How does the search functionality work on OnePetro?


The search engine within OnePetro can be found at the top of each OnePetro page.  OnePetro uses the Solr search engine. 

Solr is an enterprise search platform from the Apache Lucene project and one of the most popular enterprise search engines available.

The search syntax used within OnePetro is Solr's edismax which is much like Google's search syntax.
The results page always includes the search performed, how many results it returned and how long it took in milliseconds.






Feature Syntax Example



Boolean Operators:

This OR that
This AND that
This NOT that

Must Include, Must Not Include and Optional:

This that
+This +that
This -that

Phrase Search:

"This that"

Specifying Fields Title:

"This that"
author:(Smith OR Jones)

Grouping:

(this AND that) OR (that AND other)
















Users can select one of the following to sort their search results by:

  • Relevance
  • Most recent
  • Highest rated (by rating 1.0 – 5.0,including at least two significant digits.)  
  • Note: Documents without a rating will be sorted as if having a rating of 0.
By default, results will be sorted by relevance.  


























Relevance takes into account the following fields and weightings:

























Field

  Phrase Weighting  

  Individual Term  

Document Title

10

5

Abstract

5

2

 Keywords, Categories and Disciplines 

2

1

PDF Full-Text

2

1


























Phrase weighting means wherever a document's fields matches the user's search terms, in the exact order they entered them, that part of the scoring will be boosted by the amount in the table.


Individual term weighting applies when terms are matched but not necessarily in the same order the user entered them. For individual terms, by default we do not apply strict boolean operators (OR/AND), rather we have a 'minimum match'.

For searches with one or two terms, both terms MUST match.  For searches with more than three terms, 25% of the terms are optional. Searches matching more than 75% of the terms will still rank higher.

By selecting 'Most recent' users will be able to sort their results by publication date, newest first 

By selecting 'Highest rated' users will be able to sort their results by the technical rating of the documents.

Users will be able to select between 25, 50 and 100 results per page, with 25 results as a default.