payload.extraPayload,
         },
         idempotencyKey: payload.idempotencyKey,
       },

backend/src/modules/ai/ai.controller.ts

@@ -40,6 +40,8 @@ import {
   ApiHeader,
   ApiParam,
   ApiQuery,
+  ApiConsumes,
+  ApiBody,
 } from '@nestjs/swagger';
 import { AiService, ExtractionResult, PaginatedResult } from './ai.service';
 import { AiSettingsService } from './ai-settings.service';
@@ -508,6 +510,77 @@ export class AiController {
     return { requestPublicId, jobId, status: 'queued' };
   }
 
+  // --- Step 1: OCR Only (สำหรับตรวจคุณภาพ OCR ก่อนทดสอบ AI) ---
+
+  @Post('admin/sandbox/ocr')
+  @UseGuards(JwtAuthGuard, RbacGuard)
+  @RequirePermission('system.manage_all')
+  @ApiOperation({
+    summary: 'Step 1: Run OCR Only — สำหรับตรวจคุณภาพ OCR ก่อนทดสอบ AI',
+    description:
+      'Upload PDF และรัน OCR เท่านั้น ไม่เรียก LLM — ผลลัพธ์ cache ไว้สำหรับ Step 2',
+  })
+  @ApiConsumes('multipart/form-data')
+  @ApiBody({
+    schema: {
+      type: 'object',
+      properties: {
+        file: {
+          type: 'string',
+          format: 'binary',
+        },
+      },
+    },
+  })
+  async submitSandboxOcr(
+    @UploadedFile(
+      new ParseFilePipe({
+        validators: [
+          new MaxFileSizeValidator({ maxSize: 50 * 1024 * 1024 }),
+          new FileTypeValidator({ fileType: 'pdf' }),
+        ],
+      })
+    )
+    file: Express.Multer.File,
+    @CurrentUser() user: User
+  ): Promise<{ requestPublicId: string; jobId: string; status: string }> {
+    const attachment = await this.fileStorageService.upload(file, user.user_id);
+    const requestPublicId = uuidv7();
+    const jobId = await this.aiQueueService.enqueueSandboxJob(
+      'sandbox-ocr-only',
+      {
+        idempotencyKey: requestPublicId,
+        pdfPath: attachment.filePath,
+      }
+    );
+    return { requestPublicId, jobId, status: 'queued' };
+  }
+
+  // --- Step 2: AI Extraction (ใช้ OCR text ที่ cache จาก Step 1) ---
+
+  @Post('admin/sandbox/ai-extract')
+  @UseGuards(JwtAuthGuard, RbacGuard)
+  @RequirePermission('system.manage_all')
+  @ApiOperation({
+    summary: 'Step 2: Run AI Extraction — ใช้ OCR text ที่ cache จาก Step 1',
+    description:
+      'รับ requestPublicId จาก Step 1 และ optional promptVersion แล้ว run LLM extraction',
+  })
+  async submitSandboxAiExtract(
+    @Body() dto: { requestPublicId: string; promptVersion?: number }
+  ): Promise<{ requestPublicId: string; jobId: string; status: string }> {
+    const { requestPublicId, promptVersion } = dto;
+    const jobId = await this.aiQueueService.enqueueSandboxJob(
+      'sandbox-ai-extract',
+      {
+        idempotencyKey: requestPublicId,
+        projectPublicId: 'default', // Sandbox ใช้ default project
+        extraPayload: { promptVersion },
+      }
+    );
+    return { requestPublicId, jobId, status: 'queued' };
+  }
+
   // --- Webhook Callback จาก n8n (Service Account) ---
 
