👤  extract-size-chart

Content

- type: clear
- type: ask
  param: department
  message: Please provide Department (e.g mens, womens, boys, girls etc)
  options: null
  default: ''
  vision:
    enabled: false
    mode: area
    hint: ''
    send: true
- type: ask
  param: category
  message: Please provide Category (e.g tops, bottoms, shoes etc)
  options: null
  default: ''
  vision:
    enabled: false
    mode: area
    hint: ''
    send: true
- type: ask
  param: additionalPrompt
  message: Do you want to add additional prompts
  options: null
  default: ''
  vision:
    enabled: false
    mode: area
    hint: ''
    send: true
- type: gpt
  prompt: >-
    Role: You are a highly precise Data Engineering Assistant specializing in
    e-commerce schema mapping.


    Task: Extract and normalize size chart data from the current view of the
    webpage into a strictly validated JSON format.


    Strict Normalization Rules:


    * Measurement Conversion: All numeric values must be in centimeters (cm). If
    the source is in inches, multiply by 2.54.


    * Size Region: Must be one of: ['us', 'uk', 'eu', 'de', 'fr', 'it', 'jp',
    'cn', 'in', 'au', 'mx']. Default to us.


    * POM (Point of Measure): Map all headers to exactly one of: ['waist',
    'chest', 'bust', 'hip', 'thigh', 'sleeveLength', 'neck', 'length', 'height',
    'highHip', 'lowHip', 'inseam', 'underBust', 'footLength', 'footWidth',
    'ankleGirth', 'calfWidth'].


    * **SellingVariantDimension** (JSON object): Key = "sizeType", Value should
    be one of: ['regular', 'petite', 'tall', 'plus', 'bigandtall']. Default to
    `{"sizeType": "regular"}` if unspecified. Use double quotes for the string
    value (e.g., `{"sizeType": "petite"}`).


    * Hierarchy Labels: Use a semicolon (;) as the internal separator to ensure
    CSV safety. Example: ["mens"; "bottoms"; "jeans"].


    * Selling Size: Maintain precision. '0', '00', and '000' are unique. Convert
    mixed fractions to decimals (e.g., 30 1/2 to 30.5).


    * {{additionalPrompt}}


    * Data Structure Per Entry: Produce a JSON array where each object
    represents one measurement for one size:

    ```json

    {
        "storeKey": "tfc",
        "storeStyle": "",
        "brand": "Extract from page in smallcase",
        "department": "{{department}}",
        "category": "{{category}}",
        "sellingSize": "Quote exact value",
        "sellingVariantDimensions": {"sizeType": "regular"},
        "hierarchyLabels": ["{{department}}", "{{category}}"],
        "uiLocale": "",
        "sizeRegion": "size region value",
        "source": "LEGACY_SIZE_CHART",
        "pom": "normalized_camelCase_value",
        "generalMeasurement": null,
        "low": float_value_in_cm,
        "high": float_value_in_cm,
        "needsReview": false,
        "reviewReason": null,
        "sourceUrl": "should be the current web page URL",
        "uom": "cm",
        "lastUpdated": "2026-01-30 18:00:00.000"
    }

    ```

    Constraint: Return ONLY the JSON array. Do not include conversational text,
    markdown explanations, or apologies. If a measurement is a single number,
    set both low and high to that value.
  silent: false
  param: ''