Extension talk:SphinxSearch/Feature requests

From MediaWiki.org
Jump to navigation Jump to search

Feature requests[edit]

Feature requests should generally be created as bugzilla report and marked as enhancement. This process helps to consolidate and prioritize any request.

MW Search and advanced search form feature[edit]

With version 0.8+, the SphinxSearch uses the standard MW search interface. Despite the simple interface which works most of the time, we do believe that an advanced search functionality for the MW search should be an important feature to ensure information can be correctly disseminated. Unfortunately still in MW 1.17, 1.18 it is not possible for the advanced search screen to have more options than just to select to namespaces.

Any website that stores large amount of information provides advanced search input features. This bugzilla report is open since 2010, so people who feel compelled that such a feature should be implemented in the core MW search interface should vote and raise their voices in bugzilla.

The newly created SpecialSearch.php hooks in MW 1.18 allows to create a new search profile and where one can add additional search options and logic that can be transferred to the Sphinx search engine. The Extension:Translate shows how such search profile can be developed (Extension:Translate does not make use of SphinxSearch but it shows on how to implement such search profile)

feature request: show categories in result list[edit]

it would be great to be able to list the categories of articles in search results. --Nilsja 01:29, 1 June 2010 (UTC)

This would be a neat feature, and one my users would like. I looked in SphinxSearch_Body.php as this is where the excerpt is generated. I am not php expert, but couldn't easily see where the categories of the result articles were stored, or how to add these to the generated excerpt, perhaps as a list just before the text excerpt is shown. This would be really useful. Happy to add this to a hack in My current version if someone would suggest what needs to get added. -- Brett.tyson 13:20, 18 April 2011 (UTC)

Sphinx and PDF indexing[edit]

PDF indexing could be neat feature in future (PDF's stored in Mediawiki's image folder). This section works as preliminary research section in how this might be achieved. Notes from the sphinx forum suggest that a script to output an XML dump from the files on the filesystem and turn it into a large file for the xmlpipe2 method ... or output it to multiple files which can be indexed with Sphinx

Faceted Search[edit]

Faceted Search is known as a possibility for dynamic clustering of items or search results into categories that let users drill into search results (or even skip searching entirely) by any value in any field search engines such as Solr support this kind of search categorization.

Facets, multi-queries, and searching ... describes multi-queries that can make the searches faster and optimize common parts between queries and further suggests to use multi-queries to allow sending a number of queries in one batch: one for main search results, a few others with different grouping and sorting settings (top-3 authors, top-5 vendors, top-20 sub-category counts, etc) and so perform a faceted search. Discussion around this topic can be found on the Sphinx forum 3326, 1761, 6045, 7822, 8376.

Bugzilla: 32203

See also