---
name: categorization_golden_v1_light
description: Lightweight golden dataset for quick transaction categorization evaluation
eval_type: categorization
version: "1.0"
metadata:
  created_at: "2025-12-04"
  updated_at: "2025-12-04"
  source: manual_curation
  notes: |
    A compact 50-sample dataset designed for quick evaluation runs.
    Includes a balanced mix across:
    - All difficulty levels (easy, medium, hard, edge_case)
    - All major category types
    - Both US and European merchants
    - Representative edge cases

    Difficulty distribution:
    - easy: 20 samples
    - medium: 15 samples
    - hard: 10 samples
    - edge_case: 5 samples

context:
  categories:
    - id: "income"
      name: "Income"
      classification: "income"
      is_subcategory: false
    - id: "salary"
      name: "Salary"
      classification: "income"
      is_subcategory: true
      parent_id: "income"
    - id: "food_and_drink"
      name: "Food & Drink"
      classification: "expense"
      is_subcategory: false
    - id: "restaurants"
      name: "Restaurants"
      classification: "expense"
      is_subcategory: true
      parent_id: "food_and_drink"
    - id: "fast_food"
      name: "Fast Food"
      classification: "expense"
      is_subcategory: true
      parent_id: "food_and_drink"
    - id: "groceries"
      name: "Groceries"
      classification: "expense"
      is_subcategory: true
      parent_id: "food_and_drink"
    - id: "coffee_shops"
      name: "Coffee Shops"
      classification: "expense"
      is_subcategory: true
      parent_id: "food_and_drink"
    - id: "shopping"
      name: "Shopping"
      classification: "expense"
      is_subcategory: false
    - id: "clothing"
      name: "Clothing"
      classification: "expense"
      is_subcategory: true
      parent_id: "shopping"
    - id: "electronics"
      name: "Electronics"
      classification: "expense"
      is_subcategory: true
      parent_id: "shopping"
    - id: "transportation"
      name: "Transportation"
      classification: "expense"
      is_subcategory: false
    - id: "gas"
      name: "Gas & Fuel"
      classification: "expense"
      is_subcategory: true
      parent_id: "transportation"
    - id: "rideshare"
      name: "Rideshare"
      classification: "expense"
      is_subcategory: true
      parent_id: "transportation"
    - id: "public_transit"
      name: "Public Transit"
      classification: "expense"
      is_subcategory: true
      parent_id: "transportation"
    - id: "entertainment"
      name: "Entertainment"
      classification: "expense"
      is_subcategory: false
    - id: "streaming"
      name: "Streaming Services"
      classification: "expense"
      is_subcategory: true
      parent_id: "entertainment"
    - id: "utilities"
      name: "Utilities"
      classification: "expense"
      is_subcategory: false
    - id: "housing"
      name: "Housing"
      classification: "expense"
      is_subcategory: false
    - id: "rent"
      name: "Rent"
      classification: "expense"
      is_subcategory: true
      parent_id: "housing"
    - id: "health"
      name: "Health & Wellness"
      classification: "expense"
      is_subcategory: false
    - id: "pharmacy"
      name: "Pharmacy"
      classification: "expense"
      is_subcategory: true
      parent_id: "health"
    - id: "gym"
      name: "Gym & Fitness"
      classification: "expense"
      is_subcategory: true
      parent_id: "health"
    - id: "travel"
      name: "Travel"
      classification: "expense"
      is_subcategory: false
    - id: "flights"
      name: "Flights"
      classification: "expense"
      is_subcategory: true
      parent_id: "travel"
    - id: "hotels"
      name: "Hotels"
      classification: "expense"
      is_subcategory: true
      parent_id: "travel"
    - id: "subscriptions"
      name: "Subscriptions"
      classification: "expense"
      is_subcategory: false
    - id: "personal_care"
      name: "Personal Care"
      classification: "expense"
      is_subcategory: false
    - id: "gifts"
      name: "Gifts & Donations"
      classification: "expense"
      is_subcategory: false

samples:
  # =============================================================================
  # EASY SAMPLES (20 samples) - Clear, unambiguous merchants
  # =============================================================================

  # Fast Food
  - id: cat_light_easy_001
    difficulty: easy
    tags: [fast_food, us]
    input:
      id: txn_light_001
      amount: 12.99
      classification: expense
      description: "MCDONALD'S #12345"
    expected:
      category_name: "Fast Food"

  - id: cat_light_easy_002
    difficulty: easy
    tags: [fast_food, us]
    input:
      id: txn_light_002
      amount: 14.50
      classification: expense
      description: "CHIPOTLE MEXICAN GRILL"
    expected:
      category_name: "Fast Food"

