Backend
Backoffice
Messor

Messor - Retail Management Agent

Overview

Messor is the Retail Management agent in the TKM AI Agency Platform, responsible for managing brands, products, venues, and receipt processing. It handles the loyalty program's retail operations, including store management, product tracking, and points assignment for purchases.

Directory Structure

Backend/Backoffice/Messor/
├── data/                # Retail data storage
├── messor.py           # Main agent implementation
├── api_messor.py       # FastAPI endpoints
├── messor_tools.py     # Retail utilities
├── messor_schema.py    # Data models and schemas
└── test_messor.py      # Unit tests

Main Components

MessorAgent Class

The core component that handles retail operations:

  • Brand management
  • Product tracking
  • Venue operations
  • Receipt processing

Processing Pipeline

  1. Data Management

    • Brand registration
    • Product cataloging
    • Venue tracking
    • Inventory updates

  2. Receipt Processing

    • Image validation
    • OCR processing
    • Product matching
    • Points calculation

  3. Location Services

    • Venue geocoding
    • Proximity search
    • Address validation
    • Map integration

Key Features

Retail Management

  • Brand registration
  • Product cataloging
  • Venue tracking
  • Inventory management

Receipt Processing

  • Image validation
  • OCR integration
  • Product matching
  • Points assignment

Location Services

  • Venue geocoding
  • Proximity search
  • Map integration
  • Address validation

Admin Operations

  • Brand management
  • Product updates
  • Venue configuration
  • Mapping management

Integration

Database Connection

  • SingleStore integration
  • Schema management
  • Query optimization
  • Data archival

External Services

  • Google Places API
  • OCR services
  • Geocoding services
  • Map services

Event System

  • Receipt processing events
  • Points assignment
  • Inventory updates
  • Location updates

Error Handling

Processing Errors

  • Image validation
  • OCR failures
  • Product matching
  • Points calculation

Recovery Mechanisms

  • Retry logic
  • Manual review queue
  • Error notification
  • Data recovery

Integration with Other Agents

Core Dependencies

  • Niger: Data persistence
  • Barbatus: Authentication
  • Gulosa: Points management
  • Lyra: Notifications

Optional Integration

  • Alfari: Configuration
  • Gigantea: Feedback processing
  • Sessile: Cache management
AlfariVulcan