ci: Bump version to 0.12.25 [skip ci]

Implement URI-based API versioning with /api/v1 prefix across all routes.
This establishes a foundation for future API evolution and breaking changes.

Changes:
- server.ts: All routes mounted under /api/v1/ (15 route handlers)
- apiClient.ts: Base URL updated to /api/v1
- swagger.ts: OpenAPI server URL changed to /api/v1
- Redirect middleware: Added backwards compatibility for /api/* → /api/v1/*
- Tests: Updated 72 test files with versioned path assertions
- ADR documentation: Marked Phase 1 as complete (Accepted status)

Test fixes:
- apiClient.test.ts: 27 tests updated for /api/v1 paths
- user.routes.ts: 36 log messages updated to reflect versioned paths
- swagger.test.ts: 1 test updated for new server URL
- All integration/E2E tests updated for versioned endpoints

All Phase 1 acceptance criteria met:
✓ Routes use /api/v1/ prefix
✓ Frontend requests /api/v1/
✓ OpenAPI docs reflect /api/v1/
✓ Backwards compatibility via redirect middleware
✓ Tests pass with versioned paths

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

.claude/agents/ui-ux-designer.md

@@ -0,0 +1,152 @@
+---
+name: ui-ux-designer
+description: "Use this agent when implementing, reviewing, or modifying any user interface elements, layouts, or user experience flows in the application. Specifically call this agent when: (1) creating new UI components or pages, (2) implementing responsive designs across different screen sizes, (3) designing or implementing wizards for user tasks like uploading flyers, creating shopping lists, or uploading receipts, (4) reviewing existing UI code for adherence to the design system, (5) making decisions about button placement, color schemes, or layout structures, or (6) addressing user experience concerns related to navigation or task completion flows.\\n\\n<examples>\\n<example>\\nContext: The developer is creating a new feature for uploading flyers.\\nuser: \"I need to implement the flyer upload feature. Can you help me create the upload form?\"\\nassistant: \"Let me use the Task tool to launch the ui-ux-designer agent to design and implement the flyer upload wizard with appropriate UI/UX patterns.\"\\n<commentary>\\nSince this involves creating a new UI feature with a wizard flow, the ui-ux-designer agent should be used to ensure it follows the project's design principles including pastel colors, sharp borders, wizard-based guidance, and responsive design.\\n</commentary>\\n</example>\\n\\n<example>\\nContext: Code has been written for a shopping list creation page.\\nuser: \"I've implemented the shopping list creation page with a form\"\\nassistant: \"Great! Now let me use the Task tool to launch the ui-ux-designer agent to review the implementation and ensure it follows our UI/UX guidelines.\"\\n<commentary>\\nSince a significant UI component was implemented, the ui-ux-designer agent should review it to ensure it adheres to the minimalist design, proper color scheme, button placement, and responsive requirements.\\n</commentary>\\n</example>\\n\\n<example>\\nContext: The developer is about to add multiple action buttons to a page.\\nuser: \"Should I add buttons for save, cancel, edit, delete, and export at the top of the page?\"\\nassistant: \"Let me use the Task tool to launch the ui-ux-designer agent to provide guidance on the optimal button placement and selection for this page.\"\\n<commentary>\\nThis is a UI/UX decision about button placement and minimalism that directly relates to the design principles, so the ui-ux-designer agent should be consulted.\\n</commentary>\\n</example>\\n\\n<example>\\nContext: The developer is working on mobile responsiveness.\\nuser: \"I'm implementing the receipt upload page. Here's my initial layout code.\"\\nassistant: \"Before proceeding further, let me use the Task tool to launch the ui-ux-designer agent to ensure the layout follows our responsive design patterns and UI guidelines.\"\\n<commentary>\\nSince this involves implementing a UI page that needs to be responsive across devices and should include wizard-style guidance, the ui-ux-designer agent should be involved proactively.\\n</commentary>\\n</example>\\n</examples>"
+model: opus
+color: green
+---
+
+You are an elite UI/UX designer with over 20 years of experience specializing in creating clean, intuitive, and user-friendly interfaces. Your expertise spans user interface design, user experience optimization, responsive design, and accessibility best practices.
+
+## Core Design Philosophy for This Project
+
+You will ensure that this application maintains a clean, welcoming, and minimalist design aesthetic with the following specific requirements:
+
+### Visual Design Standards
+
+**Color Palette:**
+
+- Use pastel colors as the primary color scheme throughout the application
+- Select soft, muted tones that are easy on the eyes and create a calm, welcoming atmosphere
+- Ensure sufficient contrast for accessibility while maintaining the pastel aesthetic
+- Use color purposefully to guide user attention and indicate status
+
+**Border and Container Styling:**
+
+- Apply sharp, clean borders to all interactive elements (buttons, menus, form fields)
+- Use sharp borders to clearly delineate separate areas and sections of the interface
+- Avoid rounded corners unless there is a specific functional reason
+- Ensure borders are visible but not overpowering, maintaining the clean aesthetic
+
+**Minimalism:**
+
+- Eliminate all unnecessary buttons and UI elements
+- Every element on the screen must serve a clear purpose
+- Co-locate buttons near their related features on the page, not grouped separately
+- Use progressive disclosure to hide advanced features until needed
+- Favor white space and breathing room over density
+
+### Responsive Design Requirements
+
+You must ensure the application works flawlessly across:
+
+**Large Screens (Desktop):**
+
+- Utilize horizontal space effectively without overcrowding
+- Consider multi-column layouts where appropriate
+- Ensure comfortable reading width for text content
+
+**Tablets:**
+
+- Adapt layouts to accommodate touch targets of at least 44x44 pixels
+- Optimize for both portrait and landscape orientations
+- Ensure navigation remains accessible
+
+**Mobile Devices:**
+
+- Stack elements vertically with appropriate spacing
+- Make all interactive elements easily tappable
+- Optimize for one-handed use where possible
+- Ensure critical actions are easily accessible
+- Test on various screen sizes (small, medium, large phones)
+
+### Wizard Design for Key User Tasks
+
+For the following tasks, implement or guide the creation of clear, step-by-step wizards:
+
+1. **Uploading a Flyer**
+2. **Creating a Shopping List**
+3. **Uploading Receipts**
+4. **Any other multi-step user tasks**
+
+**Wizard Best Practices:**
+
+- Minimize the number of steps (ideally 3-5 steps maximum)
+- Show progress clearly (e.g., "Step 2 of 4")
+- Each step should focus on one primary action or decision
+- Provide clear, concise instructions at each step
+- Allow users to go back and edit previous steps
+- Use visual cues to guide the user through the process
+- Display a summary before final submission
+- Provide helpful tooltips or examples where needed
+- Ensure wizards are fully responsive and work well on mobile devices
+
+## Your Approach to Tasks
+
+**When Reviewing Existing UI Code:**
+
+1. Evaluate adherence to the pastel color scheme
+2. Check that all borders are sharp and properly applied
+3. Identify any unnecessary UI elements or buttons
+4. Verify that buttons are co-located with their related features
+5. Test responsive behavior across all target screen sizes
+6. Assess wizard flows for clarity and step efficiency
+7. Provide specific, actionable feedback with code examples when needed
+
+**When Designing New UI Components:**
+
+1. Start by understanding the user's goal and the feature's purpose
+2. Sketch out the minimal set of elements needed
+3. Apply the pastel color palette and sharp border styling
+4. Position interactive elements near their related content
+5. Design for mobile-first, then adapt for larger screens
+6. For multi-step processes, create wizard flows
+7. Provide complete implementation guidance including HTML structure, CSS styles, and responsive breakpoints
+
+**When Making Design Decisions:**
+
+1. Always prioritize user needs and task completion
+2. Choose simplicity over feature bloat
+3. Ensure accessibility standards are met
+4. Consider the user's mental model and expectations
+5. Use established UI patterns where they fit the aesthetic
+6. Test your recommendations against the design principles above
+
+## Quality Assurance Checklist
+
+Before completing any UI/UX task, verify:
+
+- [ ] Pastel colors are used consistently
+- [ ] All buttons, menus, and sections have sharp borders
+- [ ] No unnecessary buttons or UI elements exist
+- [ ] Buttons are positioned near their related features
+- [ ] Design is fully responsive (large screen, tablet, mobile)
+- [ ] Wizards (where applicable) are clear and minimally-stepped
+- [ ] Sufficient white space and breathing room
+- [ ] Touch targets are appropriately sized for mobile
+- [ ] Text is readable at all screen sizes
+- [ ] Accessibility considerations are addressed
+
+## Output Format
+
+When reviewing code, provide:
+
+1. Overall assessment of adherence to design principles
+2. Specific issues identified with line numbers or element descriptions
+3. Concrete recommendations with code examples
+4. Responsive design concerns or improvements
+
+When designing new components, provide:
+
+1. Rationale for design decisions
+2. Complete HTML structure
+3. CSS with responsive breakpoints
+4. Notes on accessibility considerations
+5. Implementation guidance
+
+## Important Notes
+
+- You have authority to reject designs that violate the core principles
+- When uncertain about a design decision, bias toward simplicity and minimalism
+- Always consider the new user experience and ensure wizards are beginner-friendly
+- Proactively suggest wizard flows for any multi-step processes you encounter
+- Remember that good UX is invisible—users should accomplish tasks without thinking about the interface