  # Coffee Shops
  - id: cat_light_easy_003
    difficulty: easy
    tags: [coffee_shops, us]
    input:
      id: txn_light_003
      amount: 5.75
      classification: expense
      description: "STARBUCKS STORE #9876"
    expected:
      category_name: "Coffee Shops"

  - id: cat_light_easy_004
    difficulty: easy
    tags: [coffee_shops, europe, uk]
    input:
      id: txn_light_004
      amount: 4.50
      classification: expense
      description: "COSTA COFFEE LTD"
    expected:
      category_name: "Coffee Shops"

  # Groceries
  - id: cat_light_easy_005
    difficulty: easy
    tags: [groceries, us]
    input:
      id: txn_light_005
      amount: 156.32
      classification: expense
      description: "WHOLE FOODS MKT #10234"
    expected:
      category_name: "Groceries"

  - id: cat_light_easy_006
    difficulty: easy
    tags: [groceries, europe, uk]
    input:
      id: txn_light_006
      amount: 87.50
      classification: expense
      description: "TESCO STORES LTD"
    expected:
      category_name: "Groceries"

  - id: cat_light_easy_007
    difficulty: easy
    tags: [groceries, europe, germany]
    input:
      id: txn_light_007
      amount: 78.90
      classification: expense
      description: "LIDL DIENSTLEISTUNG"
    expected:
      category_name: "Groceries"

  # Gas & Fuel
  - id: cat_light_easy_008
    difficulty: easy
    tags: [gas, us]
    input:
      id: txn_light_008
      amount: 45.00
      classification: expense
      description: "SHELL OIL 573849234"
    expected:
      category_name: "Gas & Fuel"

  - id: cat_light_easy_009
    difficulty: easy
    tags: [gas, europe, uk]
    input:
      id: txn_light_009
      amount: 75.00
      classification: expense
      description: "BP OIL UK LTD"
    expected:
      category_name: "Gas & Fuel"

  # Rideshare
  - id: cat_light_easy_010
    difficulty: easy
    tags: [rideshare, us]
    input:
      id: txn_light_010
      amount: 23.50
      classification: expense
      description: "UBER *TRIP HELP.UBER.COM"
    expected:
      category_name: "Rideshare"

  # Streaming
  - id: cat_light_easy_011
    difficulty: easy
    tags: [streaming, us]
    input:
      id: txn_light_011
      amount: 15.99
      classification: expense
      description: "NETFLIX.COM"
    expected:
      category_name: "Streaming Services"

  - id: cat_light_easy_012
    difficulty: easy
    tags: [streaming, us]
    input:
      id: txn_light_012
      amount: 10.99
      classification: expense
      description: "SPOTIFY USA"
    expected:
      category_name: "Streaming Services"

  # Electronics
  - id: cat_light_easy_013
    difficulty: easy
    tags: [electronics, us]
    input:
      id: txn_light_013
      amount: 299.99
      classification: expense
      description: "BEST BUY 00000456"
    expected:
      category_name: "Electronics"
      acceptable_alternatives: ["Shopping"]

  # Clothing
  - id: cat_light_easy_014
    difficulty: easy
    tags: [clothing, europe, spain]
    input:
      id: txn_light_014
      amount: 79.99
      classification: expense
      description: "ZARA ESPANA SA"
    expected:
      category_name: "Clothing"
      acceptable_alternatives: ["Shopping"]

  # Pharmacy
  - id: cat_light_easy_015
    difficulty: easy
    tags: [pharmacy, us]
    input:
      id: txn_light_015
      amount: 24.99
      classification: expense
      description: "CVS/PHARMACY #4567"
    expected:
      category_name: "Pharmacy"

  # Flights
  - id: cat_light_easy_016
    difficulty: easy
    tags: [flights, us]
    input:
      id: txn_light_016
      amount: 345.00
      classification: expense
      description: "UNITED AIRLINES 0162345678"
    expected:
      category_name: "Flights"

  - id: cat_light_easy_017
    difficulty: easy
    tags: [flights, europe, ireland]
    input:
      id: txn_light_017
      amount: 89.99
      classification: expense
      description: "RYANAIR DAC"
    expected:
      category_name: "Flights"

  # Hotels
  - id: cat_light_easy_018
    difficulty: easy
    tags: [hotels, us]
    input:
      id: txn_light_018
      amount: 189.00
      classification: expense
      description: "MARRIOTT HOTELS NYC"
    expected:
      category_name: "Hotels"

  # Gym
  - id: cat_light_easy_019
    difficulty: easy
    tags: [gym, us]
    input:
      id: txn_light_019
      amount: 39.99
      classification: expense
      description: "PLANET FITNESS MONTHLY"
    expected:
      category_name: "Gym & Fitness"

  # Income
  - id: cat_light_easy_020
    difficulty: easy
    tags: [income, salary, us]
    input:
      id: txn_light_020
      amount: 3500.00
      classification: income
      description: "ACME CORP PAYROLL"
    expected:
      category_name: "Salary"

