ci: Bump version to 0.7.3 [skip ci]

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "flyer-crawler",
-  "version": "0.6.4",
+  "version": "0.7.3",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "flyer-crawler",
-      "version": "0.6.4",
+      "version": "0.7.3",
       "dependencies": {
         "@bull-board/api": "^6.14.2",
         "@bull-board/express": "^6.14.2",

package.json

@@ -1,7 +1,7 @@
 {
   "name": "flyer-crawler",
   "private": true,
-  "version": "0.6.4",
+  "version": "0.7.3",
   "type": "module",
   "scripts": {
     "dev": "concurrently \"npm:start:dev\" \"vite\"",

src/providers/AuthProvider.tsx

@@ -15,7 +15,7 @@ export const AuthProvider: React.FC<{ children: ReactNode }> = ({ children }) =>
   // FIX: Stabilize the apiFunction passed to useApi.
   // By wrapping this in useCallback, we ensure the same function instance is passed to
   // useApi on every render. This prevents the `execute` function returned by `useApi`
-  // from being recreated, which in turn breaks the infinite re-render loop in the useEffect below.
+  // from being recreated, which in turn breaks the infinite re-render loop in the useEffect.
   const getProfileCallback = useCallback(() => apiClient.getAuthenticatedUserProfile(), []);
 
   const { execute: checkTokenApi } = useApi<UserProfile, []>(getProfileCallback);

src/providers/FlyersProvider.tsx

@@ -4,17 +4,21 @@ import { FlyersContext, FlyersContextType } from '../contexts/FlyersContext';
 import type { Flyer } from '../types';
 import * as apiClient from '../services/apiClient';
 import { useInfiniteQuery } from '../hooks/useInfiniteQuery';
+import { useCallback } from 'react';
 
 export const FlyersProvider: React.FC<{ children: ReactNode }> = ({ children }) => {
+  // Memoize the fetch function to ensure stability for the useInfiniteQuery hook.
+  const fetchFlyersFn = useCallback(apiClient.fetchFlyers, []);
+
   const {
     data: flyers,
-    isLoading: isLoadingFlyers,
+    isLoading: isLoadingFlyers,    
     error: flyersError,
     fetchNextPage: fetchNextFlyersPage,
     hasNextPage: hasNextFlyersPage,
     refetch: refetchFlyers,
     isRefetching: isRefetchingFlyers,
-  } = useInfiniteQuery<Flyer>(apiClient.fetchFlyers);
+  } = useInfiniteQuery<Flyer>(fetchFlyersFn);
 
   const value: FlyersContextType = {
     flyers: flyers || [],
@@ -26,5 +30,5 @@ export const FlyersProvider: React.FC<{ children: ReactNode }> = ({ children })
     refetchFlyers,
   };
 
-  return <FlyersContext.Provider value={value}>{children}</FlyersContext.Provider>;
+ return <FlyersContext.Provider value={value}>{children}</FlyersContext.Provider>;
 };

src/providers/MasterItemsProvider.tsx

@@ -1,14 +1,22 @@
 // src/providers/MasterItemsProvider.tsx
-import React, { ReactNode, useMemo } from 'react';
+import React, { ReactNode, useMemo, useEffect, useCallback } from 'react';
 import { MasterItemsContext } from '../contexts/MasterItemsContext';
 import type { MasterGroceryItem } from '../types';
 import * as apiClient from '../services/apiClient';
 import { useApiOnMount } from '../hooks/useApiOnMount';
+import { logger } from '../services/logger.client';
 
 export const MasterItemsProvider: React.FC<{ children: ReactNode }> = ({ children }) => {
-  const { data, loading, error } = useApiOnMount<MasterGroceryItem[], []>(() =>
-    apiClient.fetchMasterItems(),
-  );
+  // LOGGING: Check if the provider is unmounting/remounting repeatedly
+  useEffect(() => {
+    logger.debug('MasterItemsProvider: MOUNTED');
+    return () => logger.debug('MasterItemsProvider: UNMOUNTED');
+  }, []);
+
+  // Memoize the fetch function to ensure stability for the useApiOnMount hook.
+  const fetchFn = useCallback(() => apiClient.fetchMasterItems(), []);
+
+  const { data, loading, error } = useApiOnMount<MasterGroceryItem[], []>(fetchFn);
 
   const value = useMemo(
     () => ({

src/providers/UserDataProvider.tsx

@@ -1,5 +1,6 @@
 // src/providers/UserDataProvider.tsx
-import React, { useState, useEffect, useMemo, ReactNode } from 'react';
+import { logger } from '../services/logger.client';
+import React, { useState, useEffect, useMemo, ReactNode, useCallback } from 'react';
 import { UserDataContext } from '../contexts/UserDataContext';
 import type { MasterGroceryItem, ShoppingList } from '../types';
 import * as apiClient from '../services/apiClient';
@@ -9,18 +10,25 @@ import { useAuth } from '../hooks/useAuth';
 export const UserDataProvider: React.FC<{ children: ReactNode }> = ({ children }) => {
   const { userProfile } = useAuth();
 
