System Outline & Architecture
WiSE Standard is a BST’s core packae software product which has various optional features, such as thesaurus and synonym management, faceted navigation, Web crawler (internet robot), etc. WiSE EC is designed specifically with E-commerce and portals sites.
In order to develop search applications, indexing protocol (named ECP: Enhanced connector Protocol), and search API are available to developers, who can develop their own crawler, filter, and search application easily.
Basic Search Features
N-gram (n-successive character strings) based indexing method. Two types of text cutting methods are applied to obtain best search results.
N-gram cutting is best suited for Japanese, Korean, and Chinese texts.
Word cutting by white space for other languages
(e.g. English, French, Arabic, Russian, …)
Range specification for numeric (integer & Floating
point number)、bit calculations. Attribute definition can be made up to 1000 by data type.
5 dimensional sorting
Search support features
Keyword driven recommendation links promotion links, Query cash, Faceted navigation, User defined categories.
Simple Search API
Application can throw search requests via HTTP, and receive results in XML format.
Indexer makes index data base from ECP effectively and speedy.
・ Incremental index update
・ Double index for nonstop operation
documents from various sources. WiSE has the following crawler on default basis.
RDB data can be handled by using SQL select and simple setting. Major RDB can be handled using JCDB driver, such as Oracle, My SQL Server, IBM DB2, PostgresSQL, MySQL.)
The crawler collects data on the machine WiSE installed.
CSV and TSV data can be also collected.
Administrator views custom reports and makes settings
・ Search Log Analysis (top queries, zero-hit queries, etc.)
・ Thesaurus and synonym dictionary management with export/import function
・ Keyword driven recommendation links (promotion links) management with export/import function.
・ Registration of administrator related settings
Additional Features (Options)
In addition to default features, the following options are available.
WiSE Web Crawler (Internet robot)
Web Crawler gathers and categorizes information on the Internet.
Thesaurus and Synonym Management *1
User specific synonym and related words dictionary can be made by thesaurus and synonym management option.
Faceted Navigation *1
Faceted navigation option can provide distribution (histogram) over numeric and date attributes, configurable multiple tree categories, and query candidate words.
Office Document Filter *2
Through office document filter, Microsoft Office, PDF, and other major office documents can be handled.
(optional for WiSE EC only. default function for WiSE Standard)
PDF Navi can provide excellent search experience for PDF documents. Search result can be shown as PDF page basis (instead of PDF file name basis), and contents can be viewed without any downloads.
*1: Option for WiSE Standard. Default function for WiSE EC.
*2: Option for WiSE EC. Default function for WiSE Standard.
Multi Server Support
WiSE is designed to scale among distributed server environments. This feature is called “WiSE Clustering”.
Multiple search servers can be configured to provide hardware redundancy.
Search servers can be scaled out to support massive amount of documents.
N+1(or any) Hardware redundancy can be configured.