  # =============================================================================
  # MEDIUM SAMPLES (15 samples) - Requires domain knowledge
  # =============================================================================

  # Restaurants
  - id: cat_light_med_001
    difficulty: medium
    tags: [restaurants, us]
    input:
      id: txn_light_med_001
      amount: 67.50
      classification: expense
      description: "OLIVE GARDEN #456"
    expected:
      category_name: "Restaurants"

  - id: cat_light_med_002
    difficulty: medium
    tags: [restaurants, europe, uk]
    input:
      id: txn_light_med_002
      amount: 78.50
      classification: expense
      description: "WAGAMAMA LTD LONDON"
    expected:
      category_name: "Restaurants"

  # Warehouse stores
  - id: cat_light_med_003
    difficulty: medium
    tags: [groceries, us, warehouse]
    input:
      id: txn_light_med_003
      amount: 234.56
      classification: expense
      description: "COSTCO WHSE #1234"
    expected:
      category_name: "Groceries"
      acceptable_alternatives: ["Shopping"]

  # Utilities
  - id: cat_light_med_004
    difficulty: medium
    tags: [utilities, us]
    input:
      id: txn_light_med_004
      amount: 125.00
      classification: expense
      description: "CON EDISON PAYMENT"
    expected:
      category_name: "Utilities"

  - id: cat_light_med_005
    difficulty: medium
    tags: [utilities, europe, uk]
    input:
      id: txn_light_med_005
      amount: 156.00
      classification: expense
      description: "BRITISH GAS SERVICES"
    expected:
      category_name: "Utilities"

  - id: cat_light_med_006
    difficulty: medium
    tags: [utilities, us]
    input:
      id: txn_light_med_006
      amount: 89.00
      classification: expense
      description: "AT&T WIRELESS"
    expected:
      category_name: "Utilities"

  # Public Transit
  - id: cat_light_med_007
    difficulty: medium
    tags: [public_transit, us]
    input:
      id: txn_light_med_007
      amount: 127.00
      classification: expense
      description: "MTA *METROCARD"
    expected:
      category_name: "Public Transit"

  - id: cat_light_med_008
    difficulty: medium
    tags: [public_transit, europe, uk]
    input:
      id: txn_light_med_008
      amount: 156.50
      classification: expense
      description: "TFL TRAVEL LONDON"
    expected:
      category_name: "Public Transit"

  # Housing
  - id: cat_light_med_009
    difficulty: medium
    tags: [rent, us]
    input:
      id: txn_light_med_009
      amount: 2100.00
      classification: expense
      description: "AVALON APARTMENTS RENT"
    expected:
      category_name: "Rent"
      acceptable_alternatives: ["Housing"]

  # Subscriptions
  - id: cat_light_med_010
    difficulty: medium
    tags: [subscriptions, us]
    input:
      id: txn_light_med_010
      amount: 9.99
      classification: expense
      description: "APPLE.COM/BILL"
    expected:
      category_name: "Subscriptions"

  # Gifts & Donations
  - id: cat_light_med_011
    difficulty: medium
    tags: [gifts, us, donation]
    input:
      id: txn_light_med_011
      amount: 50.00
      classification: expense
      description: "RED CROSS DONATION"
    expected:
      category_name: "Gifts & Donations"

  # Entertainment
  - id: cat_light_med_012
    difficulty: medium
    tags: [entertainment, us]
    input:
      id: txn_light_med_012
      amount: 89.00
      classification: expense
      description: "TICKETMASTER *EVENT"
    expected:
      category_name: "Entertainment"

  # Travel
  - id: cat_light_med_013
    difficulty: medium
    tags: [hotels, us]
    input:
      id: txn_light_med_013
      amount: 234.00
      classification: expense
      description: "AIRBNB *HMQT5J6QQJ"
    expected:
      category_name: "Hotels"
      acceptable_alternatives: ["Travel"]

  # Personal Care
  - id: cat_light_med_014
    difficulty: medium
    tags: [personal_care, us]
    input:
      id: txn_light_med_014
      amount: 45.00
      classification: expense
      description: "SUPERCUTS #1234"
    expected:
      category_name: "Personal Care"

  # Income
  - id: cat_light_med_015
    difficulty: medium
    tags: [income, us]
    input:
      id: txn_light_med_015
      amount: 500.00
      classification: income
      description: "VENMO CASHOUT"
    expected:
      category_name: "Income"

  # =============================================================================
  # HARD SAMPLES (10 samples) - Ambiguous, multiple interpretations
  # =============================================================================