+  // Wrap the API calls in useCallback to prevent unnecessary re-renders.
+  const fetchWatchedItemsFn = useCallback(
+    () => apiClient.fetchWatchedItems(),
+    [],
+  );
+  const fetchShoppingListsFn = useCallback(() => apiClient.fetchShoppingLists(), []);
+
   const {
     data: watchedItemsData,
     loading: isLoadingWatched,
     error: watchedItemsError,
-  } = useApiOnMount<MasterGroceryItem[], []>(() => apiClient.fetchWatchedItems(), [userProfile], {
+  } = useApiOnMount<MasterGroceryItem[], []>(fetchWatchedItemsFn, [userProfile], {
     enabled: !!userProfile,
   });
   const {
     data: shoppingListsData,
-    loading: isLoadingShoppingLists,
+    loading: isLoadingShoppingLists,    
     error: shoppingListsError,
-  } = useApiOnMount<ShoppingList[], []>(() => apiClient.fetchShoppingLists(), [userProfile], {
+  } = useApiOnMount<ShoppingList[], []>(fetchShoppingListsFn, [userProfile], {
     enabled: !!userProfile,
   });
 
@@ -32,7 +40,7 @@ export const UserDataProvider: React.FC<{ children: ReactNode }> = ({ children }
   useEffect(() => {
     // When the user logs out (user becomes null), immediately clear all user-specific data.
     // This also serves to clear out old data when a new user logs in, before their new data arrives.
-    if (!userProfile) {
+ if (!userProfile) {
       setWatchedItems([]);
       setShoppingLists([]);
       return;
@@ -60,7 +68,7 @@ export const UserDataProvider: React.FC<{ children: ReactNode }> = ({ children }
       watchedItemsError,
       shoppingListsError,
     ],
-  );
+ );
 
   return <UserDataContext.Provider value={value}>{children}</UserDataContext.Provider>;
 };

src/routes/auth.routes.test.ts

@@ -165,6 +165,38 @@ describe('Auth Routes (/api/auth)', () => {
       );
     });
 
+    it('should allow registration with an empty string for avatar_url', async () => {
+      // Arrange
+      const email = 'avatar-user@test.com';
+      const mockNewUser = createMockUserProfile({
+        user: { user_id: 'avatar-user-id', email },
+      });
+      mockedAuthService.registerAndLoginUser.mockResolvedValue({
+        newUserProfile: mockNewUser,
+        accessToken: 'avatar-access-token',
+        refreshToken: 'avatar-refresh-token',
+      });
+
+      // Act
+      const response = await supertest(app).post('/api/auth/register').send({
+        email,
+        password: strongPassword,
+        full_name: 'Avatar User',
+        avatar_url: '', // Send an empty string
+      });
+
+      // Assert
+      expect(response.status).toBe(201);
+      expect(response.body.message).toBe('User registered successfully!');
+      expect(mockedAuthService.registerAndLoginUser).toHaveBeenCalledWith(
+        email,
+        strongPassword,
+        'Avatar User',
+        undefined, // The preprocess step in the Zod schema should convert '' to undefined
+        mockLogger,
+      );
+    });
+
     it('should set a refresh token cookie on successful registration', async () => {
       const mockNewUser = createMockUserProfile({
         user: { user_id: 'new-user-id', email: 'cookie@test.com' },

src/routes/auth.routes.ts

@@ -51,7 +51,11 @@ const registerSchema = z.object({
       }),
     // Sanitize optional string inputs.
     full_name: z.string().trim().optional(),
-    avatar_url: z.string().trim().url().optional(),
+    // Allow empty string or valid URL. If empty string is received, convert to undefined.
+    avatar_url: z.preprocess(
+      (val) => (val === '' ? undefined : val),
+      z.string().trim().url().optional(),
+    ),
   }),
 });
 

src/routes/personalization.routes.ts

@@ -19,6 +19,12 @@ router.get(
   validateRequest(emptySchema),
   async (req: Request, res: Response, next: NextFunction) => {
     try {
+      // LOGGING: Track how often this heavy DB call is actually made vs served from cache
+      req.log.info('Fetching master items list from database...');
+
+      // Optimization: This list changes rarely. Instruct clients to cache it for 1 hour (3600s).
+      res.set('Cache-Control', 'public, max-age=3600');
+
       const masterItems = await db.personalizationRepo.getAllMasterItems(req.log);
       res.json(masterItems);
     } catch (error) {

src/routes/user.routes.test.ts

@@ -585,6 +585,27 @@ describe('User Routes (/api/users)', () => {
         expect(response.body).toEqual(updatedProfile);
       });
 