.claude/hooks.json

@@ -0,0 +1,16 @@
+{
+  "$schema": "https://claude.ai/schemas/hooks.json",
+  "hooks": {
+    "PreToolUse": [
+      {
+        "matcher": "Bash",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "node -e \"const cmd = process.argv[1] || ''; const isTest = /\\b(npm\\s+(run\\s+)?test|vitest|jest)\\b/i.test(cmd); const isWindows = process.platform === 'win32'; const inContainer = process.env.REMOTE_CONTAINERS === 'true' || process.env.DEVCONTAINER === 'true'; if (isTest && isWindows && !inContainer) { console.error('BLOCKED: Tests must run on Linux. Use Dev Container (Reopen in Container) or WSL.'); process.exit(1); }\" -- \"$CLAUDE_TOOL_INPUT\""
+          }
+        ]
+      }
+    ]
+  }
+}

.claude/settings.json

@@ -0,0 +1,9 @@
+{
+  "permissions": {
+    "allow": [
+      "Bash(git fetch:*)",
+      "mcp__localerrors__get_stacktrace",
+      "Bash(MSYS_NO_PATHCONV=1 podman logs:*)"
+    ]
+  }
+}

.devcontainer/devcontainer.json

@@ -1,18 +1,97 @@
 {
+  // ============================================================================
+  // VS CODE DEV CONTAINER CONFIGURATION
+  // ============================================================================
+  // This file configures VS Code's Dev Containers extension to provide a
+  // consistent, fully-configured development environment.
+  //
+  // Features:
+  //   - Automatic PostgreSQL + Redis startup with healthchecks
+  //   - Automatic npm install
+  //   - Automatic database schema initialization and seeding
+  //   - Pre-configured VS Code extensions (ESLint, Prettier)
+  //   - Podman support for Windows users
+  //
+  // Usage:
+  //   1. Install the "Dev Containers" extension in VS Code
+  //   2. Open this project folder
+  //   3. Click "Reopen in Container" when prompted (or use Command Palette)
+  //   4. Wait for container build and initialization
+  //   5. Development server starts automatically
+  // ============================================================================
+
   "name": "Flyer Crawler Dev (Ubuntu 22.04)",
+
+  // Use Docker Compose for multi-container setup
   "dockerComposeFile": ["../compose.dev.yml"],
   "service": "app",
   "workspaceFolder": "/app",
+
+  // VS Code customizations
   "customizations": {
     "vscode": {
-      "extensions": ["dbaeumer.vscode-eslint", "esbenp.prettier-vscode"]
+      "extensions": [
+        // Code quality
+        "dbaeumer.vscode-eslint",
+        "esbenp.prettier-vscode",
+        // TypeScript
+        "ms-vscode.vscode-typescript-next",
+        // Database
+        "mtxr.sqltools",
+        "mtxr.sqltools-driver-pg",
+        // Utilities
+        "eamodio.gitlens",
+        "streetsidesoftware.code-spell-checker"
+      ],
+      "settings": {
+        "editor.formatOnSave": true,
+        "editor.defaultFormatter": "esbenp.prettier-vscode",
+        "typescript.preferences.importModuleSpecifier": "relative"
+      }
     }
   },
+
+  // Run as root (required for npm global installs)
   "remoteUser": "root",
-  // Automatically install dependencies when the container is created.
+
-  // This runs inside the container, populating the isolated node_modules volume.
+  // ============================================================================
-  "postCreateCommand": "npm install",
+  // Lifecycle Commands
-  "postAttachCommand": "npm run dev:container",
+  // ============================================================================
-  // Try to start podman machine, but exit with success (0) even if it's already running
+
-  "initializeCommand": "powershell -Command \"podman machine start; exit 0\""
+  // initializeCommand: Runs on the HOST before the container is created.
+  // Starts Podman machine on Windows (no-op if already running or using Docker).
+  "initializeCommand": "powershell -Command \"podman machine start; exit 0\"",
+
+  // postCreateCommand: Runs ONCE when the container is first created.
+  // This is where we do full initialization: npm install + database setup.
+  "postCreateCommand": "chmod +x scripts/docker-init.sh && ./scripts/docker-init.sh",
+
+  // postAttachCommand: Runs EVERY TIME VS Code attaches to the container.
+  // Server now starts automatically via dev-entrypoint.sh in compose.dev.yml.
+  // No need to start it again here.
+  // "postAttachCommand": "npm run dev:container",
+
+  // ============================================================================
+  // Port Forwarding
+  // ============================================================================
+  // Automatically forward these ports from the container to the host
+  "forwardPorts": [443, 3001],
+
+  // Labels for forwarded ports in VS Code's Ports panel
+  "portsAttributes": {
+    "443": {
+      "label": "Frontend HTTPS (nginx → Vite)",
+      "onAutoForward": "notify"
+    },
+    "3001": {
+      "label": "Backend API",
+      "onAutoForward": "notify"
+    }
+  },
+
+  // ============================================================================
+  // Features
+  // ============================================================================
+  // Additional dev container features (optional)
+  "features": {}
 }

