EVS Infrastructure Environment
Developers should use the
caCORE EVS API for access to EVS services. The underlying software code, object model, use cases, and user documentation are
available for use under an open-source license. Developers are encouraged visit the EVS download page for the relevant files and licensing information.
This drawing portrays some of the major components of the
EVS server environment.
The LexBIG terminology server
is deployed in the same JBOSS container used to deploy caCORE EVS.
caCORE EVS 4.x utilizes LexBIG terminology server to
provide access to all terminology content. Terminology content is stored in a
MySQL database while the LexBIG index files, which enable fast lexical seraching,
are located on the same platform as caCORE EVS server. One of the major goals of
caCORE EVS 4.x is to provide distributed access to LexBIG server. The
Distributed LexBIG API was made available for the first time in the 4.0 release of the caCORE EVS API.
It is based on the LexGRID data model and relies on a LexBIG Proxy and Distributed LexBIG Adapter to
provide remote clients access via the caCORE EVS API to the native LexBIG API.
We anticipate that the Distributed LexBIG API will be the EVS interface used
by new applications across CBIIT and caBIG. Like the other caCORE components,
the caCORE EVS API supports query-by-example (QBE) search. Because
QBE searching of terminology data can in some circumstances be tedious, the
Distributed LexBIG Adapter implements a number of client-side convenience methods
that streamline certain common query patterns.
Block Diagram of EVS Environment
Because of the QBE style of query; availability currently only in Java; and use of the LexGRID model that is very
different than the EVS 3.2 model, it will take some time for developers of EVS-dependent services
to adopt the new Distributed API. Therefore EVS will continue to support the 3.2 API. The drawing shows that
EVS 3.2 support will continue to be provided using two technology stacks. EVS will continue
throughout the caCORE EVS life cycle to operate the legacy caCORE EVS stack that relies on the
DTS and Metaphrase terminology servers. In addition we will support the 3.2 EVS APIs on the caCORE EVS 4.x stack.
Support of the EVS 3.2 model and APIs in caCORE EVS 4.x is achieved by use of a server-side Distributed LexBIG Adapter.
We are operating two stacks supporting the EVS 3.2 model because, while the LexBIG based APIs
in caCORE EVS 4.x operate identically to those in the legacy stack, the differences in the data model are so
substantial that the data returned by LexBIG is in may cases more voluminous than that returned by the legacy service.
Legacy services and applications should continue to rely on the legacy EVS 3.2 stack; new development efforts should use the
caCORE EVS 4.x API; developers upgrading existing services and applications are encouraged to try the 3.2 API in the caCORE EVS 4.x
service as a bridge to the new LexGRID model and distributed API.
EVS Content Production
The NCI Metathesaurus is intended to provide rich synonymy and is useful
in mapping among the codes and terms used in various biomedical vocabularies.
The NCI Thesaurus is the NCI reference vocabulary that provides the concepts used in caCORE and caBIG to establish
data semantics.
We produce two or three "builds" of the NCI Metathesaurus each year. Each build is based
on the most recent release
of the UMLS Metathesaurus, supplemented with NCI-specific vocabulary sources. In
EVS, a vocabulary build corresponds roughly to a major releases of a software
product. Builds may contain wholly new sources; existing sources may be dropped or supplanted. The
availability of new builds, and the differences between builds are announced on the NCI EVS
list server .
In addition
to the build cycle, NCI Metathesaurus is updated monthly. The monthly updates correspond to
minor software releases. They contain new or modified concepts, spelling corrections and other minor changes. 
NCI Thesaurus content comes from
NCI, NCI-sponsored groups and an ongoing review of the open literature. We do not attempt to model
aspects of biology or medicine that fall outside the Institute's needs. The NCI
Thesaurus is updated monthly. The differences between one month's release and
the next are summarized on the NCI EVS list server .
API access to the complete
history of every concept in the Thesaurus is supported in caCORE EVS 3.2 and caCORE EVS 4.x. Real-time access to the complete history of
every concept provides NCI Thesaurus users capability to deal with change over time due to
the way that information is coded or classified.
In addition to the NCI
Metathesaurus, the NCI Thesaurus, the EVS servers provide access to a variety of other vocabulary products
such as SNOMED/CT, LOINC and VA/NDF.
To request technical support or additional information about the NCI EVS Project, visit the Application Support page, or contact ncicb@pop.nci.nih.gov.
|