Searching in GtR

The GtR search uses the open-source product ElasticSearch that in turn leverages the Apache Lucene search engine to provide full text searching. This provides very expressive search functionality including wildcards, Boolean operators etc. Below is a brief introduction to the search vocabulary with some examples of how this can be used on GtR to find what you are looking for. Each resource type (a Project, Person, Organisation or Publication) contains a number of Fields, such as title, abstract, first name and surname. When using exact phrase searches and proximity searches it will look in the fields independently.

When looking at search results you will notice that projects are initially ordered by relevance. Relevance is based upon the match between the search term and text fields for the resource type, the number of matches found, the weighting applied to specific fields like titles and the % of total text relevant to the search term.

Advanced search will vary depending on the tab selected. This allows you to restrict the fields used in the search. Within person and organisation detailed pages you can search related projects. The projects returned will include any with a relationship to the person (i.e. PI, or CoI) or organisation (i.e. Project Partner, Collaborator, Co-Funder or Lead RO).

Facets are ordered alphabetically and will not change order after a search has been completed. Searches can be refined further by selecting one or more facets within a given category and across categories. Performing a new search or changing tabs will clear all facets from the previous searches. Guidance on how this works can be found here

A reset button is available on the organisation and person detailed pages to restore the screen to show all of the projects for that organisation/person.

To share your search you can send the web address to another person.

Simple Term
A single word (may include wildcards or other special operators). Not Case Sensitive.
Test
All things in the search index that contain the words "Test", "TEST" or "TeSt" etc.
Phrase
A Set of terms. Not Case Sensitive.
Test Blood
All things in the search index that contain the words "Test" or "TesT" or "blood" or "BloOD" etc.
Quoted Phrase
Enclosing a phrase with speech marks means match the entire phrase rather than the individual terms.
"Blood Test"
All things in the search index that contain the two words together.
Wildcard ?
A "?" represents any single character in a term.
T?nt
All things in the search index containing words such as "tent" or "tint" or "tnt"
Wildcard *
A "*" matches 0 or more characters.
T*nt
All things in the index containing words such as "Tent", "tint", "TNT", "taint", "taunt", "talent" etc.
Fuzzy Search ~
Performs a likeness search for a single term.(Levenshtein Distance, or Edit Distance). For example, "term"~[similarity] where "term" is the search string and (optional) similarity is a number between 0 and 1. The default value is 0.5 and the higher the number the nearer the returned items will be to the search term. See http://en.wikipedia.org/wiki/Fuzzy_search.
Test~0.8
Things that are within a certain fuzzy distance of the word Test e.g, "Tests", "text" etc.
Proximity Search~
Using the tilda "~" after a quoted searcher looks for terms that are a maximum of the given number of words from each other.
"Test blood"~10
Looks for things in the index that contain both the words "Test" and "blood" within 10 words of each other, such as "blood filled test tube".
Boosting Terms^
You can provide a weighted relevance factor to a term or phrase using the caret character ^. The higher the boost value the more relevant the term or phrase is. Values are from greater than zero e.g. 0.1, 1, 10.
"Test blood"^3 "Test Sugar"^0.5
Finds things in the index related to "test blood" or "test sugar" with "test blood" matches given a higher rating or relevance score.
Boolean AND
Return things contain both Terms or phrases (Boolean operators must be uppercase)
Test AND blood
Finds things in the index related that have both the words Blood and Test.
Boolean OR
Return things contain either or both of the Terms or phrases (Boolean operators must be uppercase)
Test OR blood
Finds things in the index related that contain either of the words Blood or Test.
Boolean NOT
Return things that exclude the subsequent term or phrase (Boolean operators must be uppercase)
Test NOT blood
Finds things in the index related that contain the word Test but not the word blood.
Exclusion -
Return things that exclude the subsequent term or phrase
Test -blood
Finds things in the index related that contain that exclude the word blood and may contain the word Test.
Inclusion +
Return things that include the subsequent term or phrase
Test +blood
Finds things in the index related that contain the word Blood and may also contain the word Test.
Grouping Terms
You can use parentheses to enclose terms or phrases
(Test AND blood) OR "Sugar Level"
Finds things in the index that contain either the words "Test" and "Blood" OR the phrase "Sugar Level".
Escape Character \
These are used to include special characters such as + - && || ! ( ) { } [ ] ^ " ~ * ? : \ in search terms and phrases.
If you wish to search for the phrase (A+B):C use \(A\+B\)\:C
Returns documents containing "(A+B):C".