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

backend/src/modules/rag/__tests__/rag.service.spec.ts

@@ -6,7 +6,7 @@ import { getQueueToken } from '@nestjs/bullmq';
 import { RagService } from '../rag.service';
 import { QdrantService } from '../qdrant.service';
 import { EmbeddingService } from '../embedding.service';
-import { TyphoonService } from '../typhoon.service';
+import { LocalLlmService } from '../local-llm.service';
 import { IngestionService } from '../ingestion.service';
 import { DocumentChunk } from '../entities/document-chunk.entity';
 import { QUEUE_AI_VECTOR_DELETION } from '../../common/constants/queue.constants';
@@ -23,7 +23,7 @@ const mockEmbedding = {
   embed: jest.fn(),
 };
 
-const mockTyphoon = {
+const mockLocalLlm = {
   generate: jest.fn(),
   sanitizeInput: jest.fn((t: string) => t),
 };
@@ -56,7 +56,7 @@ describe('RagService', () => {
         RagService,
         { provide: QdrantService, useValue: mockQdrant },
         { provide: EmbeddingService, useValue: mockEmbedding },
-        { provide: TyphoonService, useValue: mockTyphoon },
+        { provide: LocalLlmService, useValue: mockLocalLlm },
         { provide: IngestionService, useValue: mockIngestion },
         { provide: getRepositoryToken(DocumentChunk), useValue: mockChunkRepo },
         { provide: DEFAULT_REDIS_TOKEN, useValue: mockRedis },
@@ -95,7 +95,7 @@ describe('RagService', () => {
           score: 0.92,
         },
       ]);
-      mockTyphoon.generate.mockResolvedValue({
+      mockLocalLlm.generate.mockResolvedValue({
         answer: 'คำตอบ',
         usedFallbackModel: false,
       });
@@ -129,20 +129,17 @@ describe('RagService', () => {
       mockQdrant.isReady.mockReturnValue(true);
       mockEmbedding.embed.mockResolvedValue(new Array(768).fill(0.1));
       mockQdrant.hybridSearch.mockResolvedValue([]);
-      mockTyphoon.generate.mockResolvedValue({
+      mockLocalLlm.generate.mockResolvedValue({
         answer: 'ลับมาก',
-        usedFallbackModel: true,
+        usedFallbackModel: false,
       });
 
       const result = await service.query(dto, adminPerms);
 
       expect(mockRedis.get).not.toHaveBeenCalled();
       expect(mockRedis.setex).not.toHaveBeenCalled();
-      expect(mockTyphoon.generate).toHaveBeenCalledWith(
-        expect.any(String),
-        true
-      );
-      expect(result.usedFallbackModel).toBe(true);
+      expect(mockLocalLlm.generate).toHaveBeenCalledWith(expect.any(String));
+      expect(result.usedFallbackModel).toBe(false);
     });
 
     it('collectionReady=false → throw ServiceUnavailableException RAG_NOT_READY', async () => {
@@ -158,7 +155,7 @@ describe('RagService', () => {
       mockRedis.get.mockResolvedValue(null);
       mockEmbedding.embed.mockResolvedValue(new Array(768).fill(0.1));
       mockQdrant.hybridSearch.mockResolvedValue([]);
-      mockTyphoon.generate.mockResolvedValue({
+      mockLocalLlm.generate.mockResolvedValue({
         answer: 'A',
         usedFallbackModel: false,
       });
@@ -181,7 +178,7 @@ describe('RagService', () => {
       mockRedis.get.mockResolvedValue(null);
       mockEmbedding.embed.mockResolvedValue(new Array(768).fill(0.1));
       mockQdrant.hybridSearch.mockResolvedValue([]);
-      mockTyphoon.generate.mockResolvedValue({
+      mockLocalLlm.generate.mockResolvedValue({
         anwer: 'ok',
         usedFallbackModel: false,
       });
@@ -199,9 +196,9 @@ describe('RagService', () => {
       mockRedis.get.mockResolvedValue(null);
       mockEmbedding.embed.mockResolvedValue(new Array(768).fill(0.1));
       mockQdrant.hybridSearch.mockResolvedValue([]);
-      mockTyphoon.generate.mockResolvedValue({
+      mockLocalLlm.generate.mockResolvedValue({
         answer: 'ok',
-        usedFallbackModel: true,
+        usedFallbackModel: false,
       });
 
       await service.query(dto, adminPerms);

