Vertical search engine that mines data for one specific narrow niche. The prototype is [url removed, login to view], this site represents the best what we are looking for. Search engine is supposed to crawl and index websites defined in admin area. Output - html listing and RSS 2.0 compatible xml. The whole package should be lightweight, installation and administration friendly and easily customizable for other projects (niches). Probably the search engine will be based on Lucene ([url removed, login to view] - preferably C++ version).
Another posibility is adaptation of swish-e or swish++
- Output of search results in RSS2.0 compatible XML
- Backend GUI to define URL's for crawling/searching/indexing