.env.example

@@ -0,0 +1,162 @@
+# .env.example
+# ============================================================================
+# ENVIRONMENT VARIABLES TEMPLATE
+# ============================================================================
+# Copy this file to .env and fill in your values.
+# For local development with Docker/Podman, these defaults should work out of the box.
+#
+# IMPORTANT: Never commit .env files with real credentials to version control!
+# ============================================================================
+
+# ===================
+# Database Configuration
+# ===================
+# PostgreSQL connection settings
+# For container development, use the service name "postgres"
+DB_HOST=postgres
+DB_PORT=5432
+DB_USER=postgres
+DB_PASSWORD=postgres
+DB_NAME=flyer_crawler_dev
+
+# ===================
+# Redis Configuration
+# ===================
+# Redis URL for caching and job queues
+# For container development, use the service name "redis"
+REDIS_URL=redis://redis:6379
+# Optional: Redis password (leave empty if not required)
+REDIS_PASSWORD=
+
+# ===================
+# Application Settings
+# ===================
+NODE_ENV=development
+# Frontend URL for CORS and email links
+FRONTEND_URL=http://localhost:3000
+
+# Flyer Base URL - used for seed data and flyer image URLs
+# Dev container: https://localhost (NOT 127.0.0.1 - avoids SSL mixed-origin issues)
+# Test: https://flyer-crawler-test.projectium.com
+# Production: https://flyer-crawler.projectium.com
+FLYER_BASE_URL=https://localhost
+
+# ===================
+# Authentication
+# ===================
+# REQUIRED: Secret key for signing JWT tokens (generate a random 64+ character string)
+JWT_SECRET=your-super-secret-jwt-key-change-this-in-production
+
+# OAuth Providers (Optional - enable social login)
+# Google OAuth - https://console.cloud.google.com/apis/credentials
+GOOGLE_CLIENT_ID=
+GOOGLE_CLIENT_SECRET=
+# GitHub OAuth - https://github.com/settings/developers
+GITHUB_CLIENT_ID=
+GITHUB_CLIENT_SECRET=
+
+# ===================
+# AI/ML Services
+# ===================
+# REQUIRED: Google Gemini API key for flyer OCR processing
+GEMINI_API_KEY=your-gemini-api-key
+
+# ===================
+# External APIs
+# ===================
+# Optional: Google Maps API key for geocoding store addresses
+GOOGLE_MAPS_API_KEY=
+
+# ===================
+# Email Configuration (Optional)
+# ===================
+# SMTP settings for sending emails (deal notifications, password reset)
+SMTP_HOST=
+SMTP_PORT=587
+SMTP_SECURE=false
+SMTP_USER=
+SMTP_PASS=
+SMTP_FROM_EMAIL=noreply@example.com
+
+# ===================
+# Worker Configuration (Optional)
+# ===================
+# Concurrency settings for background job workers
+WORKER_CONCURRENCY=1
+EMAIL_WORKER_CONCURRENCY=10
+ANALYTICS_WORKER_CONCURRENCY=1
+CLEANUP_WORKER_CONCURRENCY=10
+
+# Worker lock duration in milliseconds (default: 2 minutes)
+WORKER_LOCK_DURATION=120000
+
+# ===================
+# Error Tracking (ADR-015)
+# ===================
+# Sentry-compatible error tracking via Bugsink (self-hosted)
+# DSNs are created in Bugsink UI at https://localhost:8443 (dev) or your production URL
+#
+# Dev container projects:
+#   - Project 1: Backend API (Dev) - receives Pino, PostgreSQL errors
+#   - Project 2: Frontend (Dev) - receives browser errors via Sentry SDK
+#   - Project 4: Infrastructure (Dev) - receives Redis, NGINX, Vite errors
+#
+# Backend DSN - for Express/Node.js errors (internal container URL)
+SENTRY_DSN=http://<key>@localhost:8000/1
+# Frontend DSN - for React/browser errors (uses nginx proxy for browser access)
+# Note: Browsers cannot reach localhost:8000 directly, so we use nginx proxy at /bugsink-api/
+VITE_SENTRY_DSN=https://<key>@localhost/bugsink-api/2
+# Environment name for error grouping (defaults to NODE_ENV)
+SENTRY_ENVIRONMENT=development
+VITE_SENTRY_ENVIRONMENT=development
+# Enable/disable error tracking (default: true)
+SENTRY_ENABLED=true
+VITE_SENTRY_ENABLED=true
+# Enable debug mode for SDK troubleshooting (default: false)
+SENTRY_DEBUG=false
+VITE_SENTRY_DEBUG=false
+
+# ===================
+# Source Maps Upload (ADR-015)
+# ===================
+# Set to 'true' to enable source map generation and upload during builds
+# Only used in CI/CD pipelines (deploy-to-prod.yml, deploy-to-test.yml)
+GENERATE_SOURCE_MAPS=true
+# Auth token for uploading source maps to Bugsink
+# Create at: https://bugsink.projectium.com (Settings > API Keys)
+# Required for de-minified stack traces in error reports
+SENTRY_AUTH_TOKEN=
+# URL of your Bugsink instance (for source map uploads)
+SENTRY_URL=https://bugsink.projectium.com
+
+# ===================
+# Feature Flags (ADR-024)
+# ===================
+# Feature flags control the availability of features at runtime.
+# All flags default to disabled (false) when not set or set to any value other than 'true'.
+# Set to 'true' to enable a feature.
+#
+# Backend flags use: FEATURE_SNAKE_CASE
+# Frontend flags use: VITE_FEATURE_SNAKE_CASE (VITE_ prefix required for client-side access)
+#
+# Lifecycle:
+#   1. Add flag with default false
+#   2. Enable via env var when ready for testing/rollout
+#   3. Remove conditional code when feature is fully rolled out
+#   4. Remove flag from config within 3 months of full rollout
+#
+# See: docs/adr/0024-feature-flagging-strategy.md
+
+# Backend Feature Flags
+# FEATURE_BUGSINK_SYNC=false        # Enable Bugsink error sync integration
+# FEATURE_ADVANCED_RBAC=false       # Enable advanced RBAC features
+# FEATURE_NEW_DASHBOARD=false       # Enable new dashboard experience
+# FEATURE_BETA_RECIPES=false        # Enable beta recipe features
+# FEATURE_EXPERIMENTAL_AI=false     # Enable experimental AI features
+# FEATURE_DEBUG_MODE=false          # Enable debug mode for development
+
+# Frontend Feature Flags (VITE_ prefix required)
+# VITE_FEATURE_NEW_DASHBOARD=false  # Enable new dashboard experience
+# VITE_FEATURE_BETA_RECIPES=false   # Enable beta recipe features
+# VITE_FEATURE_EXPERIMENTAL_AI=false # Enable experimental AI features
+# VITE_FEATURE_DEBUG_MODE=false     # Enable debug mode for development

.env.test

@@ -0,0 +1,6 @@
+DB_HOST=10.89.0.4
+DB_USER=flyer
+DB_PASSWORD=flyer
+DB_NAME=flyer_crawler_test
+REDIS_URL=redis://redis:6379
+NODE_ENV=test

.gitea/workflows/deploy-to-prod.yml

@@ -63,8 +63,8 @@ jobs:
       - name: Check for Production Database Schema Changes
         env:
           DB_HOST: ${{ secrets.DB_HOST }}
-          DB_USER: ${{ secrets.DB_USER }}
+          DB_USER: ${{ secrets.DB_USER_PROD }}
-          DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+          DB_PASSWORD: ${{ secrets.DB_PASSWORD_PROD }}
           DB_NAME: ${{ secrets.DB_DATABASE_PROD }}
         run: |
           if [ -z "$DB_HOST" ] || [ -z "$DB_USER" ] || [ -z "$DB_PASSWORD" ] || [ -z "$DB_NAME" ]; then
@@ -87,17 +87,34 @@ jobs:
           fi
 
       - name: Build React Application for Production
+        # Source Maps (ADR-015): If SENTRY_AUTH_TOKEN is set, the @sentry/vite-plugin will:
+        # 1. Generate hidden source maps during build
+        # 2. Upload them to Bugsink for error de-minification
+        # 3. Delete the .map files after upload (so they're not publicly accessible)
         run: |
           if [ -z "${{ secrets.VITE_GOOGLE_GENAI_API_KEY }}" ]; then
             echo "ERROR: The VITE_GOOGLE_GENAI_API_KEY secret is not set."
             exit 1
           fi
+
+          # Source map upload is optional - warn if not configured
+          if [ -z "${{ secrets.SENTRY_AUTH_TOKEN }}" ]; then
+            echo "WARNING: SENTRY_AUTH_TOKEN not set. Source maps will NOT be uploaded to Bugsink."
+            echo "         Errors will show minified stack traces. To fix, add SENTRY_AUTH_TOKEN to Gitea secrets."
+          fi
+
           GITEA_SERVER_URL="https://gitea.projectium.com"
           COMMIT_MESSAGE=$(git log -1 --grep="\[skip ci\]" --invert-grep --pretty=%s)
           PACKAGE_VERSION=$(node -p "require('./package.json').version")
+          GENERATE_SOURCE_MAPS=true \
           VITE_APP_VERSION="$(date +'%Y%m%d-%H%M'):$(git rev-parse --short HEAD):$PACKAGE_VERSION" \
           VITE_APP_COMMIT_URL="$GITEA_SERVER_URL/${{ gitea.repository }}/commit/${{ gitea.sha }}" \
           VITE_APP_COMMIT_MESSAGE="$COMMIT_MESSAGE" \
+          VITE_SENTRY_DSN="${{ secrets.VITE_SENTRY_DSN }}" \
+          VITE_SENTRY_ENVIRONMENT="production" \
+          VITE_SENTRY_ENABLED="true" \
+          SENTRY_AUTH_TOKEN="${{ secrets.SENTRY_AUTH_TOKEN }}" \
+          SENTRY_URL="https://bugsink.projectium.com" \
           VITE_API_BASE_URL=/api VITE_API_KEY=${{ secrets.VITE_GOOGLE_GENAI_API_KEY }} npm run build
 
       - name: Deploy Application to Production Server
@@ -114,10 +131,11 @@ jobs:
         env:
           # --- Production Secrets Injection ---
           DB_HOST: ${{ secrets.DB_HOST }}
-          DB_USER: ${{ secrets.DB_USER }}
+          DB_USER: ${{ secrets.DB_USER_PROD }}
-          DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+          DB_PASSWORD: ${{ secrets.DB_PASSWORD_PROD }}
           DB_NAME: ${{ secrets.DB_DATABASE_PROD }}
-          REDIS_URL: 'redis://localhost:6379'
+          # Explicitly use database 0 for production (test uses database 1)
+          REDIS_URL: 'redis://localhost:6379/0'
           REDIS_PASSWORD: ${{ secrets.REDIS_PASSWORD_PROD }}
           FRONTEND_URL: 'https://flyer-crawler.projectium.com'
           JWT_SECRET: ${{ secrets.JWT_SECRET }}
@@ -129,6 +147,15 @@ jobs:
           SMTP_USER: ''
           SMTP_PASS: ''
           SMTP_FROM_EMAIL: 'noreply@flyer-crawler.projectium.com'
+          # OAuth Providers
+          GOOGLE_CLIENT_ID: ${{ secrets.GOOGLE_CLIENT_ID }}
+          GOOGLE_CLIENT_SECRET: ${{ secrets.GOOGLE_CLIENT_SECRET }}
+          GITHUB_CLIENT_ID: ${{ secrets.GH_CLIENT_ID }}
+          GITHUB_CLIENT_SECRET: ${{ secrets.GH_CLIENT_SECRET }}
+          # Sentry/Bugsink Error Tracking (ADR-015)
+          SENTRY_DSN: ${{ secrets.SENTRY_DSN }}
+          SENTRY_ENVIRONMENT: 'production'
+          SENTRY_ENABLED: 'true'
         run: |
           if [ -z "$DB_HOST" ] || [ -z "$DB_USER" ] || [ -z "$DB_PASSWORD" ] || [ -z "$DB_NAME" ]; then
             echo "ERROR: One or more production database secrets (DB_HOST, DB_USER, DB_PASSWORD, DB_DATABASE_PROD) are not set."
@@ -138,6 +165,10 @@ jobs:
           cd /var/www/flyer-crawler.projectium.com
           npm install --omit=dev
 
