feat(rfa-ai): Complete RFA Approval Refactor and AI Model Revision

2026-05-16 10:59:53 +07:00
parent 6cb3ae10ee
commit 1a162bf320
105 changed files with 5088 additions and 1083 deletions
@@ -0,0 +1,113 @@
+// File: src/modules/ai/processors/ai-batch.processor.ts
+// Change Log
+// - 2026-05-15: เพิ่ม processor สำหรับ ai-batch queue ตาม ADR-023A.
+// - 2026-05-15: เพิ่ม EmbeddingService สำหรับ embed-document logic (T022).
+
+import { Processor, WorkerHost } from '@nestjs/bullmq';
+import { Logger } from '@nestjs/common';
+import { Job } from 'bullmq';
+import { InjectRepository } from '@nestjs/typeorm';
+import { Repository } from 'typeorm';
+import { Attachment } from '../../../common/file-storage/entities/attachment.entity';
+import { QUEUE_AI_BATCH } from '../../common/constants/queue.constants';
+import { EmbeddingService } from '../services/embedding.service';
+
+export type AiBatchJobType = 'ocr' | 'extract-metadata' | 'embed-document';
+
+export interface AiBatchJobData {
+  jobType: AiBatchJobType;
+  documentPublicId: string;
+  projectPublicId: string;
+  payload: Record<string, unknown>;
+  batchId?: string;
+  idempotencyKey: string;
+}
+
+/** Processor สำหรับงาน AI batch ที่รันทีละงานเพื่อคุม VRAM */
+@Processor(QUEUE_AI_BATCH, { concurrency: 1 })
+export class AiBatchProcessor extends WorkerHost {
+  private readonly logger = new Logger(AiBatchProcessor.name);
+
+  constructor(
+    @InjectRepository(Attachment)
+    private readonly attachmentRepo: Repository<Attachment>,
+    private readonly embeddingService: EmbeddingService
+  ) {
+    super();
+  }
+
+  /** Dispatch งาน batch ตาม jobType */
+  async process(job: Job<AiBatchJobData>): Promise<void> {
+    await this.setAiProcessingStatus(job.data.documentPublicId, 'PROCESSING');
+    try {
+      switch (job.data.jobType) {
+        case 'ocr':
+          this.logger.log(`OCR batch job processing — jobId=${String(job.id)}`);
+          // OCR logic handled by OcrService in ai-realtime processor
+          await this.setAiProcessingStatus(job.data.documentPublicId, 'DONE');
+          return;
+        case 'extract-metadata':
+          this.logger.log(
+            `Metadata extraction job processing — jobId=${String(job.id)}`
+          );
+          // Metadata extraction handled in ai-realtime processor
+          await this.setAiProcessingStatus(job.data.documentPublicId, 'DONE');
+          return;
+        case 'embed-document':
+          this.logger.log(`Embedding job processing — jobId=${String(job.id)}`);
+          await this.processEmbedDocument(job.data);
+          await this.setAiProcessingStatus(job.data.documentPublicId, 'DONE');
+          return;
+        default: {
+          const unreachable: never = job.data.jobType;
+          throw new Error(
+            `Unsupported ai-batch jobType: ${String(unreachable)}`
+          );
+        }
+      }
+    } catch (err) {
+      this.logger.error(
+        `Batch job failed — jobType=${job.data.jobType}, documentPublicId=${job.data.documentPublicId}`,
+        err instanceof Error ? err.stack : String(err)
+      );
+      await this.setAiProcessingStatus(job.data.documentPublicId, 'FAILED');
+      throw err;
+    }
+  }
+
+  /** ประมวลผล embed-document job ด้วย EmbeddingService (T022) */
+  private async processEmbedDocument(data: AiBatchJobData): Promise<void> {
+    const { documentPublicId, projectPublicId, payload } = data;
+    const pdfPath = payload.pdfPath as string;
+    const extractedText = payload.extractedText as string | undefined;
+
+    if (!pdfPath) {
+      throw new Error('pdfPath is required for embed-document job');
+    }
+
+    const result = await this.embeddingService.embedDocument(
+      pdfPath,
+      documentPublicId,
+      projectPublicId,
+      extractedText
+    );
+
+    if (!result.success) {
+      throw new Error(`Embedding failed: ${result.error ?? 'Unknown error'}`);
+    }
+
+    this.logger.log(
+      `Embedding completed for document ${documentPublicId} — ${result.chunksEmbedded} chunks embedded`
+    );
+  }
+
+  private async setAiProcessingStatus(
+    documentPublicId: string,
+    status: 'PENDING' | 'PROCESSING' | 'DONE' | 'FAILED'
+  ): Promise<void> {
+    await this.attachmentRepo.update(
+      { publicId: documentPublicId },
+      { aiProcessingStatus: status }
+    );
+  }
+}