   @Post('callback')

backend/src/modules/ai/processors/ai-batch.processor.ts

@@ -49,6 +49,8 @@ export type AiBatchJobType =
   | 'embed-document'
   | 'sandbox-rag'
   | 'sandbox-extract'
+  | 'sandbox-ocr-only'
+  | 'sandbox-ai-extract'
   | 'migrate-document';
 
 export interface AiBatchJobData {
@@ -197,6 +199,18 @@ export class AiBatchProcessor extends WorkerHost {
           );
           await this.processSandboxExtract(job.data);
           return;
+        case 'sandbox-ocr-only':
+          this.logger.log(
+            `Sandbox OCR-Only job processing — jobId=${String(job.id)}`
+          );
+          await this.processSandboxOcrOnly(job.data);
+          return;
+        case 'sandbox-ai-extract':
+          this.logger.log(
+            `Sandbox AI-Extract job processing — jobId=${String(job.id)}`
+          );
+          await this.processSandboxAiExtract(job.data);
+          return;
         case 'migrate-document':
           this.logger.log(
             `Migrate document job processing — jobId=${String(job.id)}`
@@ -369,6 +383,186 @@ export class AiBatchProcessor extends WorkerHost {
     }
   }
 
+  /** Step 1: OCR เท่านั้น — สำหรับตรวจคุณภาพ OCR ก่อนทดสอบ AI */
+  private async processSandboxOcrOnly(data: AiBatchJobData): Promise<void> {
+    const { idempotencyKey, payload } = data;
+    const pdfPath = payload.pdfPath as string;
+
+    if (!pdfPath) {
+      throw new Error('pdfPath is required for sandbox-ocr-only job');
+    }
+
+    await this.redis.setex(
+      `ai:rag:result:${idempotencyKey}`,
+      3600,
+      JSON.stringify({
+        requestPublicId: idempotencyKey,
+        status: 'processing',
+      })
+    );
+
+    try {
+      const ocrResult = await this.ocrService.detectAndExtract({ pdfPath });
+
+      // Cache OCR text สำหรับ Step 2
+      await this.redis.setex(
+        `ai:sandbox:ocr:${idempotencyKey}`,
+        3600,
+        JSON.stringify({
+          ocrText: ocrResult.text,
+          ocrUsed: ocrResult.ocrUsed,
+          timestamp: new Date().toISOString(),
+        })
+      );
+
+      await this.redis.setex(
+        `ai:rag:result:${idempotencyKey}`,
+        3600,
+        JSON.stringify({
+          requestPublicId: idempotencyKey,
+          status: 'completed',
+          ocrText: ocrResult.text,
+          ocrUsed: ocrResult.ocrUsed,
+          completedAt: new Date().toISOString(),
+        })
+      );
+    } catch (err: unknown) {
+      const errMsg = err instanceof Error ? err.message : String(err);
+      this.logger.error(`Sandbox OCR-only failed: ${errMsg}`);
+      await this.redis.setex(
+        `ai:rag:result:${idempotencyKey}`,
+        3600,
+        JSON.stringify({
+          requestPublicId: idempotencyKey,
+          status: 'failed',
+          errorMessage: errMsg,
+          completedAt: new Date().toISOString(),
+        })
+      );
+      throw err;
+    }
+  }
+
+  /** Step 2: AI Extraction — ใช้ OCR text ที่ cache จาก Step 1 */
+  private async processSandboxAiExtract(data: AiBatchJobData): Promise<void> {
+    const { idempotencyKey, payload, projectPublicId } = data;
+    const promptVersion = (payload.promptVersion as number) || undefined;
+
+    await this.redis.setex(
+      `ai:rag:result:${idempotencyKey}`,
+      3600,
+      JSON.stringify({
+        requestPublicId: idempotencyKey,
+        status: 'processing',
+      })
+    );
+
+    try {
+      // ดึง OCR text จาก cache
+      const cachedOcr = await this.redis.get(
+        `ai:sandbox:ocr:${idempotencyKey}`
+      );
+      if (!cachedOcr) {
+        throw new Error(
+          'OCR text not found or expired, please run Step 1 first'