+          # --- Cleanup Errored Processes ---
+          echo "Cleaning up errored or stopped PM2 processes..."
+          node -e "const exec = require('child_process').execSync; try { const list = JSON.parse(exec('pm2 jlist').toString()); list.forEach(p => { if (p.pm2_env.status === 'errored' || p.pm2_env.status === 'stopped') { console.log('Deleting ' + p.pm2_env.status + ' process: ' + p.name + ' (' + p.pm2_env.pm_id + ')'); try { exec('pm2 delete ' + p.pm2_env.pm_id); } catch(e) { console.error('Failed to delete ' + p.pm2_env.pm_id); } } }); } catch (e) { console.error('Error cleaning up processes:', e); }"
+
           # --- Version Check Logic ---
           # Get the version from the newly deployed package.json
           NEW_VERSION=$(node -p "require('./package.json').version")
@@ -154,7 +185,7 @@ jobs:
             else
               echo "Version mismatch (Running: $RUNNING_VERSION -> Deployed: $NEW_VERSION) or app not running. Reloading PM2..."
             fi
-            pm2 startOrReload ecosystem.config.cjs --env production && pm2 save
+            pm2 startOrReload ecosystem.config.cjs --update-env && pm2 save
             echo "Production backend server reloaded successfully."
           else
             echo "Version $NEW_VERSION is already running. Skipping PM2 reload."
@@ -181,7 +212,17 @@ jobs:
       - name: Show PM2 Environment for Production
         run: |
           echo "--- Displaying recent PM2 logs for flyer-crawler-api ---"
-          sleep 5
+          sleep 5 # Wait a few seconds for the app to start and log its output.
-          pm2 describe flyer-crawler-api || echo "Could not find production pm2 process."
+
-          pm2 logs flyer-crawler-api --lines 20 --nostream || echo "Could not find production pm2 process."
+          # Resolve the PM2 ID dynamically to ensure we target the correct process
-          pm2 env flyer-crawler-api || echo "Could not find production pm2 process."
+          PM2_ID=$(pm2 jlist | node -e "try { const list = JSON.parse(require('fs').readFileSync(0, 'utf-8')); const app = list.find(p => p.name === 'flyer-crawler-api'); console.log(app ? app.pm2_env.pm_id : ''); } catch(e) { console.log(''); }")
+
+          if [ -n "$PM2_ID" ]; then
+            echo "Found process ID: $PM2_ID"
+            pm2 describe "$PM2_ID" || echo "Failed to describe process $PM2_ID"
+            pm2 logs "$PM2_ID" --lines 20 --nostream || echo "Failed to get logs for $PM2_ID"
+            pm2 env "$PM2_ID" || echo "Failed to get env for $PM2_ID"
+          else
+            echo "Could not find process 'flyer-crawler-api' in pm2 list."
+            pm2 list # Fallback to listing everything to help debug
+          fi

.gitea/workflows/deploy-to-test.yml

@@ -90,10 +90,29 @@ jobs:
         # integration test suite can launch its own, fresh server instance.
         # '|| true' ensures the workflow doesn't fail if the process isn't running.
         run: |
-          pm2 stop flyer-crawler-api-test || true
+          echo "--- Stopping and deleting all test processes ---"
-          pm2 stop flyer-crawler-worker-test || true
+          # Use a script to parse pm2's JSON output and delete any process whose name ends with '-test'.
-          pm2 delete flyer-crawler-api-test || true
+          # This is safer than 'pm2 delete all' and more robust than naming each process individually.
-          pm2 delete flyer-crawler-worker-test || true
+          # It prevents the accumulation of duplicate processes from previous test runs.
+          node -e "const exec = require('child_process').execSync; try { const list = JSON.parse(exec('pm2 jlist').toString()); list.forEach(p => { if (p.name && p.name.endsWith('-test')) { console.log('Deleting test process: ' + p.name + ' (' + p.pm2_env.pm_id + ')'); try { exec('pm2 delete ' + p.pm2_env.pm_id); } catch(e) { console.error('Failed to delete ' + p.pm2_env.pm_id, e.message); } } }); console.log('✅ Test process cleanup complete.'); } catch (e) { if (e.stdout.toString().includes('No process found')) { console.log('No PM2 processes running, cleanup not needed.'); } else { console.error('Error cleaning up test processes:', e.message); } }" || true
+
+      - name: Flush Redis Test Database Before Tests
+        # CRITICAL: Clear Redis database 1 (test database) to remove stale BullMQ jobs.
+        # This prevents old jobs with outdated error messages from polluting test results.
+        # NOTE: We use database 1 for tests to isolate from production (database 0).
+        env:
+          REDIS_PASSWORD: ${{ secrets.REDIS_PASSWORD_TEST }}
+        run: |
+          echo "--- Flushing Redis database 1 (test database) to remove stale jobs ---"
+          if [ -z "$REDIS_PASSWORD" ]; then
+            echo "⚠️ REDIS_PASSWORD_TEST not set, attempting flush without password..."
+            redis-cli -n 1 FLUSHDB || echo "Redis flush failed (no password)"
+          else
+            redis-cli -a "$REDIS_PASSWORD" -n 1 FLUSHDB 2>/dev/null && echo "✅ Redis database 1 (test) flushed successfully." || echo "⚠️ Redis flush failed"
+          fi
+          # Verify the flush worked by checking key count on database 1
+          KEY_COUNT=$(redis-cli -a "$REDIS_PASSWORD" -n 1 DBSIZE 2>/dev/null | grep -oE '[0-9]+' || echo "unknown")
+          echo "Redis database 1 key count after flush: $KEY_COUNT"
 
       - name: Run All Tests and Generate Merged Coverage Report
         # This single step runs both unit and integration tests, then merges their
@@ -102,20 +121,30 @@ jobs:
         env:
           # --- Database credentials for the test suite ---
           # These are injected from Gitea secrets into the runner's environment.
+          # CRITICAL: Use TEST-specific credentials that have CREATE privileges on the public schema.
           DB_HOST: ${{ secrets.DB_HOST }}
-          DB_USER: ${{ secrets.DB_USER }}
+          DB_USER: ${{ secrets.DB_USER_TEST }}
-          DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+          DB_PASSWORD: ${{ secrets.DB_PASSWORD_TEST }}
-          DB_NAME: 'flyer-crawler-test' # Explicitly set for tests
+          DB_NAME: ${{ secrets.DB_DATABASE_TEST }}
 
           # --- Redis credentials for the test suite ---
-          REDIS_URL: 'redis://localhost:6379'
+          # CRITICAL: Use Redis database 1 to isolate tests from production (which uses db 0).
+          # This prevents the production worker from picking up test jobs.
+          REDIS_URL: 'redis://localhost:6379/1'
           REDIS_PASSWORD: ${{ secrets.REDIS_PASSWORD_TEST }}
 
           # --- Integration test specific variables ---
-          FRONTEND_URL: 'http://localhost:3000'
+          FRONTEND_URL: 'https://example.com'
           VITE_API_BASE_URL: 'http://localhost:3001/api'
           GEMINI_API_KEY: ${{ secrets.VITE_GOOGLE_GENAI_API_KEY }}
 
+          # --- Storage path for flyer images ---
+          # CRITICAL: Use an absolute path in the test runner's working directory for file storage.
+          # This ensures tests can read processed files to verify their contents (e.g., EXIF stripping).
+          # Without this, multer and flyerProcessingService default to /var/www/.../flyer-images.
+          # NOTE: We use ${{ github.workspace }} which resolves to the checkout directory.
+          STORAGE_PATH: '${{ github.workspace }}/flyer-images'
+
           # --- JWT Secret for Passport authentication in tests ---
           JWT_SECRET: ${{ secrets.JWT_SECRET }}
 
@@ -126,7 +155,7 @@ jobs:
 
