Release Notes: This version moves Yioop! from using a bag of words index to a positional index. Proximity scores are calculated when a multi-word query is done. The host of an inlink is now also incorporated in a ranking. An indexing plugin architecture has been created to make it easier to customize the crawl process; an example recipe plugin is included. An SEO cache view and improved test query statistics have been added. This version finally includes bug fixes for memcached and bmp handling.
Release Notes: This version expands the capabilities of the crawl mixing. Now you can take a query and specify that the first result should come from a particular crawl such as an open Web crawl, the second result should come from Web pages from Wikipedia, the next result should be an image, and all remaining results should come again from the open Web crawl results. This version also reworks the index dictionary to try to make search results appear faster. It also has improvements to the way pages are ranked when indexing raw arc, ODF-RDF, or media-wiki dumps.
Release Notes: This version provides preliminary support for archive crawling of arc, media wiki, and open directory RDF files. It allows re-crawls of previously created Yioop! WebArchives. It also makes it easier to add stemmers for languages other than English (for which there is already a stemmer). Finally, it fixes several bugs in indexing and improves the group by iterator.