Working with search attributes¶
The product, ad, and category attributes in Elevate 3 Enterprise can have configured search attributes. Search attributes defines what attributes are usable in panel queries for searching. Attributes that do not have search attributes configured cannot be searched. Configuration is done in the Voyado Elevate manager.
Panels used for search accepts the argument search_attributes. This argument defines which search attributes that will be searched by Elevate.
Search attributes¶
Before a product (and variant), ad, or category attribute can be used for search, autocomplete, or did-you-mean, it must be defined as a search attribute.
| Search property | Description |
|---|---|
Name | The name of the search attribute. Required property |
Locale | The locale in which the attribute text is written. A locale identifier usually consists of at least a language identifier and a region identifier. Defaults to <Use product locale> for products, <Use ad locale> for ads, and <Use category locale> for categories. A variant inherits Locale from its parent product unless it is explicitly set at the variant level. |
Match suffix | Setting to allow a word to also match as the suffix of another word, e.g. a search for board will match with skateboard and surfboard, as long as the word board is also recognised as a stand-alone word in the product data. Commonly used for markets with locales such as sv-SE, de-DE, da-DK, and no-NO where words are affixed (concatenated) to form new words. Does not match as infix or prefix. Defaults to Disabled. |
Suggest | Setting to allow the attribute to be used as a suggestion with the Autocomplete panel, the Did-you-mean panel, the Product suggestions panel, and the Category suggestions panel. Attributes with Suggest enabled are ranked higher when ranking with relevance. Defaults to Disabled. |
Simple format¶
The standard format for search attributes are comma-separated lists.
| Example attributes | Description |
|---|---|
title | Search will be performed in the title search attribute of products. |
title,author | Search will be performed in the title and author search attributes of products. |
Typed format¶
The typed format allows for searching in the attributes of a product's categories. The attributes to must be specified for each entity type. The different entity types are separated by a comma (,) and each entity type is specified as type[comma-separated list of attributes].
It is not possible to use both the simple and the typed format in the same argument.
| Example attributes | Description |
|---|---|
product[title,author] | Search will be performed in the title and author search attributes of products. |
product[title,author],category[display_name] | Search will be performed in the title and author search attributes of products, and in the display_name search attribute of their categories. |
category[display_name] | Search will be performed in the display_name search attribute of categories. |
Refinements¶
Search phrases used in Elevate are always treated as text. This means that no specific refinement can be set for search attributes. However, the text in the product attributes is tokenised (split into words) with special rules for:
- Words combining letters and digits such as model designations, e.g.
iPhone 8 256GB - Words combined with hyphens, e.g.
blue-green - Words with camel case, e.g.
MasterCard
East asian languages
East asian languages requires special rules to split text. The following locales are supported in Elevate:
th-TH, zh-CN, zh-HK, zh-TW, ja-JP, ko-KR, en-HK, en-CN, en-TW, en-KR, en-JP
The resulting set of words are normalised by making the text lower case and removing the following special characters:
`-=[];'\,./·~!@$%^&*()_+{}"|><?§´¨½¼¾¤£€©®’
An extended version of the Snowball stemming algorithm is then used on the set of words. Stemming is a process to produce the stem of a word. It is performed to enable Elevate to match queries to conjugations and inflections of the words in the set.
Stemming availability
Stemming of data and queries is available for locales with the following language components:
ar, ca, da, de, el, en, es, fi, fr, hi, hu, id, in, it, ga, lt, nl, no, pt, ro, ru, sr, sv, tr
Stop words¶
Stop words are common words that are ignored in searches. These words are often added to a search hit, but are not necessary for a relevant hit. Voyado Elevate utilize lists of stop words for several languages. The lists are based on lists of stop words created for use with the Snowball stemming algorithm.
English stop words
The following words are designated stop words for locales with the English language component en:
a, an, in, the, which, that, and, or, of, by, for, with, about, to, from, on, over, under, within, between
Stop word availability
Stop words are available for locales with the following language components:
da, de, en, es, fi, it, nl, no, pt, ru, sv