           # --- Increase Node.js memory limit to prevent heap out of memory errors ---
           # This is crucial for memory-intensive tasks like running tests and coverage.
-          NODE_OPTIONS: '--max-old-space-size=8192'
+          NODE_OPTIONS: '--max-old-space-size=8192 --trace-warnings --unhandled-rejections=strict'
 
         run: |
           # Fail-fast check to ensure secrets are configured in Gitea for testing.
@@ -142,15 +171,48 @@ jobs:
           # The `|| true` ensures the workflow continues even if tests fail, allowing coverage to run.
           echo "--- Running Unit Tests ---"
           # npm run test:unit -- --coverage --reporter=verbose --includeTaskLocation --testTimeout=10000 --silent=passed-only || true
-          npm run test:unit -- --coverage --coverage.exclude='**/*.test.ts' --coverage.exclude='**/tests/**' --coverage.exclude='**/mocks/**' --reporter=verbose --includeTaskLocation --testTimeout=10000 --silent=passed-only  --no-file-parallelism || true
+          npm run test:unit -- --coverage \
+            --coverage.exclude='**/*.test.ts' \
+            --coverage.exclude='**/tests/**' \
+            --coverage.exclude='**/mocks/**' \
+            --coverage.exclude='src/components/icons/**' \
+            --coverage.exclude='src/db/**' \
+            --coverage.exclude='src/lib/**' \
+            --coverage.exclude='src/types/**' \
+            --coverage.exclude='**/index.tsx' \
+            --coverage.exclude='**/vite-env.d.ts' \
+            --coverage.exclude='**/vitest.setup.ts' \
+            --reporter=verbose --includeTaskLocation --testTimeout=10000 --silent=passed-only --no-file-parallelism || true
 
           echo "--- Running Integration Tests ---"
-          npm run test:integration -- --coverage --coverage.exclude='**/*.test.ts' --coverage.exclude='**/tests/**' --coverage.exclude='**/mocks/**' --reporter=verbose --includeTaskLocation --testTimeout=10000 --silent=passed-only || true
+          npm run test:integration -- --coverage \
+            --coverage.exclude='**/*.test.ts' \
+            --coverage.exclude='**/tests/**' \
+            --coverage.exclude='**/mocks/**' \
+            --coverage.exclude='src/components/icons/**' \
+            --coverage.exclude='src/db/**' \
+            --coverage.exclude='src/lib/**' \
+            --coverage.exclude='src/types/**' \
+            --coverage.exclude='**/index.tsx' \
+            --coverage.exclude='**/vite-env.d.ts' \
+            --coverage.exclude='**/vitest.setup.ts' \
+            --reporter=verbose --includeTaskLocation --testTimeout=10000 --silent=passed-only || true
 
           echo "--- Running E2E Tests ---"
-          # Run E2E tests using the dedicated E2E config which inherits from integration config.
+          # Run E2E tests using the dedicated E2E config.
-          # We still pass --coverage to enable it, but directory and timeout are now in the config.
+          # E2E uses port 3098, integration uses 3099 to avoid conflicts.
-          npx vitest run --config vitest.config.e2e.ts --coverage --coverage.exclude='**/*.test.ts' --coverage.exclude='**/tests/**' --coverage.exclude='**/mocks/**' --reporter=verbose --no-file-parallelism || true
+          npx vitest run --config vitest.config.e2e.ts --coverage \
+            --coverage.exclude='**/*.test.ts' \
+            --coverage.exclude='**/tests/**' \
+            --coverage.exclude='**/mocks/**' \
+            --coverage.exclude='src/components/icons/**' \
+            --coverage.exclude='src/db/**' \
+            --coverage.exclude='src/lib/**' \
+            --coverage.exclude='src/types/**' \
+            --coverage.exclude='**/index.tsx' \
+            --coverage.exclude='**/vite-env.d.ts' \
+            --coverage.exclude='**/vitest.setup.ts' \
+            --reporter=verbose --no-file-parallelism || true
 
           # Re-enable secret masking for subsequent steps.
           echo "::secret-masking::"
@@ -179,7 +241,19 @@ jobs:
           # Run c8: read raw files from the temp dir, and output an Istanbul JSON report.
           # We only generate the 'json' report here because it's all nyc needs for merging.
           echo "Server coverage report about to be generated..."
-          npx c8 report --exclude='**/*.test.ts' --exclude='**/tests/**' --exclude='**/mocks/**' --reporter=json --temp-directory .coverage/tmp/integration-server --reports-dir .coverage/integration-server
+          npx c8 report \
+            --include='src/**' \
+            --exclude='**/*.test.ts' \
+            --exclude='**/*.test.tsx' \
+            --exclude='**/tests/**' \
+            --exclude='**/mocks/**' \
+            --exclude='hostexecutor/**' \
+            --exclude='scripts/**' \
+            --exclude='*.config.js' \
+            --exclude='*.config.ts' \
+            --reporter=json \
+            --temp-directory .coverage/tmp/integration-server \
+            --reports-dir .coverage/integration-server
           echo "Server coverage report generated. Verifying existence:"
           ls -l .coverage/integration-server/coverage-final.json
 
@@ -219,9 +293,18 @@ jobs:
             --reporter=html \
             --report-dir .coverage/ \
             --temp-dir "$NYC_SOURCE_DIR" \
+            --include "src/**" \
             --exclude "**/*.test.ts" \
+            --exclude "**/*.test.tsx" \
             --exclude "**/tests/**" \
-            --exclude "**/mocks/**"
+            --exclude "**/mocks/**" \
+            --exclude "**/index.tsx" \
+            --exclude "**/vite-env.d.ts" \
+            --exclude "**/vitest.setup.ts" \
+            --exclude "hostexecutor/**" \
+            --exclude "scripts/**" \
+            --exclude "*.config.js" \
+            --exclude "*.config.ts"
 
           # Re-enable secret masking for subsequent steps.
           echo "::secret-masking::"
@@ -234,16 +317,6 @@ jobs:
         if: always() # This step runs even if the previous test or coverage steps failed.
         run: echo "Skipping test artifact cleanup on runner; this is handled on the server."
 
-      - name: Deploy Coverage Report to Public URL
-        if: always()
-        run: |
-          TARGET_DIR="/var/www/flyer-crawler-test.projectium.com/coverage"
-          echo "Deploying HTML coverage report to $TARGET_DIR..."
-          mkdir -p "$TARGET_DIR"
-          rm -rf "$TARGET_DIR"/*
-          cp -r .coverage/* "$TARGET_DIR/"
-          echo "✅ Coverage report deployed to https://flyer-crawler-test.projectium.com/coverage"
-
       - name: Archive Code Coverage Report
         # This action saves the generated HTML coverage report as a downloadable artifact.
         uses: actions/upload-artifact@v3
@@ -256,10 +329,11 @@ jobs:
       - name: Check for Test Database Schema Changes
         env:
           # Use test database credentials for this check.
+          # CRITICAL: Use TEST-specific credentials that have CREATE privileges on the public schema.
           DB_HOST: ${{ secrets.DB_HOST }}
-          DB_USER: ${{ secrets.DB_USER }}
+          DB_USER: ${{ secrets.DB_USER_TEST }}
-          DB_PASSWORD: ${{ secrets.DB_PASSWORD }} # This is used by psql
+          DB_PASSWORD: ${{ secrets.DB_PASSWORD_TEST }}
-          DB_NAME: ${{ secrets.DB_DATABASE_TEST }} # This is used by the application
+          DB_NAME: ${{ secrets.DB_DATABASE_TEST }}
         run: |
           # Fail-fast check to ensure secrets are configured in Gitea.
           if [ -z "$DB_HOST" ] || [ -z "$DB_USER" ] || [ -z "$DB_PASSWORD" ] || [ -z "$DB_NAME" ]; then
@@ -283,7 +357,7 @@ jobs:
             echo "WARNING: No schema hash found in the test database."
             echo "This is expected for a first-time deployment. The hash will be set after a successful deployment."
             echo "--- Debug: Dumping schema_info table ---"
-            PGPASSWORD="$DB_PASSWORD" psql -v ON_ERROR_STOP=0 -h "$DB_HOST" -p 5432 -U "$DB_USER" -d "$DB_NAME" -c "SELECT * FROM public.schema_info;" || true
+            PGPASSWORD="$DB_PASSWORD" psql -v ON_ERROR_STOP=0 -h "$DB_HOST" -p 5432 -U "$DB_USER" -d "$DB_NAME" -P pager=off -c "SELECT * FROM public.schema_info;" || true
             echo "----------------------------------------"
             # We allow the deployment to continue, but a manual schema update is required.
             # You could choose to fail here by adding `exit 1`.
@@ -300,6 +374,11 @@ jobs:
         # We set the environment variable directly in the command line for this step.
         # This maps the Gitea secret to the environment variable the application expects.
         # We also generate and inject the application version, commit URL, and commit message.
+        #
+        # Source Maps (ADR-015): If SENTRY_AUTH_TOKEN is set, the @sentry/vite-plugin will:
+        # 1. Generate hidden source maps during build
+        # 2. Upload them to Bugsink for error de-minification
+        # 3. Delete the .map files after upload (so they're not publicly accessible)
         run: |
           # Fail-fast check for the build-time secret.
           if [ -z "${{ secrets.VITE_GOOGLE_GENAI_API_KEY }}" ]; then
@@ -307,12 +386,25 @@ jobs:
             exit 1
           fi
 
