IBM OmniFind Enterprise Search, Yahoo! Edition

Note: this is not the Business Intelligence product, it's mid-range fulltext search engine

Product Information

Platform: Java in Linux (RedHat 4.3, SUSE 10), Java in Windows XP (SP2, 2003 Server SP1)
Price: Free (up to 500,000 documents) , the very different OmniFind Enterprise Edition starts around $20,000

Features

• Simple installation and configuration using the browser admin
• Can index documents in the file system, including remote file system crawling
• Web robot indexing crawler handles complex sites, with wildcard include and exclude options, can handle IBM Websphere and blog formats.
• Proxy server support for the crawler.
• Crawler does continuous checking, adapting to document change frequency
• No current support for meta robots tags.
• Access control via Basic Authentication (HTTP user name and password), form-based authentication
• Indexes over 400 file types, including text, HTML, rtf, Microsoft Word, Excel, PowerPoint, WordPerfect, etc. (uses Stellent Outside In file converters), up to 50 MB per document
• Extracts HTML title, keyword and description tags, title from documents with properties
• Index/search engine based on Lucene
• Uses Unstructured Information Management Architecture (UIMA) for document indexing processing
• Indexes and searches Arabic, Czech, Danish, German, Greek, English, Spanish, Finnish, French, Hebrew, Italian, Japanese, Korean, Dutch, Norwegian, Polish, Portuguese, Russian, Swedish, Simplified Chinese, and Traditional Chinese.
• Document language detection for better tokenization and stemming
• Automated duplicate detection and deletion.
• Index stores stemmed form of words.
• Default search finds pages with all match terms, query processing supports Internet Query Operators (-, ""), Boolean AND, OR, NOT and parentheses, and metadata field tags for URL, doctype, title, keyword and description.
• Option to search the Internet via Yahoo!
• Query language defined by browser settings or language parameter
• Spellchecker and synonym functions (which can be imported from XML)
• Search suggestions (called "Featured Links" and "Shortcuts") editing via the admin interface or import XML file
• Search results UI customization via the admin interface, as minimal content, ATOM feed mode, HTML with XSLT, or as XML for local processing.
• Can integrate with PHP, ASP, REST (HTTP with parameters), etc.
• Default relevance uses keyword match (TF:IDF), date, path depth and links analysis.
• Relevance weighting can be disabled for any element but keyword match
• Documents are cached in original format, converted to HTML on demand for viewing.
• Admin interface for configuring and checking indexes.
• Reports include general metrics, crawled URLs, response times, popular queries and no-matches.
• Active development and online user forums.
• Enterprise support available from IBM

Articles & Reviews

• PHP Sample and "PHP API" (Reusable Functions) OmniFind Blog : March 6, 2007 , by Sean Johnson
  
  Using PHP 5 to and oye.php to call the OmniFind search engine using REST, with screenshots and a listing of the functions available.
  
  
  
• Adding IBM’s OmniFind Yahoo! Edition to Your Web Site IBM.yahoo.net: December 13, 2006 , by Todd Leyba
  
  Describes several ways to integrate OmniFind, specifically with Blogger sites. Includes details on how to use REST to send the query, get the results in XML
  and format with XSLT.
  
  
• IBM and Yahoo give the gift of free search InfoWorld Test Center Daily : December 13, 2006, by Mike Heck
  
  Quick review praises the easy installation and administration, simple customization of UI, featured links, synonyms and ranking adjustments. The reviewer likes the speed of indexing and the familiar interface, with options to switch to Yahoo's web, audio, video, news and local search.
  
  
  

Examples

• UFO Crawler - topics include UFO Sightings, time travel, conspiracy theories and anomalies.
• Hayes County Search local government site