backend/src/modules/rag/local-llm.service.ts

@@ -0,0 +1,67 @@
+// File: src/modules/rag/local-llm.service.ts
+// Change Log
+// - 2026-05-15: แทนที่ cloud LLM API ด้วย Ollama local-only ตาม ADR-023A.
+
+import { Injectable, Logger } from '@nestjs/common';
+import { ConfigService } from '@nestjs/config';
+import axios from 'axios';
+
+export interface LlmGenerateResult {
+  answer: string;
+  usedFallbackModel: boolean;
+}
+
+/** บริการเรียก LLM ภายในองค์กรผ่าน Ollama เท่านั้น */
+@Injectable()
+export class LocalLlmService {
+  private readonly logger = new Logger(LocalLlmService.name);
+  private readonly ollamaUrl: string;
+  private readonly ollamaModel: string;
+  private readonly timeoutMs: number;
+
+  constructor(private readonly configService: ConfigService) {
+    this.ollamaUrl = this.configService.get<string>(
+      'OLLAMA_URL',
+      this.configService.get<string>('AI_HOST_URL', 'http://localhost:11434')
+    );
+    this.ollamaModel = this.configService.get<string>(
+      'OLLAMA_MODEL_MAIN',
+      this.configService.get<string>('OLLAMA_RAG_MODEL', 'gemma4:e4b')
+    );
+    this.timeoutMs = this.configService.get<number>('RAG_TIMEOUT_MS', 30000);
+  }
+
+  /** สร้างคำตอบจากโมเดล local-only โดยไม่มี cloud fallback */
+  async generate(prompt: string): Promise<LlmGenerateResult> {
+    try {
+      const response = await axios.post<{ response: string }>(
+        `${this.ollamaUrl}/api/generate`,
+        {
+          model: this.ollamaModel,
+          prompt,
+          stream: false,
+        },
+        { timeout: this.timeoutMs }
+      );
+      return {
+        answer: response.data.response ?? '',
+        usedFallbackModel: false,
+      };
+    } catch (err) {
+      this.logger.error(
+        'Local Ollama generation failed',
+        err instanceof Error ? err.stack : String(err)
+      );
+      throw err;
+    }
+  }
+
+  /** ทำความสะอาด prompt injection pattern พื้นฐานก่อนส่งเข้าโมเดล */
+  sanitizeInput(text: string): string {
+    return text
+      .replace(/<CONTEXT_START>|<CONTEXT_END>/gi, '')
+      .replace(/ignore previous instructions/gi, '')
+      .replace(/system:/gi, '')
+      .slice(0, 1000);
+  }
+}

backend/src/modules/rag/rag.module.ts

@@ -7,7 +7,7 @@ import { DocumentChunk } from './entities/document-chunk.entity';
 import { QUEUE_AI_VECTOR_DELETION } from '../common/constants/queue.constants';
 import { EmbeddingService } from './embedding.service';
 import { QdrantService } from './qdrant.service';
-import { TyphoonService } from './typhoon.service';
+import { LocalLlmService } from './local-llm.service';
 import { RagService } from './rag.service';
 import { RagController } from './rag.controller';
 import { IngestionService } from './ingestion.service';
@@ -40,7 +40,7 @@ const DLQ_DEFAULTS = {
   providers: [
     EmbeddingService,
     QdrantService,
-    TyphoonService,
+    LocalLlmService,
     RagService,
     IngestionService,
     OcrProcessor,
@@ -50,7 +50,7 @@ const DLQ_DEFAULTS = {
   exports: [
     EmbeddingService,
     QdrantService,
-    TyphoonService,
+    LocalLlmService,
     RagService,
     IngestionService,
   ],

backend/src/modules/rag/rag.service.ts

@@ -16,7 +16,7 @@ import { createHash } from 'crypto';
 
 import { QdrantService } from './qdrant.service';
 import { EmbeddingService } from './embedding.service';
-import { TyphoonService } from './typhoon.service';
+import { LocalLlmService } from './local-llm.service';
 import { IngestionService } from './ingestion.service';
 import { DocumentChunk } from './entities/document-chunk.entity';
 import { RagQueryDto } from './dto/rag-query.dto';
@@ -32,7 +32,7 @@ export class RagService {
   constructor(
     private readonly qdrant: QdrantService,
     private readonly embedding: EmbeddingService,
-    private readonly typhoon: TyphoonService,
+    private readonly localLlm: LocalLlmService,
     private readonly ingestionService: IngestionService,
     @InjectRepository(DocumentChunk)
     private readonly chunkRepo: Repository<DocumentChunk>,
@@ -84,13 +84,10 @@ export class RagService {
 
     const context = this.buildContext(reranked);
 
