Attributes

Structure

The attribute field is an array of JSON elements. Each attribute must declare the following fields.

Name	Type	Mandatory	Description
type	type as string	yes	See available types below
localizedNames	list	no	List of objects providing available translations per locale: { "locale": "<langcode_countrycode>", "name": "<translation>" }
isRequired	Boolean	no	The attribute is required or not on item creation
isReturned	Boolean	no	XO configuration
isIndexed	Boolean	no	XO configuration
isExported	Boolean	no	XO configuration
attributes	list	yes if type is Object or ObjectList	type of elements in object
listSubType	Boolean	yes if type is List	type of elements in list
isFhrValue	Boolean	no	Define if the attribute is related to the FHR Value attribute
isFhrValueId	Boolean	no	Define if the attribute is related to FHR Value Id attribute

Attributes name can be freely defined _**_but certain attributes are reserved for use.

Available types

Name	Description
BOOLEAN	true/false
FLOAT	Floating numbers
INTEGER	Integer numbers
TEXT	String
TIMESTAMP	Must be in RFC 3339 timestamp format
LIST	List of values, each value with the same type
OBJECT	Key value object, where each value as its own schema.
OBJECTLIST	List of objects (only object of the same schema are allowed)
LOCALIZEDTEXT	Key value object (see below)

Localized Text

The LOCALIZEDTEXT type allows you to add translations to the given field defined in the attributes. It's a key-value object where:

• key is based on the Java locale format <languagecode>_<countrycode>

  • languagecode is based on the ISO 639-1

  • countrycode is defined by ISO 3166-1

• values are strings

Let's say that the title field of your item has a French and German translation:

// LocalizedText attribute definition in itemSchema
{
    "name": "title",
    "type": "LOCALIZEDTEXT"
},‌

// final attribute
{
    "title": {
        "fr_FR": "mon titre",
        "en_GB": "my title",
        "de_DE": "mein Titel"
    }
}

LOCALIZEDTEXT can be used where text fields are available (as is, in a LIST, in an OBJECT or OBJECTLIST)

XO features (recommendation, product tagging) need a default locale to work. This default locale must be set in this api.

Type decision tree

FHR data modeling

To make use of FHR specific features refer to below table on how to model item's attributesype

Types

Type	Usage	Notes
BOOLEAN	Flag out of stock or on sale items.	Implicitly converted to INTEGER, False - 0, True - 1
INTEGER	For (bracketed) facets and sorting on e.g. stock amounts, ratings, etc.	Only positive integer values are allowed.
FLOAT	Typically used for bracketed facets and sorting on prices, margin, discounts etc.	Only positive float values are allowed. Negative values will be considered as undefined.
TEXT	Not multi lingual but possible to navigate on. It is typically used for product codes that are locale independent and need to be displayed or free text searchable	Data type without locale specific values, i.e. its value is the same across all locales.
TIMESTAMP	Same as text	Implicitly converted to datetime.
LOCALIZEDTEXT	Meant for free text search and display purposes. Typical examples of this are product descriptions or image references.	An asset value does not have a locale independent identifier. It is not possible to create facets or sort on assets
LIST	Used exclusively for providing product item's categorization (categories attribute) as list of leaf category ids the item belongs to.	Not multi lingual, any other attribute of the type is implicitly converted to OBJECTLIST where each value is made available for all catalog localizations.
OBJECT	Typically used for facets. One value per item.	Represents fixed list of multi-lingual, locale dependent display values (values property) sharing the same locale independent value identifier (value_id property). Identifiers should be unique to the attribute and can contain only characters a-z0-9 and "_".
OBJECTLIST	Typically used for facets. Multiple values per item.	The type specifics are identical to OBJECT.

Example schema

Example item attribute

FHR functionality matrix

The following table provides an overview of some FHR functionalities and how each types can be used for that:

Type	Create facet/filter	Create ranking	Include in ranking cocktail	Translatable	Searchable	Multiple values per item
BOOLEAN	✔	✔	✔	✘	✘	✘
INTEGER	✔	✔	✔	✘	✘	✘
FLOAT	✔	✔	✔	✘	✘	✘
TEXT	✔	✔	✘	✘	✔	✘
TIMESTAMP	✔	✔	✘	✘	✔	✘
LOCALIZEDTEXT	✘	✘	✘	✔	✔	✘
LIST	✔	✔*	✘	✔	✔	✔
OBJECT	✔	✔	✘	✔	✔	✘
OBJECTLIST	✔	✘	✘	✔	✔	✔

* Not applicable for categories attribute.