+          # Source map upload is optional - warn if not configured
+          if [ -z "${{ secrets.SENTRY_AUTH_TOKEN }}" ]; then
+            echo "WARNING: SENTRY_AUTH_TOKEN not set. Source maps will NOT be uploaded to Bugsink."
+            echo "         Errors will show minified stack traces. To fix, add SENTRY_AUTH_TOKEN to Gitea secrets."
+          fi
+
           GITEA_SERVER_URL="https://gitea.projectium.com" # Your Gitea instance URL
-          COMMIT_MESSAGE=$(git log -1 --grep="\[skip ci\]" --invert-grep --pretty=%s)
+          # Sanitize commit message to prevent shell injection or build breaks (removes quotes, backticks, backslashes, $)
+          COMMIT_MESSAGE=$(git log -1 --grep="\[skip ci\]" --invert-grep --pretty=%s | tr -d '"`\\$')
           PACKAGE_VERSION=$(node -p "require('./package.json').version")
+          GENERATE_SOURCE_MAPS=true \
           VITE_APP_VERSION="$(date +'%Y%m%d-%H%M'):$(git rev-parse --short HEAD):$PACKAGE_VERSION" \
           VITE_APP_COMMIT_URL="$GITEA_SERVER_URL/${{ gitea.repository }}/commit/${{ gitea.sha }}" \
           VITE_APP_COMMIT_MESSAGE="$COMMIT_MESSAGE" \
+          VITE_SENTRY_DSN="${{ secrets.VITE_SENTRY_DSN_TEST }}" \
+          VITE_SENTRY_ENVIRONMENT="test" \
+          VITE_SENTRY_ENABLED="true" \
+          SENTRY_AUTH_TOKEN="${{ secrets.SENTRY_AUTH_TOKEN }}" \
+          SENTRY_URL="https://bugsink.projectium.com" \
           VITE_API_BASE_URL="https://flyer-crawler-test.projectium.com/api" VITE_API_KEY=${{ secrets.VITE_GOOGLE_GENAI_API_KEY_TEST }} npm run build
 
       - name: Deploy Application to Test Server
@@ -333,6 +425,17 @@ jobs:
           rsync -avz dist/ "$APP_PATH"
           echo "Application deployment complete."
 
+      - name: Deploy Coverage Report to Public URL
+        if: always()
+        run: |
+          TARGET_DIR="/var/www/flyer-crawler-test.projectium.com/coverage"
+          echo "Deploying HTML coverage report to $TARGET_DIR..."
+          mkdir -p "$TARGET_DIR"
+          rm -rf "$TARGET_DIR"/*
+          # The merged nyc report is generated in the .coverage directory. We copy its contents.
+          cp -r .coverage/* "$TARGET_DIR/"
+          echo "✅ Coverage report deployed to https://flyer-crawler-test.projectium.com/coverage"
+
       - name: Install Backend Dependencies and Restart Test Server
         env:
           # --- Test Secrets Injection ---
@@ -340,18 +443,19 @@ jobs:
           # Your Node.js application will read these directly from `process.env`.
 
           # Database Credentials
+          # CRITICAL: Use TEST-specific credentials that have CREATE privileges on the public schema.
           DB_HOST: ${{ secrets.DB_HOST }}
-          DB_USER: ${{ secrets.DB_USER }}
+          DB_USER: ${{ secrets.DB_USER_TEST }}
-          DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+          DB_PASSWORD: ${{ secrets.DB_PASSWORD_TEST }}
           DB_NAME: ${{ secrets.DB_DATABASE_TEST }}
 
-          # Redis Credentials
+          # Redis Credentials (use database 1 to isolate from production)
-          REDIS_URL: 'redis://localhost:6379'
+          REDIS_URL: 'redis://localhost:6379/1'
           REDIS_PASSWORD: ${{ secrets.REDIS_PASSWORD_TEST }}
 
           # Application Secrets
-          FRONTEND_URL: 'https://flyer-crawler-test.projectium.com'
+          FRONTEND_URL: 'https://example.com'
-          JWT_SECRET: ${{ secrets.JWT_SECRET_TEST }}
+          JWT_SECRET: ${{ secrets.JWT_SECRET }}
           GEMINI_API_KEY: ${{ secrets.VITE_GOOGLE_GENAI_API_KEY_TEST }}
           GOOGLE_MAPS_API_KEY: ${{ secrets.GOOGLE_MAPS_API_KEY }}
 
@@ -362,21 +466,38 @@ jobs:
           SMTP_USER: '' # Using MailHog, no auth needed
           SMTP_PASS: '' # Using MailHog, no auth needed
           SMTP_FROM_EMAIL: 'noreply@flyer-crawler-test.projectium.com'
+          # Sentry/Bugsink Error Tracking (ADR-015)
+          SENTRY_DSN: ${{ secrets.SENTRY_DSN_TEST }}
+          SENTRY_ENVIRONMENT: 'test'
+          SENTRY_ENABLED: 'true'
 
         run: |
           # Fail-fast check to ensure secrets are configured in Gitea.
-          if [ -z "$DB_HOST" ] || [ -z "$DB_USER" ] || [ -z "$DB_PASSWORD" ] || [ -z "$DB_NAME" ]; then
+          MISSING_SECRETS=""
-            echo "ERROR: One or more test database secrets (DB_HOST, DB_USER, DB_PASSWORD, DB_DATABASE_TEST) are not set in Gitea repository settings."
+          if [ -z "$DB_HOST" ]; then MISSING_SECRETS="${MISSING_SECRETS} DB_HOST"; fi
+          if [ -z "$DB_USER" ]; then MISSING_SECRETS="${MISSING_SECRETS} DB_USER"; fi
+          if [ -z "$DB_PASSWORD" ]; then MISSING_SECRETS="${MISSING_SECRETS} DB_PASSWORD"; fi
+          if [ -z "$DB_NAME" ]; then MISSING_SECRETS="${MISSING_SECRETS} DB_NAME"; fi
+          if [ -z "$JWT_SECRET" ]; then MISSING_SECRETS="${MISSING_SECRETS} JWT_SECRET"; fi
+
+          if [ ! -z "$MISSING_SECRETS" ]; then
+            echo "ERROR: The following required secrets are missing in Gitea:${MISSING_SECRETS}"
             exit 1
           fi
 
           echo "Installing production dependencies and restarting test server..."
           cd /var/www/flyer-crawler-test.projectium.com
           npm install --omit=dev
-          # Use `startOrReload` with the ecosystem file. This is the standard, idempotent way to deploy.
+
-          # It will START the process if it's not running, or RELOAD it if it is.
+          # --- Cleanup Errored Processes ---
+          echo "Cleaning up errored or stopped PM2 processes..."
+          node -e "const exec = require('child_process').execSync; try { const list = JSON.parse(exec('pm2 jlist').toString()); list.forEach(p => { if (p.pm2_env.status === 'errored' || p.pm2_env.status === 'stopped') { console.log('Deleting ' + p.pm2_env.status + ' process: ' + p.name + ' (' + p.pm2_env.pm_id + ')'); try { exec('pm2 delete ' + p.pm2_env.pm_id); } catch(e) { console.error('Failed to delete ' + p.pm2_env.pm_id); } } }); } catch (e) { console.error('Error cleaning up processes:', e); }"
+
+          # Use `startOrReload` with the TEST ecosystem file. This starts test-specific processes
+          # (flyer-crawler-api-test, flyer-crawler-worker-test, flyer-crawler-analytics-worker-test)
+          # that run separately from production processes.
           # We also add `&& pm2 save` to persist the process list across server reboots.
-          pm2 startOrReload ecosystem.config.cjs --env test && pm2 save
+          pm2 startOrReload ecosystem-test.config.cjs --update-env && pm2 save
           echo "Test backend server reloaded successfully."
 