+        );
+      }
+      const parsedOcr = JSON.parse(cachedOcr) as {
+        ocrText: string;
+        ocrUsed: boolean;
+        timestamp: string;
+      };
+      const { ocrText } = parsedOcr;
+
+      // ดึง prompt version
+      const activePrompt =
+        await this.aiPromptsService.getActive('ocr_extraction');
+      if (!activePrompt) {
+        throw new Error('No active ocr_extraction prompt version found');
+      }
+
+      // ถ้าระบุ promptVersion ให้ใช้ version นั้น
+      const targetPrompt = promptVersion
+        ? await this.aiPromptsService.findByVersion(
+            'ocr_extraction',
+            promptVersion
+          )
+        : activePrompt;
+
+      if (!targetPrompt) {
+        throw new Error(`Prompt version ${promptVersion} not found`);
+      }
+
+      // Resolve context และ run LLM
+      const masterDataContext = await this.aiPromptsService.resolveContext(
+        targetPrompt,
+        projectPublicId
+      );
+
+      const resolvedPrompt = targetPrompt.template
+        .replace('{{ocr_text}}', ocrText)
+        .replace(
+          '{{master_data_context}}',
+          JSON.stringify(masterDataContext, null, 2)
+        );
+
+      const response = await this.ollamaService.generate(resolvedPrompt, {
+        timeoutMs: 120000,
+      });
+
+      const cleanedResponse = response
+        .replace(/```json/g, '')
+        .replace(/```/g, '')
+        .trim();
+
+      let extractedMetadata: Record<string, unknown>;
+      try {
+        extractedMetadata = JSON.parse(cleanedResponse) as Record<
+          string,
+          unknown
+        >;
+      } catch {
+        throw new Error(
+          `Failed to parse LLM response as JSON: ${cleanedResponse}`
+        );
+      }
+
+      await this.aiPromptsService.saveTestResult(
+        'ocr_extraction',
+        targetPrompt.versionNumber,
+        extractedMetadata
+      );
+
+      await this.redis.setex(
+        `ai:rag:result:${idempotencyKey}`,
+        3600,
+        JSON.stringify({
+          requestPublicId: idempotencyKey,
+          status: 'completed',
+          answer: JSON.stringify(extractedMetadata, null, 2),
+          ocrText,
+          ocrUsed: parsedOcr.ocrUsed,
+          promptVersionUsed: targetPrompt.versionNumber,
+          completedAt: new Date().toISOString(),
+        })
+      );
+    } catch (err: unknown) {
+      const errMsg = err instanceof Error ? err.message : String(err);
+      this.logger.error(`Sandbox AI-extract failed: ${errMsg}`);
+      await this.redis.setex(
+        `ai:rag:result:${idempotencyKey}`,
+        3600,
+        JSON.stringify({
+          requestPublicId: idempotencyKey,
+          status: 'failed',
+          errorMessage: errMsg,
+          completedAt: new Date().toISOString(),
+        })
+      );
+      throw err;
+    }
+  }
+
   private async processMigrateDocument(
     job: Job<AiBatchJobData>
   ): Promise<void> {

backend/src/modules/ai/prompts/ai-prompts.service.ts

@@ -292,6 +292,21 @@ export class AiPromptsService {
     return prompt;
   }
 
+  /**
+   * ดึง Prompt version ตาม versionNumber ที่ระบุ
+   * @param promptType ประเภทของ prompt
+   * @param versionNumber เลข version ที่ต้องการ
+   * @returns Prompt version ที่ตรงกับ versionNumber หรือ null หากไม่พบ
+   */
+  async findByVersion(
+    promptType: string,
+    versionNumber: number
+  ): Promise<AiPrompt | null> {
+    return this.aiPromptRepo.findOne({
+      where: { promptType, versionNumber },
+    });
+  }
+
   /**
    * ค้นหา prompt ที่มีผลใช้งานจริง และแทนที่ placeholder {{ocr_text}} ด้วยข้อความ OCR
    * @param promptType ประเภทของ prompt