Get Taxa Actor

Brief Description

The returned name strings will include names from two different sources:

• Those names associated with concepts that have been defined to have a synonymy relationship within the authority.
• Those names that have been extracted from the Mammal Networked Information System (MaNIS ).

During import of the names extracted from MaNIS, a simple N-Gram algorithm is used by the TOS to associate each name with the best concept within each authority known to the TOS.

This actor will replace the workflow consisting of three WebService actors that was discussed during the Oct. 2005 All-Hands meeting in San Diego.

Requirements

• The TOS must be stable and publicly accessible
• A user should be able to specify the authoritative list whose data they trust
• A user should be able to limit the breadth of the results by selecting the root node within the authority's hierarchy
• A user should be able to specify any level within the authority's hierarchy from which concepts will be extracted
• A user should have the option of dealing with names that overlap (are associated with) multiple concepts in the result set
• The actor should be able to distinguish overlaps in the result set

Inputs

• name
• Type: string
• The scientific name used to determine the GUID of the concept acting as the root of the sub-tree of the hierarchy defined by the authority.
• Ex: "Mammalia"

• auth
• Type: string
• The authoritative list containing the hierarchy and synonymy information used to extract the names.
• Ex: "ITIS, 2005"

• targetLevel (string)
• Type: string
• The taxonomic rank of the concepts from which names should be extracted.
• Ex: "Species"

• checkForOverlaps
• Type: boolean
• True if the actor should check for overlaps in the names returned, false otherwise

Output

• A 2-dimensional array with the following properties:
• Each row is an association between a concept and a set of names
• Index 0 of each row is the GUID of the concept to which the names contained in the remaining elements is associated
• Ex:

[["urn:lsid:bioguid.org:concept:1", "Pan troglodytes", "chimpanzee"],
 ["urn:lsid:biobuid.org:concept:2", "Pan paniscus", "bonobo"]]

Basic Algorithm

guids = TOS.getBestConcept(name, auth)
if (guids.length > 1) then
  user interaction to pick best guid

string[][] synNames = TOS.getSynsForAuthoritativeList(guid, auth, targetLevel)

if (checkForOverlaps) then
  determine overlapsExist
  if (overlapsExist) then
    user interaction

Development Agenda

• Dec. 09 2005: Stubbed prototype, no interaction with TOS
• Feb. 2006: Stable implementation with no user interaction
• TBD: Stable implementation with user interaction