+      it('should allow updating the profile with an empty string for avatar_url', async () => {
+        // Arrange
+        const profileUpdates = { avatar_url: '' };
+        // The service should receive `undefined` after Zod preprocessing
+        const updatedProfile = createMockUserProfile({ ...mockUserProfile, avatar_url: undefined });
+        vi.mocked(db.userRepo.updateUserProfile).mockResolvedValue(updatedProfile);
+
+        // Act
+        const response = await supertest(app).put('/api/users/profile').send(profileUpdates);
+
+        // Assert
+        expect(response.status).toBe(200);
+        expect(response.body).toEqual(updatedProfile);
+        // Verify that the Zod schema preprocessed the empty string to undefined
+        expect(db.userRepo.updateUserProfile).toHaveBeenCalledWith(
+          mockUserProfile.user.user_id,
+          { avatar_url: undefined },
+          expectLogger,
+        );
+      });
+
       it('should return 500 on a generic database error', async () => {
         const dbError = new Error('DB Connection Failed');
         vi.mocked(db.userRepo.updateUserProfile).mockRejectedValue(dbError);

src/routes/user.routes.ts

@@ -26,7 +26,13 @@ const router = express.Router();
 
 const updateProfileSchema = z.object({
   body: z
-    .object({ full_name: z.string().optional(), avatar_url: z.string().url().optional() })
+    .object({
+      full_name: z.string().optional(),
+      avatar_url: z.preprocess(
+        (val) => (val === '' ? undefined : val),
+        z.string().trim().url().optional(),
+      ),
+    })
     .refine((data) => Object.keys(data).length > 0, {
       message: 'At least one field to update must be provided.',
     }),

src/services/aiService.server.test.ts

@@ -6,12 +6,13 @@ import type { FlyerStatus, MasterGroceryItem, UserProfile } from '../types';
 // Import the class, not the singleton instance, so we can instantiate it with mocks.
 import {
   AIService,
-  AiFlyerDataSchema,
   aiService as aiServiceSingleton,
   DuplicateFlyerError,
+  type RawFlyerItem,
 } from './aiService.server';
 import { createMockMasterGroceryItem } from '../tests/utils/mockFactories';
 import { ValidationError } from './db/errors.db';
+import { AiFlyerDataSchema } from '../types/ai';
 
 // Mock the logger to prevent the real pino instance from being created, which causes issues with 'pino-pretty' in tests.
 vi.mock('./logger.server', () => ({
@@ -1058,4 +1059,56 @@ describe('AI Service (Server)', () => {
       expect(aiServiceSingleton).toBeInstanceOf(AIService);
     });
   });
+
+  describe('_normalizeExtractedItems (private method)', () => {
+    it('should correctly normalize items with null or undefined price_in_cents', () => {
+      const rawItems: RawFlyerItem[] = [
+        {
+          item: 'Valid Item',
+          price_display: '$1.99',
+          price_in_cents: 199,
+          quantity: '1',
+          category_name: 'Category A',
+          master_item_id: 1,
+        },
+        {
+          item: 'Item with Null Price',
+          price_display: null,
+          price_in_cents: null, // Test case for null
+          quantity: '1',
+          category_name: 'Category B',
+          master_item_id: 2,
+        },
+        {
+          item: 'Item with Undefined Price',
+          price_display: '$2.99',
+          price_in_cents: undefined, // Test case for undefined
+          quantity: '1',
+          category_name: 'Category C',
+          master_item_id: 3,
+        },
+        {
+          item: null, // Test null item name
+          price_display: undefined, // Test undefined display price
+          price_in_cents: 50,
+          quantity: null, // Test null quantity
+          category_name: undefined, // Test undefined category
+          master_item_id: null, // Test null master_item_id
+        },
+      ];
+
+      // Access the private method for testing
+      const normalized = (aiServiceInstance as any)._normalizeExtractedItems(rawItems);
+
+      expect(normalized).toHaveLength(4);
+      expect(normalized[0].price_in_cents).toBe(199);
+      expect(normalized[1].price_in_cents).toBe(null); // null should remain null
+      expect(normalized[2].price_in_cents).toBe(null); // undefined should become null
+      expect(normalized[3].item).toBe('Unknown Item');
+      expect(normalized[3].quantity).toBe('');
+      expect(normalized[3].category_name).toBe('Other/Miscellaneous');
+      expect(normalized[3].master_item_id).toBeUndefined(); // nullish coalescing to undefined
+    });
+  });
+
 });