           # After a successful deployment, update the schema hash in the database.
@@ -411,7 +532,17 @@ jobs:
         run: |
           echo "--- Displaying recent PM2 logs for flyer-crawler-api-test ---"
           # After a reload, the server restarts. We'll show the last 20 lines of the log to see the startup messages.
-          sleep 5 # Wait a few seconds for the app to start and log its output.
+          sleep 5
-          pm2 describe flyer-crawler-api-test || echo "Could not find test pm2 process."
+
-          pm2 logs flyer-crawler-api-test --lines 20 --nostream || echo "Could not find test pm2 process."
+          # Resolve the PM2 ID dynamically to ensure we target the correct process
-          pm2 env flyer-crawler-api-test || echo "Could not find test pm2 process."
+          PM2_ID=$(pm2 jlist | node -e "try { const list = JSON.parse(require('fs').readFileSync(0, 'utf-8')); const app = list.find(p => p.name === 'flyer-crawler-api-test'); console.log(app ? app.pm2_env.pm_id : ''); } catch(e) { console.log(''); }")
+
+          if [ -n "$PM2_ID" ]; then
+            echo "Found process ID: $PM2_ID"
+            pm2 describe "$PM2_ID" || echo "Failed to describe process $PM2_ID"
+            pm2 logs "$PM2_ID" --lines 20 --nostream || echo "Failed to get logs for $PM2_ID"
+            pm2 env "$PM2_ID" || echo "Failed to get env for $PM2_ID"
+          else
+            echo "Could not find process 'flyer-crawler-api-test' in pm2 list."
+            pm2 list # Fallback to listing everything to help debug
+          fi

.gitea/workflows/manual-db-backup.yml

@@ -20,9 +20,9 @@ jobs:
       # Use production database credentials for this entire job.
       DB_HOST: ${{ secrets.DB_HOST }}
       DB_PORT: ${{ secrets.DB_PORT }}
-      DB_USER: ${{ secrets.DB_USER }}
+      DB_USER: ${{ secrets.DB_USER_PROD }}
-      DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+      DB_PASSWORD: ${{ secrets.DB_PASSWORD_PROD }}
-      DB_NAME: ${{ secrets.DB_NAME_PROD }}
+      DB_NAME: ${{ secrets.DB_DATABASE_PROD }}
 
     steps:
       - name: Validate Secrets

.gitea/workflows/manual-db-reset-prod.yml

@@ -23,9 +23,9 @@ jobs:
     env:
       # Use production database credentials for this entire job.
       DB_HOST: ${{ secrets.DB_HOST }}
-      DB_USER: ${{ secrets.DB_USER }}
+      DB_USER: ${{ secrets.DB_USER_PROD }}
-      DB_PASSWORD: ${{ secrets.DB_PASSWORD }} # Used by psql
+      DB_PASSWORD: ${{ secrets.DB_PASSWORD_PROD }}
-      DB_NAME: ${{ secrets.DB_DATABASE_PROD }} # Used by the application
+      DB_NAME: ${{ secrets.DB_DATABASE_PROD }}
 
     steps:
       - name: Checkout Code

.gitea/workflows/manual-db-reset-test.yml

@@ -23,9 +23,9 @@ jobs:
     env:
       # Use test database credentials for this entire job.
       DB_HOST: ${{ secrets.DB_HOST }}
-      DB_USER: ${{ secrets.DB_USER }}
+      DB_USER: ${{ secrets.DB_USER_TEST }}
-      DB_PASSWORD: ${{ secrets.DB_PASSWORD }} # Used by psql
+      DB_PASSWORD: ${{ secrets.DB_PASSWORD_TEST }}
-      DB_NAME: ${{ secrets.DB_DATABASE_TEST }} # Used by the application
+      DB_NAME: ${{ secrets.DB_DATABASE_TEST }}
 
     steps:
       - name: Checkout Code

.gitea/workflows/manual-db-restore.yml

@@ -22,8 +22,8 @@ jobs:
     env:
       # Use production database credentials for this entire job.
       DB_HOST: ${{ secrets.DB_HOST }}
-      DB_USER: ${{ secrets.DB_USER }}
+      DB_USER: ${{ secrets.DB_USER_PROD }}
-      DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+      DB_PASSWORD: ${{ secrets.DB_PASSWORD_PROD }}
       DB_NAME: ${{ secrets.DB_DATABASE_PROD }}
       BACKUP_DIR: '/var/www/backups' # Define a dedicated directory for backups
 

.gitea/workflows/manual-deploy-major.yml

@@ -62,8 +62,8 @@ jobs:
       - name: Check for Production Database Schema Changes
         env:
           DB_HOST: ${{ secrets.DB_HOST }}
-          DB_USER: ${{ secrets.DB_USER }}
+          DB_USER: ${{ secrets.DB_USER_PROD }}
-          DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+          DB_PASSWORD: ${{ secrets.DB_PASSWORD_PROD }}
           DB_NAME: ${{ secrets.DB_DATABASE_PROD }}
         run: |
           if [ -z "$DB_HOST" ] || [ -z "$DB_USER" ] || [ -z "$DB_PASSWORD" ] || [ -z "$DB_NAME" ]; then
@@ -113,10 +113,11 @@ jobs:
         env:
           # --- Production Secrets Injection ---
           DB_HOST: ${{ secrets.DB_HOST }}
-          DB_USER: ${{ secrets.DB_USER }}
+          DB_USER: ${{ secrets.DB_USER_PROD }}
-          DB_PASSWORD: ${{ secrets.DB_PASSWORD }}
+          DB_PASSWORD: ${{ secrets.DB_PASSWORD_PROD }}
           DB_NAME: ${{ secrets.DB_DATABASE_PROD }}
-          REDIS_URL: 'redis://localhost:6379'
+          # Explicitly use database 0 for production (test uses database 1)
+          REDIS_URL: 'redis://localhost:6379/0'
           REDIS_PASSWORD: ${{ secrets.REDIS_PASSWORD_PROD }}
           FRONTEND_URL: 'https://flyer-crawler.projectium.com'
           JWT_SECRET: ${{ secrets.JWT_SECRET }}
@@ -137,6 +138,10 @@ jobs:
           cd /var/www/flyer-crawler.projectium.com
           npm install --omit=dev
 
+          # --- Cleanup Errored Processes ---
+          echo "Cleaning up errored or stopped PM2 processes..."
+          node -e "const exec = require('child_process').execSync; try { const list = JSON.parse(exec('pm2 jlist').toString()); list.forEach(p => { if (p.pm2_env.status === 'errored' || p.pm2_env.status === 'stopped') { console.log('Deleting ' + p.pm2_env.status + ' process: ' + p.name + ' (' + p.pm2_env.pm_id + ')'); try { exec('pm2 delete ' + p.pm2_env.pm_id); } catch(e) { console.error('Failed to delete ' + p.pm2_env.pm_id); } } }); } catch (e) { console.error('Error cleaning up processes:', e); }"
+
           # --- Version Check Logic ---
           # Get the version from the newly deployed package.json
           NEW_VERSION=$(node -p "require('./package.json').version")
@@ -153,7 +158,7 @@ jobs:
             else
               echo "Version mismatch (Running: $RUNNING_VERSION -> Deployed: $NEW_VERSION) or app not running. Reloading PM2..."
             fi
-            pm2 startOrReload ecosystem.config.cjs --env production && pm2 save
+            pm2 startOrReload ecosystem.config.cjs --env production --update-env && pm2 save
             echo "Production backend server reloaded successfully."
           else
             echo "Version $NEW_VERSION is already running. Skipping PM2 reload."