FHR Object Type

To work with attribute type "Object" with FHR, the object's attribute must have one attribute with the isFhrValue flag and one with the isFhrValueId flag.

The flags are automatically set if your schema is following this structure:

[
    {
        "name": {attribute_name},
        "type": "OBJECT",
        "attributes": [
            { "name": "value", "type": "TEXT" | "LOCALIZEDTEXT"},
            { "name": "valueId", "type": "TEXT"}
        ]
    }
]

Examples

Here is an example of item schema with one field of each type:

[
    // boolean attribute
    {
        "name": "in_stock",
        "type": "BOOLEAN"
    },

    // float attribute
    {
        "name": "customer_rating",
        "type": "FLOAT"
    },

    // integer attribute
    {
        "name": "price",
        "type": "INTEGER"
    },

    // text attribute
    {
        "name": "title",
        "type": "TEXT"
    },

    // LocalizedText attribute with translation
    {
        "name": "title",
        "type": "LOCALIZEDTEXT"
        "localizedNames": [
            {
                "locale": "en_GB",
                "name": "Title"
            },
            {
                "locale": "fr_FR",
                "name": "Titre"
            }
        ]
    },

    // timestamp attribute
    {
        "name": "start_date",
        "type": "TIMESTAMP"
    },

    // list attribute
    {
        "name": "categories",
        "type": "LIST",
        "listSubType": "TEXT"
    },

    // object attribute
    // must have a nested schema in "attributes" which
    // corresponds to the schema of the object
    {
        "name": "film",
        "type": "OBJECT",
        "attributes": [
            {
                "name": "film_title"
                "type": "TEXT"
            },
            {
                "name": "film_description"
                "type": "TEXT"
            }
        ]
    },

    // FHR object attribute 
    // must have a nested schema in "attributes" which
    // corresponds to the schema of the object
    {
        "name": "brand",
        "type": "OBJECT",
        "attributes": [
            {
                "name": "value_id"
                "type": "TEXT"
            },
            {
                "name": "value"
                "type": "LOCALIZEDTEXT"
            }
        ]
    },

    // objectList attribute
    // must have a nested schema in "attributes" which
    // corresponds to the schema of the objects
    {
        "name": "film"
        "type": "OBJECTLIST",
        "attributes": [
            {
                "name": "film_title",
                "type": "TEXT"
            },
            {
                "name": "film_description",
                "type": "TEXT"
            }
        ]
    },

    // FHR objectList attribute
    // must have a nested schema in "attributes" which
    // corresponds to the schema of the objects
    {
        "name": "color"
        "type": "OBJECTLIST",
        "localizedNames": [
            {
                "locale": "en_GB",
                "name": "Colour"
            },
            {
                "locale": "fr_FR",
                "name": "Couleur"
            }
        ],
        "attributes": [
            {
                "name": "value_id",
                "type": "TEXT"
            },
            {
                "name": "value",
                "type": "LOCALIZEDTEXT"
            }
        ]
    }
]

OBJECT and OBJECTLIST types allow recursive typing:

[
    // object attribute with nested object
    {
        "name": "film"
        "type": "OBJECT"
        "attributes": [
            {
                "name": "film_title",
                "type": "TEXT"
            },
            {
                "name": "film_description",
                "type": "OBJECT",
                "attributes": [
                    {
                        "name": "scenario",
                        "type": "TEXT"
                    }
                    {
                        "name": "actors",
                        "type": "LIST",
                        "listSubType": "TEXT"
                    }
                ]
            }
        ]
    }

Validation

Additional validation:

• each attribute (and attributes of inner schemas) should have unique attribute names

• the following attributes name are reserved and cannot be used : "secondid", "universe", "countries", "itemid", "match_rate", "categories"

FHR data model validation

The query parameter fhrValidation=true can be added to enforces Fredhopper specific validation to the schema API. This includes the following additional checks to the schema:

• Attribute name (aka attribute_ids) are lowercased with only [a-z0-9_]+ i.e. lowercase Latin letters: a-z, digits: 0-9, and underscores: _

• Value (type TEXT) and value_id (type TEXT or LOCALIZEDTEXT) are mandatory for attributes of type OBJECT and OBJECTLIST

• Attribute name (aka attribute_ids) cannot start with a number

• FHR valid types for attributes are : “boolean”, “text”, “integer”, “float”, “list”, “localizedText”, “Object”, “ObjectList”, “timestamp”