src/services/aiService.server.ts

@@ -4,7 +4,6 @@
  * It is intended to be used only by the backend (e.g., server.ts) and should never be imported into client-side code.
  * The `.server.ts` naming convention helps enforce this separation.
  */
-
 import { GoogleGenAI, type GenerateContentResponse, type Content, type Tool } from '@google/genai';
 import fsPromises from 'node:fs/promises';
 import type { Logger } from 'pino';
@@ -26,29 +25,11 @@ import type { Job } from 'bullmq';
 import { createFlyerAndItems } from './db/flyer.db';
 import { generateFlyerIcon } from '../utils/imageProcessor';
 import path from 'path';
-import { ValidationError } from './db/errors.db';
-
-// Helper for consistent required string validation (handles missing/null/empty)
-const requiredString = (message: string) =>
-  z.preprocess((val) => val ?? '', z.string().min(1, message));
-
-// --- Zod Schemas for AI Response Validation (exported for the transformer) ---
-const ExtractedFlyerItemSchema = z.object({
-  item: z.string(),
-  price_display: z.string(),
-  price_in_cents: z.number().nullable(),
-  quantity: z.string(),
-  category_name: z.string(),
-  master_item_id: z.number().nullish(), // .nullish() allows null or undefined
-});
-
-export const AiFlyerDataSchema = z.object({
-  store_name: requiredString('Store name cannot be empty'),
-  valid_from: z.string().nullable(),
-  valid_to: z.string().nullable(),
-  store_address: z.string().nullable(),
-  items: z.array(ExtractedFlyerItemSchema),
-});
+import { ValidationError } from './db/errors.db'; // Keep this import for ValidationError
+import {
+  AiFlyerDataSchema,
+  ExtractedFlyerItemSchema,
+} from '../types/ai'; // Import consolidated schemas
 
 interface FlyerProcessPayload extends Partial<ExtractedCoreData> {
   checksum?: string;
@@ -89,10 +70,10 @@ interface IAiClient {
  * This type is intentionally loose to accommodate potential null/undefined values
  * from the AI before they are cleaned and normalized.
  */
-type RawFlyerItem = {
-  item: string;
+export type RawFlyerItem = {
+  item: string | null;
   price_display: string | null | undefined;
-  price_in_cents: number | null;
+  price_in_cents: number | null | undefined;
   quantity: string | null | undefined;
   category_name: string | null | undefined;
   master_item_id?: number | null | undefined;
@@ -507,7 +488,7 @@ export class AIService {
     userProfileAddress?: string,
     logger: Logger = this.logger,
   ): Promise<{
-    store_name: string;
+    store_name: string | null;
     valid_from: string | null;
     valid_to: string | null;
     store_address: string | null;
@@ -606,6 +587,8 @@ export class AIService {
         item.category_name === null || item.category_name === undefined
           ? 'Other/Miscellaneous'
           : String(item.category_name),
+      // Ensure undefined is converted to null to match the Zod schema.
+      price_in_cents: item.price_in_cents ?? null,
       master_item_id: item.master_item_id ?? undefined,
     }));
   }

src/services/apiClient.ts

@@ -283,7 +283,10 @@ export const fetchFlyerById = (flyerId: number): Promise<Response> =>
  * Fetches all master grocery items from the backend.
  * @returns A promise that resolves to an array of MasterGroceryItem objects.
  */
-export const fetchMasterItems = (): Promise<Response> => publicGet('/personalization/master-items');
+export const fetchMasterItems = (): Promise<Response> => {
+  logger.debug('apiClient: fetchMasterItems called');
+  return publicGet('/personalization/master-items');
+};
 
 /**
  * Fetches all categories from the backend.