-    const safeQuestion = this.typhoon.sanitizeInput(question);
+    const safeQuestion = this.localLlm.sanitizeInput(question);
     const prompt = this.buildPrompt(safeQuestion, context);
 
-    const { answer, usedFallbackModel } = await this.typhoon.generate(
-      prompt,
-      isConfidential
-    );
+    const { answer, usedFallbackModel } = await this.localLlm.generate(prompt);
 
     const citations: RagCitation[] = reranked.map((r) => ({
       chunkId: r.chunkId,

backend/src/modules/rag/typhoon.service.ts

@@ -1,115 +0,0 @@
-import { Injectable, Logger } from '@nestjs/common';
-import { ConfigService } from '@nestjs/config';
-import axios from 'axios';
-
-export interface LlmGenerateResult {
-  answer: string;
-  usedFallbackModel: boolean;
-}
-
-interface TyphoonChatResponse {
-  choices: Array<{ message: { content: string } }>;
-}
-
-@Injectable()
-export class TyphoonService {
-  private readonly logger = new Logger(TyphoonService.name);
-  private readonly typhoonUrl: string;
-  private readonly typhoonKey: string;
-  private readonly ollamaUrl: string;
-  private readonly ollamaModel: string;
-  private readonly timeoutMs: number;
-
-  constructor(private readonly configService: ConfigService) {
-    this.typhoonUrl = this.configService.get<string>(
-      'TYPHOON_API_URL',
-      'https://api.opentyphoon.ai/v1'
-    );
-    this.typhoonKey = this.configService.get<string>('TYPHOON_API_KEY', '');
-    this.ollamaUrl = this.configService.get<string>(
-      'OLLAMA_URL',
-      'http://localhost:11434'
-    );
-    this.ollamaModel = this.configService.get<string>(
-      'OLLAMA_RAG_MODEL',
-      'gemma3:12b'
-    );
-    this.timeoutMs = this.configService.get<number>('RAG_TIMEOUT_MS', 5000);
-  }
-
-  async generate(
-    prompt: string,
-    forceLocal: boolean
-  ): Promise<LlmGenerateResult> {
-    if (forceLocal) {
-      const answer = await this.generateOllama(prompt);
-      return { answer, usedFallbackModel: true };
-    }
-
-    try {
-      const answer = await Promise.race([
-        this.generateTyphoon(prompt),
-        this.delay(this.timeoutMs).then(() => {
-          throw new Error('Typhoon timeout');
-        }),
-      ]);
-      return { answer, usedFallbackModel: false };
-    } catch (err) {
-      this.logger.warn(
-        `Typhoon failed, falling back to Ollama: ${err instanceof Error ? err.message : String(err)}`
-      );
-      const answer = await this.generateOllama(prompt);
-      return { answer, usedFallbackModel: true };
-    }
-  }
-
-  sanitizeInput(text: string): string {
-    return text
-      .replace(/<CONTEXT_START>|<CONTEXT_END>/gi, '')
-      .replace(/ignore previous instructions/gi, '')
-      .replace(/system:/gi, '')
-      .slice(0, 1000);
-  }
-
-  private async generateTyphoon(prompt: string): Promise<string> {
-    const response = await axios.post<TyphoonChatResponse>(
-      `${this.typhoonUrl}/chat/completions`,
-      {
-        model: 'typhoon-v2.1-12b-instruct',
-        messages: [
-          {
-            role: 'user',
-            content: `<CONTEXT_START>\n${prompt}\n<CONTEXT_END>`,
-          },
-        ],
-        max_tokens: 1024,
-        temperature: 0.1,
-      },
-      {
-        headers: {
-          Authorization: `Bearer ${this.typhoonKey}`,
-          'Content-Type': 'application/json',
-        },
-        timeout: this.timeoutMs,
-      }
-    );
-    return response.data.choices[0]?.message?.content ?? '';
-  }
-
-  private async generateOllama(prompt: string): Promise<string> {
-    const response = await axios.post<{ response: string }>(
-      `${this.ollamaUrl}/api/generate`,
-      {
-        model: this.ollamaModel,
-        prompt,
-        stream: false,
-      },
-      { timeout: 30000 }
-    );
-    return response.data.response ?? '';
-  }
-
-  private delay(ms: number): Promise<void> {
-    return new Promise((resolve) => setTimeout(resolve, ms));
-  }
-}