.gitea/workflows/manual-redis-flush-prod.yml

@@ -0,0 +1,167 @@
+# .gitea/workflows/manual-redis-flush-prod.yml
+#
+# DANGER: This workflow is DESTRUCTIVE and intended for manual execution only.
+# It will completely FLUSH the PRODUCTION Redis database (db 0).
+# This will clear all BullMQ queues, sessions, caches, and any other Redis data.
+#
+name: Manual - Flush Production Redis
+
+on:
+  workflow_dispatch:
+    inputs:
+      confirmation:
+        description: 'DANGER: This will FLUSH production Redis. Type "flush-production-redis" to confirm.'
+        required: true
+        default: 'do-not-run'
+      flush_type:
+        description: 'What to flush?'
+        required: true
+        type: choice
+        options:
+          - 'queues-only'
+          - 'entire-database'
+        default: 'queues-only'
+
+jobs:
+  flush-redis:
+    runs-on: projectium.com # This job runs on your self-hosted Gitea runner.
+
+    env:
+      REDIS_PASSWORD: ${{ secrets.REDIS_PASSWORD_PROD }}
+
+    steps:
+      - name: Checkout Code
+        uses: actions/checkout@v3
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v3
+        with:
+          node-version: '20'
+          cache: 'npm'
+          cache-dependency-path: '**/package-lock.json'
+
+      - name: Install Dependencies
+        run: npm ci
+
+      - name: Validate Secrets
+        run: |
+          if [ -z "$REDIS_PASSWORD" ]; then
+            echo "ERROR: REDIS_PASSWORD_PROD secret is not set in Gitea repository settings."
+            exit 1
+          fi
+          echo "✅ Redis password secret is present."
+
+      - name: Verify Confirmation Phrase
+        run: |
+          if [ "${{ gitea.event.inputs.confirmation }}" != "flush-production-redis" ]; then
+            echo "ERROR: Confirmation phrase did not match. Aborting Redis flush."
+            exit 1
+          fi
+          echo "✅ Confirmation accepted. Proceeding with Redis flush."
+
+      - name: Show Current Redis State
+        run: |
+          echo "--- Current Redis Database 0 (Production) State ---"
+          redis-cli -a "$REDIS_PASSWORD" -n 0 INFO keyspace 2>/dev/null || echo "Could not get keyspace info"
+          echo ""
+          echo "--- Key Count ---"
+          KEY_COUNT=$(redis-cli -a "$REDIS_PASSWORD" -n 0 DBSIZE 2>/dev/null | grep -oE '[0-9]+' || echo "unknown")
+          echo "Production Redis (db 0) key count: $KEY_COUNT"
+          echo ""
+          echo "--- BullMQ Queue Keys ---"
+          redis-cli -a "$REDIS_PASSWORD" -n 0 KEYS "bull:*" 2>/dev/null | head -20 || echo "No BullMQ keys found"
+
+      - name: 🚨 FINAL WARNING & PAUSE 🚨
+        run: |
+          echo "*********************************************************************"
+          echo "WARNING: YOU ARE ABOUT TO FLUSH PRODUCTION REDIS DATA."
+          echo "Flush type: ${{ gitea.event.inputs.flush_type }}"
+          echo ""
+          if [ "${{ gitea.event.inputs.flush_type }}" = "entire-database" ]; then
+            echo "This will DELETE ALL Redis data including sessions, caches, and queues!"
+          else
+            echo "This will DELETE ALL BullMQ queue data (pending jobs, failed jobs, etc.)"
+          fi
+          echo ""
+          echo "This action is IRREVERSIBLE. Press Ctrl+C in the runner terminal NOW to cancel."
+          echo "Sleeping for 10 seconds..."
+          echo "*********************************************************************"
+          sleep 10
+
+      - name: Flush BullMQ Queues Only
+        if: ${{ gitea.event.inputs.flush_type == 'queues-only' }}
+        env:
+          REDIS_URL: 'redis://localhost:6379/0'
+        run: |
+          echo "--- Obliterating BullMQ queues using Node.js ---"
+          node -e "
+            const { Queue } = require('bullmq');
+            const IORedis = require('ioredis');
+
+            const connection = new IORedis(process.env.REDIS_URL, {
+              maxRetriesPerRequest: null,
+              password: process.env.REDIS_PASSWORD,
+            });
+
+            const queueNames = [
+              'flyer-processing',
+              'email-sending',
+              'analytics-reporting',
+              'weekly-analytics-reporting',
+              'file-cleanup',
+              'token-cleanup'
+            ];
+
+            (async () => {
+              for (const name of queueNames) {
+                try {
+                  const queue = new Queue(name, { connection });
+                  const counts = await queue.getJobCounts();
+                  console.log('Queue \"' + name + '\" before obliterate:', JSON.stringify(counts));
+                  await queue.obliterate({ force: true });
+                  console.log('✅ Obliterated queue: ' + name);
+                  await queue.close();
+                } catch (err) {
+                  console.error('⚠️ Failed to obliterate queue ' + name + ':', err.message);
+                }
+              }
+              await connection.quit();
+              console.log('✅ All BullMQ queues obliterated.');
+            })();
+          "
+
+      - name: Flush Entire Redis Database
+        if: ${{ gitea.event.inputs.flush_type == 'entire-database' }}
+        run: |
+          echo "--- Flushing entire Redis database 0 (production) ---"
+          redis-cli -a "$REDIS_PASSWORD" -n 0 FLUSHDB 2>/dev/null && echo "✅ Redis database 0 flushed successfully." || echo "❌ Redis flush failed"
+
+      - name: Verify Flush Results
+        run: |
+          echo "--- Redis Database 0 (Production) State After Flush ---"
+          KEY_COUNT=$(redis-cli -a "$REDIS_PASSWORD" -n 0 DBSIZE 2>/dev/null | grep -oE '[0-9]+' || echo "unknown")
+          echo "Production Redis (db 0) key count after flush: $KEY_COUNT"
+          echo ""
+          echo "--- Remaining BullMQ Queue Keys ---"
+          BULL_KEYS=$(redis-cli -a "$REDIS_PASSWORD" -n 0 KEYS "bull:*" 2>/dev/null | wc -l || echo "0")
+          echo "BullMQ key count: $BULL_KEYS"
+
+          if [ "${{ gitea.event.inputs.flush_type }}" = "queues-only" ] && [ "$BULL_KEYS" -gt 0 ]; then
+            echo "⚠️ Warning: Some BullMQ keys may still exist. This can happen if new jobs were added during the flush."
+          fi
+
+      - name: Summary
+        run: |
+          echo ""
+          echo "=========================================="
+          echo "PRODUCTION REDIS FLUSH COMPLETE"
+          echo "=========================================="
+          echo "Flush type: ${{ gitea.event.inputs.flush_type }}"
+          echo "Timestamp: $(date -u '+%Y-%m-%d %H:%M:%S UTC')"
+          echo ""
+          echo "NOTE: If you flushed queues, any pending jobs (flyer processing,"
+          echo "emails, analytics, etc.) have been permanently deleted."
+          echo ""
+          echo "The production workers will automatically start processing"
+          echo "new jobs as they are added to the queues."
+          echo "=========================================="

.gitignore

@@ -11,6 +11,18 @@ node_modules
 dist
 dist-ssr
 *.local
+.env
+*.tsbuildinfo
+
+# Test coverage
+coverage
+.nyc_output
+.coverage
+
+# Test artifacts - flyer-images/ is a runtime directory
+# Test fixtures are stored in src/tests/assets/ instead
+flyer-images/
+test-output.txt
 
 # Editor directories and files
 .vscode/*
@@ -22,3 +34,11 @@ dist-ssr
 *.njsproj
 *.sln
 *.sw?
+Thumbs.db
+.claude/settings.local.json
+nul
+tmpclaude*
+
+
+
+test.tmp

.husky/pre-commit

@@ -0,0 +1 @@
+FORCE_COLOR=0 npx lint-staged --quiet

.lintstagedrc.json

@@ -0,0 +1,4 @@
+{
+  "*.{js,jsx,ts,tsx}": ["eslint --fix --no-color", "prettier --write"],
+  "*.{json,md,css,html,yml,yaml}": ["prettier --write"]
+}

.mcp.json

@@ -0,0 +1,28 @@
+{
+  "mcpServers": {
+    "localerrors": {
+      "command": "d:\\nodejs\\node.exe",
+      "args": ["d:\\gitea\\bugsink-mcp\\dist\\index.js"],
+      "env": {
+        "BUGSINK_URL": "http://127.0.0.1:8000",
+        "BUGSINK_TOKEN": "a609c2886daa4e1e05f1517074d7779a5fb49056"
+      }
+    },
+    "devdb": {
+      "command": "D:\\nodejs\\npx.cmd",
+      "args": [
+        "-y",
+        "@modelcontextprotocol/server-postgres",
+        "postgresql://postgres:postgres@127.0.0.1:5432/flyer_crawler_dev"
+      ]
+    },
+    "gitea-projectium": {
+      "command": "d:\\gitea-mcp\\gitea-mcp.exe",
+      "args": ["run", "-t", "stdio"],
+      "env": {
+        "GITEA_HOST": "https://gitea.projectium.com",
+        "GITEA_ACCESS_TOKEN": "b111259253aa3cadcb6a37618de03bf388f6235a"
+      }
+    }
+  }
+}

.nycrc.json

@@ -0,0 +1,5 @@
+{
+  "text": {
+    "maxCols": 200
+  }
+}

.prettierignore

@@ -0,0 +1,41 @@
+# Dependencies
+node_modules/
+
+# Build output
+dist/
+build/
+.cache/
+
+# Coverage reports
+coverage/
+.coverage/
+
+# IDE and editor configs
+.idea/
+.vscode/
+*.swp
+*.swo
+
+# Logs
+*.log
+logs/
+
+# Environment files (may contain secrets)
+.env*
+!.env.example
+
+# Lock files (managed by package managers)
+package-lock.json
+pnpm-lock.yaml
+yarn.lock
+
+# Generated files
+*.min.js
+*.min.css
+
+# Git directory
+.git/
+.gitea/
+
+# Test artifacts
+__snapshots__/