src/services/flyerAiProcessor.server.test.ts

@@ -2,7 +2,7 @@
 import { describe, it, expect, vi, beforeEach } from 'vitest';
 import { FlyerAiProcessor } from './flyerAiProcessor.server';
 import { AiDataValidationError } from './processingErrors';
-import { logger } from './logger.server';
+import { logger } from './logger.server'; // Keep this import for the logger instance
 import type { AIService } from './aiService.server';
 import type { PersonalizationRepository } from './db/personalization.db';
 import type { FlyerJobData } from '../types/job-data';
@@ -127,4 +127,98 @@ describe('FlyerAiProcessor', () => {
     expect(result.needsReview).toBe(true);
     expect(logger.warn).toHaveBeenCalledWith(expect.any(Object), expect.stringContaining('contains no items. The flyer will be saved with an item_count of 0. Flagging for review.'));
   });
+
+  describe('Batching Logic', () => {
+    it('should process images in batches and merge the results correctly', async () => {
+      // Arrange
+      const jobData = createMockJobData({});
+      // 5 images, with BATCH_SIZE = 4, should result in 2 batches.
+      const imagePaths = [
+        { path: 'page1.jpg', mimetype: 'image/jpeg' },
+        { path: 'page2.jpg', mimetype: 'image/jpeg' },
+        { path: 'page3.jpg', mimetype: 'image/jpeg' },
+        { path: 'page4.jpg', mimetype: 'image/jpeg' },
+        { path: 'page5.jpg', mimetype: 'image/jpeg' },
+      ];
+
+      const mockAiResponseBatch1 = {
+        store_name: 'Batch 1 Store',
+        valid_from: '2025-01-01',
+        valid_to: '2025-01-07',
+        store_address: '123 Batch St',
+        items: [
+          { item: 'Item A', price_display: '$1', price_in_cents: 100, quantity: '1', category_name: 'Cat A', master_item_id: 1 },
+          { item: 'Item B', price_display: '$2', price_in_cents: 200, quantity: '1', category_name: 'Cat B', master_item_id: 2 },
+        ],
+      };
+
+      const mockAiResponseBatch2 = {
+        store_name: 'Batch 2 Store', // This should be ignored in the merge
+        valid_from: null,
+        valid_to: null,
+        store_address: null,
+        items: [
+          { item: 'Item C', price_display: '$3', price_in_cents: 300, quantity: '1', category_name: 'Cat C', master_item_id: 3 },
+        ],
+      };
+
+      // Mock the AI service to return different results for each batch call
+      vi.mocked(mockAiService.extractCoreDataFromFlyerImage)
+        .mockResolvedValueOnce(mockAiResponseBatch1)
+        .mockResolvedValueOnce(mockAiResponseBatch2);
+
+      // Act
+      const result = await service.extractAndValidateData(imagePaths, jobData, logger);
+
+      // Assert
+      // 1. AI service was called twice (for 2 batches)
+      expect(mockAiService.extractCoreDataFromFlyerImage).toHaveBeenCalledTimes(2);
+
+      // 2. Check the arguments for each call
+      expect(mockAiService.extractCoreDataFromFlyerImage).toHaveBeenNthCalledWith(1, imagePaths.slice(0, 4), [], undefined, undefined, logger);
+      expect(mockAiService.extractCoreDataFromFlyerImage).toHaveBeenNthCalledWith(2, imagePaths.slice(4, 5), [], undefined, undefined, logger);
+
+      // 3. Check the merged data
+      expect(result.data.store_name).toBe('Batch 1 Store'); // Metadata from the first batch
+      expect(result.data.valid_from).toBe('2025-01-01');
+      expect(result.data.valid_to).toBe('2025-01-07');
+      expect(result.data.store_address).toBe('123 Batch St');
+
+      // 4. Check that items from both batches are merged
+      expect(result.data.items).toHaveLength(3);
+      expect(result.data.items).toEqual(expect.arrayContaining([
+        expect.objectContaining({ item: 'Item A' }),
+        expect.objectContaining({ item: 'Item B' }),
+        expect.objectContaining({ item: 'Item C' }),
+      ]));
+
+      // 5. Check that the job is not flagged for review
+      expect(result.needsReview).toBe(false);
+    });
+
+    it('should fill in missing metadata from subsequent batches', async () => {
+      // Arrange
+      const jobData = createMockJobData({});
+      const imagePaths = [
+        { path: 'page1.jpg', mimetype: 'image/jpeg' }, { path: 'page2.jpg', mimetype: 'image/jpeg' }, { path: 'page3.jpg', mimetype: 'image/jpeg' }, { path: 'page4.jpg', mimetype: 'image/jpeg' }, { path: 'page5.jpg', mimetype: 'image/jpeg' },
+      ];
+
+      const mockAiResponseBatch1 = { store_name: null, valid_from: '2025-01-01', valid_to: '2025-01-07', store_address: null, items: [{ item: 'Item A', price_display: '$1', price_in_cents: 100, quantity: '1', category_name: 'Cat A', master_item_id: 1 }] };
+      const mockAiResponseBatch2 = { store_name: 'Batch 2 Store', valid_from: '2025-01-02', valid_to: null, store_address: '456 Subsequent St', items: [{ item: 'Item C', price_display: '$3', price_in_cents: 300, quantity: '1', category_name: 'Cat C', master_item_id: 3 }] };
+
+      vi.mocked(mockAiService.extractCoreDataFromFlyerImage)
+        .mockResolvedValueOnce(mockAiResponseBatch1)
+        .mockResolvedValueOnce(mockAiResponseBatch2);
+
+      // Act
+      const result = await service.extractAndValidateData(imagePaths, jobData, logger);
+
+      // Assert
+      expect(result.data.store_name).toBe('Batch 2 Store'); // Filled from batch 2
+      expect(result.data.valid_from).toBe('2025-01-01'); // Kept from batch 1
+      expect(result.data.valid_to).toBe('2025-01-07'); // Kept from batch 1
+      expect(result.data.store_address).toBe('456 Subsequent St'); // Filled from batch 2
+      expect(result.data.items).toHaveLength(2);
+    });
+  });
 });