  # Big-box stores
  - id: cat_light_hard_001
    difficulty: hard
    tags: [ambiguous, us, multi_purpose_retailer]
    input:
      id: txn_light_hard_001
      amount: 156.78
      classification: expense
      description: "TARGET #1234"
    expected:
      category_name: "Shopping"
      acceptable_alternatives: ["Groceries"]

  - id: cat_light_hard_002
    difficulty: hard
    tags: [ambiguous, europe, uk, multi_purpose_retailer]
    input:
      id: txn_light_hard_002
      amount: 156.00
      classification: expense
      description: "MARKS & SPENCER PLC"
    expected:
      category_name: "Shopping"
      acceptable_alternatives: ["Groceries", "Clothing"]

  # Online marketplaces
  - id: cat_light_hard_003
    difficulty: hard
    tags: [ambiguous, us, online_marketplace]
    input:
      id: txn_light_hard_003
      amount: 89.99
      classification: expense
      description: "AMAZON.COM*1A2B3C4D"
    expected:
      category_name: "Shopping"

  # Payment processors (should be null)
  - id: cat_light_hard_004
    difficulty: hard
    tags: [ambiguous, us, payment_processor]
    input:
      id: txn_light_hard_004
      amount: 78.00
      classification: expense
      description: "PAYPAL *JOHNSMITH"
    expected:
      category_name: null

  # Fast-casual
  - id: cat_light_hard_005
    difficulty: hard
    tags: [ambiguous, us, fast_casual]
    input:
      id: txn_light_hard_005
      amount: 34.50
      classification: expense
      description: "PANERA BREAD #567"
    expected:
      category_name: "Restaurants"
      acceptable_alternatives: ["Fast Food"]

  # Delivery services
  - id: cat_light_hard_006
    difficulty: hard
    tags: [ambiguous, us, delivery_service]
    input:
      id: txn_light_hard_006
      amount: 45.00
      classification: expense
      description: "DOORDASH*CHIPOTLE"
    expected:
      category_name: "Fast Food"
      acceptable_alternatives: ["Restaurants"]

  - id: cat_light_hard_007
    difficulty: hard
    tags: [ambiguous, europe, uk, delivery_service]
    input:
      id: txn_light_hard_007
      amount: 32.50
      classification: expense
      description: "DELIVEROO UK LTD"
    expected:
      category_name: "Restaurants"
      acceptable_alternatives: ["Fast Food"]

  # Amazon Prime
  - id: cat_light_hard_008
    difficulty: hard
    tags: [ambiguous, us, amazon]
    input:
      id: txn_light_hard_008
      amount: 14.99
      classification: expense
      description: "AMAZON PRIME*1A2B3C"
    expected:
      category_name: "Subscriptions"

  # Convenience store
  - id: cat_light_hard_009
    difficulty: hard
    tags: [ambiguous, us, convenience_store]
    input:
      id: txn_light_hard_009
      amount: 12.50
      classification: expense
      description: "7-ELEVEN #34567"
    expected:
      category_name: "Groceries"
      acceptable_alternatives: ["Fast Food"]

  # Streaming vs Subscription
  - id: cat_light_hard_010
    difficulty: hard
    tags: [ambiguous, us, streaming_subscription]
    input:
      id: txn_light_hard_010
      amount: 15.99
      classification: expense
      description: "HBO MAX"
    expected:
      category_name: "Streaming Services"
      acceptable_alternatives: ["Subscriptions"]

  # =============================================================================
  # EDGE CASES (5 samples) - Should return null
  # =============================================================================

  # Generic POS
  - id: cat_light_edge_001
    difficulty: edge_case
    tags: [should_be_null, generic_pos]
    input:
      id: txn_light_edge_001
      amount: 15.00
      classification: expense
      description: "POS DEBIT 12345"
    expected:
      category_name: null

  # ACH transfer
  - id: cat_light_edge_002
    difficulty: edge_case
    tags: [should_be_null, transfer]
    input:
      id: txn_light_edge_002
      amount: 100.00
      classification: expense
      description: "ACH WITHDRAWAL"
    expected:
      category_name: null

  # ATM
  - id: cat_light_edge_003
    difficulty: edge_case
    tags: [should_be_null, atm]
    input:
      id: txn_light_edge_003
      amount: 200.00
      classification: expense
      description: "ATM WITHDRAWAL 12345"
    expected:
      category_name: null

  # Check
  - id: cat_light_edge_004
    difficulty: edge_case
    tags: [should_be_null, check]
    input:
      id: txn_light_edge_004
      amount: 350.00
      classification: expense
      description: "CHECK #1234"
    expected:
      category_name: null

  # Cryptic
  - id: cat_light_edge_005
    difficulty: edge_case
    tags: [should_be_null, cryptic]
    input:
      id: txn_light_edge_005
      amount: 45.67
      classification: expense
      description: "TXN*89234*AUTH"
    expected:
      category_name: null