src/services/flyerAiProcessor.server.ts

@@ -5,28 +5,11 @@ import type { AIService } from './aiService.server';
 import type { PersonalizationRepository } from './db/personalization.db';
 import { AiDataValidationError } from './processingErrors';
 import type { FlyerJobData } from '../types/job-data';
-
-// Helper for consistent required string validation (handles missing/null/empty)
-const requiredString = (message: string) =>
-  z.preprocess((val) => val ?? '', z.string().min(1, message));
-
-// --- Zod Schemas for AI Response Validation ---
-const ExtractedFlyerItemSchema = z.object({
-  item: z.string().nullable(),
-  price_display: z.string().nullable(),
-  price_in_cents: z.number().nullable(),
-  quantity: z.string().nullable(),
-  category_name: z.string().nullable(),
-  master_item_id: z.number().nullish(),
-});
-
-export const AiFlyerDataSchema = z.object({
-  store_name: z.string().nullable(),
-  valid_from: z.string().nullable(),
-  valid_to: z.string().nullable(),
-  store_address: z.string().nullable(),
-  items: z.array(ExtractedFlyerItemSchema),
-});
+import {
+  AiFlyerDataSchema,
+  ExtractedFlyerItemSchema,
+  requiredString,
+} from '../types/ai'; // Import consolidated schemas and helper
 
 export type ValidatedAiDataType = z.infer<typeof AiFlyerDataSchema>;
 
@@ -94,19 +77,64 @@ export class FlyerAiProcessor {
     jobData: FlyerJobData,
     logger: Logger,
   ): Promise<AiProcessorResult> {
-    logger.info(`Starting AI data extraction.`);
+    logger.info(`Starting AI data extraction for ${imagePaths.length} pages.`);
     const { submitterIp, userProfileAddress } = jobData;
     const masterItems = await this.personalizationRepo.getAllMasterItems(logger);
     logger.debug(`Retrieved ${masterItems.length} master items for AI matching.`);
 
-    const extractedData = await this.ai.extractCoreDataFromFlyerImage(
-      imagePaths,
-      masterItems,
-      submitterIp,
-      userProfileAddress,
-      logger,
-    );
+    // BATCHING LOGIC: Process images in chunks to avoid hitting AI payload/token limits.
+    const BATCH_SIZE = 4;
+    const batches = [];
+    for (let i = 0; i < imagePaths.length; i += BATCH_SIZE) {
+      batches.push(imagePaths.slice(i, i + BATCH_SIZE));
+    }
 
-    return this._validateAiData(extractedData, logger);
+    // Initialize container for merged data
+    const mergedData: ValidatedAiDataType = {
+      store_name: null,
+      valid_from: null,
+      valid_to: null,
+      store_address: null,
+      items: [],
+    };
+
+    logger.info(`Processing ${imagePaths.length} pages in ${batches.length} batches (Batch Size: ${BATCH_SIZE}).`);
+
+    for (const [index, batch] of batches.entries()) {
+      logger.info(`Processing batch ${index + 1}/${batches.length} (${batch.length} pages)...`);
+
+      // The AI service handles rate limiting internally (e.g., max 5 RPM).
+      // Processing these sequentially ensures we respect that limit.
+      const batchResult = await this.ai.extractCoreDataFromFlyerImage(
+        batch,
+        masterItems,
+        submitterIp,
+        userProfileAddress,
+        logger,
+      );
+
+      // MERGE LOGIC:
+      // 1. Metadata (Store Name, Dates): Prioritize the first batch (usually the cover page).
+      //    If subsequent batches have data and the current is null, fill it in.
+      if (index === 0) {
+        mergedData.store_name = batchResult.store_name;
+        mergedData.valid_from = batchResult.valid_from;
+        mergedData.valid_to = batchResult.valid_to;
+        mergedData.store_address = batchResult.store_address;
+      } else {
+        if (!mergedData.store_name && batchResult.store_name) mergedData.store_name = batchResult.store_name;
+        if (!mergedData.valid_from && batchResult.valid_from) mergedData.valid_from = batchResult.valid_from;
+        if (!mergedData.valid_to && batchResult.valid_to) mergedData.valid_to = batchResult.valid_to;
+        if (!mergedData.store_address && batchResult.store_address) mergedData.store_address = batchResult.store_address;
+      }
+
+      // 2. Items: Append all found items to the master list.
+      mergedData.items.push(...batchResult.items);
+    }
+
+    logger.info(`Batch processing complete. Total items extracted: ${mergedData.items.length}`);
+
+    // Validate the final merged dataset
+    return this._validateAiData(mergedData, logger);
   }
 }

src/services/flyerDataTransformer.ts

@@ -2,8 +2,9 @@
 import path from 'path';
 import type { z } from 'zod';
 import type { Logger } from 'pino';
-import type { FlyerInsert, FlyerItemInsert, FlyerStatus } from '../types';
-import type { AiFlyerDataSchema, AiProcessorResult } from './flyerAiProcessor.server';
+import type { FlyerInsert, FlyerItemInsert } from '../types';
+import type { AiProcessorResult } from './flyerAiProcessor.server'; // Keep this import for AiProcessorResult
+import { AiFlyerDataSchema } from '../types/ai'; // Import consolidated schema
 import { generateFlyerIcon } from '../utils/imageProcessor';
 
 /**

src/services/flyerProcessingService.server.ts

@@ -133,6 +133,12 @@ export class FlyerProcessingService {
       return { flyerId: flyer.flyer_id };
     } catch (error) {
       logger.warn('Job failed. Temporary files will NOT be cleaned up to allow for manual inspection.');
+      // Add detailed logging of the raw error object
+      if (error instanceof Error) {
+        logger.error({ err: error, stack: error.stack }, 'Raw error object in processJob catch block');
+      } else {
+        logger.error({ error }, 'Raw non-Error object in processJob catch block');
+      }
       // This private method handles error reporting and re-throwing.
       await this._reportErrorAndThrow(error, job, logger, stages);
       // This line is technically unreachable because the above method always throws,

src/services/workers.server.ts

@@ -1,3 +1,4 @@
+// src/services/workers.server.ts
 import { Worker, Job, UnrecoverableError } from 'bullmq';
 import fsPromises from 'node:fs/promises';
 import { exec } from 'child_process';

src/tests/integration/auth.integration.test.ts

@@ -21,16 +21,18 @@ const request = supertest(app);
 describe('Authentication API Integration', () => {
   let testUserEmail: string;
   let testUser: UserProfile;
+  const createdUserIds: string[] = [];
 
   beforeAll(async () => {
     // Use a unique email for this test suite to prevent collisions with other tests.
     const email = `auth-integration-test-${Date.now()}@example.com`;
     ({ user: testUser } = await createAndLoginUser({ email, fullName: 'Auth Test User', request }));
     testUserEmail = testUser.user.email;
+    createdUserIds.push(testUser.user.user_id);
   });
 
   afterAll(async () => {
-    await cleanupDb({ userIds: testUser ? [testUser.user.user_id] : [] });
+    await cleanupDb({ userIds: createdUserIds });
   });
 
   // This test migrates the logic from the old DevTestRunner.tsx component.
@@ -83,6 +85,38 @@ describe('Authentication API Integration', () => {
     expect(errorData.message).toBe('Incorrect email or password.');
   });
 
+  it('should allow registration with an empty string for avatar_url and save it as null', async () => {
+    // Arrange: Define user data with an empty avatar_url.
+    const email = `empty-avatar-user-${Date.now()}@example.com`;
+    const userData = {
+      email,
+      password: TEST_PASSWORD,
+      full_name: 'Empty Avatar',
+      avatar_url: '',
+    };
+
+    // Act: Register the new user.
+    const registerResponse = await request.post('/api/auth/register').send(userData);
+
+    // Assert 1: Check that the registration was successful and the returned profile is correct.
+    expect(registerResponse.status).toBe(201);
+    const registeredProfile = registerResponse.body.userprofile;
+    const registeredToken = registerResponse.body.token;
+    expect(registeredProfile.user.email).toBe(email);
+    expect(registeredProfile.avatar_url).toBeNull(); // The API should return null for the avatar_url.
+
+    // Add the newly created user's ID to the array for cleanup in afterAll.
+    createdUserIds.push(registeredProfile.user.user_id);
+
+    // Assert 2 (Verification): Fetch the profile using the new token to confirm the value in the DB is null.
+    const profileResponse = await request
+      .get('/api/users/profile')
+      .set('Authorization', `Bearer ${registeredToken}`);
+
+    expect(profileResponse.status).toBe(200);
+    expect(profileResponse.body.avatar_url).toBeNull();
+  });
+
   it('should successfully refresh an access token using a refresh token cookie', async () => {
     // Arrange: Log in to get a fresh, valid refresh token cookie for this specific test.
     // This ensures the test is self-contained and not affected by other tests.

src/tests/integration/flyer-processing.integration.test.ts

@@ -82,6 +82,7 @@ describe('Flyer Processing Background Job Integration Test', () => {
     let jobStatus;
     const maxRetries = 30; // Poll for up to 90 seconds (30 * 3s)
     for (let i = 0; i < maxRetries; i++) {
+      console.log(`Polling attempt ${i + 1}...`);
       await new Promise((resolve) => setTimeout(resolve, 3000)); // Wait 3 seconds between polls
       const statusReq = request.get(`/api/ai/jobs/${jobId}/status`);
       if (token) {
@@ -89,6 +90,7 @@ describe('Flyer Processing Background Job Integration Test', () => {
       }
       const statusResponse = await statusReq;
       jobStatus = statusResponse.body;
+      console.log(`Job status: ${JSON.stringify(jobStatus)}`);
       if (jobStatus.state === 'completed' || jobStatus.state === 'failed') {
         break;
       }

src/tests/integration/public.routes.integration.test.ts

@@ -38,6 +38,7 @@ describe('Public API Routes Integration Tests', () => {
       email: userEmail,
       password: 'a-Very-Strong-Password-123!',
       fullName: 'Public Routes Test User',
+      request,
     });
     testUser = createdUser;
 

src/tests/integration/user.integration.test.ts

@@ -75,6 +75,32 @@ describe('User API Routes Integration Tests', () => {
     expect(refetchedProfile.full_name).toBe('Updated Test User');
   });
 
+  it('should allow updating the profile with an empty string for avatar_url', async () => {
+    // Arrange: Define the profile updates.
+    const profileUpdates = {
+      full_name: 'Empty Avatar User',
+      avatar_url: '',
+    };
+
+    // Act: Call the update endpoint with the new data and the auth token.
+    const response = await request
+      .put('/api/users/profile')
+      .set('Authorization', `Bearer ${authToken}`)
+      .send(profileUpdates);
+    const updatedProfile = response.body;
+
+    // Assert: Check that the returned profile reflects the changes.
+    expect(response.status).toBe(200);
+    expect(updatedProfile.full_name).toBe('Empty Avatar User');
+    expect(updatedProfile.avatar_url).toBeNull();
+
+    // Also, fetch the profile again to ensure the change was persisted in the database as NULL.
+    const refetchResponse = await request
+      .get('/api/users/profile')
+      .set('Authorization', `Bearer ${authToken}`);
+    expect(refetchResponse.body.avatar_url).toBeNull();
+  });
+
   it('should update user preferences via PUT /api/users/profile/preferences', async () => {
     // Arrange: Define the preference updates.
     const preferenceUpdates = {

src/tests/utils/cleanup.ts

@@ -43,8 +43,6 @@ export const cleanupDb = async (ids: TestResourceIds) => {
       await pool.query('DELETE FROM public.user_watched_items WHERE user_id = ANY($1::uuid[])', [userIds]);
       await pool.query('DELETE FROM public.user_achievements WHERE user_id = ANY($1::uuid[])', [userIds]);
       await pool.query('DELETE FROM public.activity_log WHERE user_id = ANY($1::uuid[])', [userIds]);
-      await pool.query('DELETE FROM public.user_refresh_tokens WHERE user_id = ANY($1::uuid[])', [userIds]);
-      await pool.query('DELETE FROM public.password_reset_tokens WHERE user_id = ANY($1::uuid[])', [userIds]);
     }
 
     // --- Stage 2: Delete parent records that other things depend on ---

src/types/ai.ts

@@ -0,0 +1,29 @@
+// src/types/ai.ts
+
+import { z } from 'zod';
+
+// Helper for consistent required string validation (handles missing/null/empty)
+// This is moved here as it's directly related to the schemas.
+export const requiredString = (message: string) =>
+  z.preprocess((val) => val ?? '', z.string().min(1, message));
+
+// --- Zod Schemas for AI Response Validation ---
+// These schemas define the expected structure of data returned by the AI.
+// They are used for validation and type inference across multiple services.
+
+export const ExtractedFlyerItemSchema = z.object({
+  item: z.string().nullable(),
+  price_display: z.string().nullable(),
+  price_in_cents: z.number().nullable(),
+  quantity: z.string().nullable(),
+  category_name: z.string().nullable(),
+  master_item_id: z.number().nullish(), // .nullish() allows null or undefined
+});
+
+export const AiFlyerDataSchema = z.object({
+  store_name: z.string().nullable(),
+  valid_from: z.string().nullable(),
+  valid_to: z.string().nullable(),
+  store_address: z.string().nullable(),
+  items: z.array(ExtractedFlyerItemSchema),
+});