NIH | National Cancer Institute | NCI Wiki  

Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Note
titleNote

The NCI Thesaurus has grown large enough that it can no longer be loaded on many typical desktop machines. We recommend a 64-bit operating system running on a multiprocessor computer with a minimum of 4g of memory. Server class Linux machines are the typical target for these loads. The time to load NCI Thesaurus will vary depending on machine, memory, and disk speed. Expect a couple of hours for a higher end machine.

Take a look at Best Practices for recommendations for loading large terminologies like NCI Thesaurus in optimal way.

Step

Action

1

Using a web or ftp client go to the URL: ftp://ftp1.nci.nih.gov/pub/cacore/EVS/
screenshot of FTP directory

2

Select the version of NCI Thesaurus OWL file you wish to download. Save the file to a directory on your machine.

3

Extract the OWL file from the ZIP download and save in a directory on your machine. This directory will be referred to as NCI_THESAURUS_DIRECTORY in script examples.

4

Create Manifest and Preferences files. (optional)

Manifests update or fill empty terminology metadata and can make adjustments to names and alternate names, versions and many other things that the source terminology normally wants to say. This is very useful when the source terminology itself does not supply this information, but the user needs to record it in the terminology service representation. Some versions of the NCI Thesaurus may not load without a manifest. What follows is a sample manifest file used to update alternate names, language designations, versions and other metadata. Details of manifest elements are found in the Administration Guide to the manifest file.
image of manifest file
Preference files control how data is loaded and, while they do not add anything to the source, they can change the representation of a terminology by making choices as to what is loaded as a property, entity, or association from the source terminology. The following preference file sets root nodes for the terminology and processes a set of complex properties not handled by the OWL processor, among other things. The full preferences definitions are described in the Administration Guide to the preferences file.
image of preference file
Since the Thesaurus may be released as either a "by code" or "by name" formatted source, preference files can serve the purpose of adjusting the entity code to its intended place in LexEVS. Here are some differences in an OWL formatted Thesaurus source.

  • "By Code" formatted Thesaurus OWL class
    Image of "By Code" formatted Thesaurus OWL class
    which has its entity code formatted as an rdf:ID. This loads as the unique identifier without using a preferences file.
  • "By name" version
    Image of By name formatted Thesaurus OWL class
  • With the code found as a "<code>" tagged attribute of the class

    In this case we can insure that the value "C1324" will be loaded into LexEVS as an entity code by supplying a preferences file with the following values. Loading this type of terminology without the preferences file may cause data truncation errors on the longer names and cause data loss on load.
    Image of values
    Finally, post processing is also available for this and other sources, as explained in the Administrative guide post processor section.

5

Using the LexEVS command line, load the NCI Thesaurus with no options:
Change to the LexEVS home directory.

Code Block
cd {LEXEVS_HOME}/admin

For Windows installation use the following command:

Code Block
LoadOWL.bat –in "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_10.10d.owl

For Linux installation use the following command:

Code Block
LoadOWL.sh –in "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_10.10d.owl

This should work best with a "by code" type Thesaurus source terminology.

6

Using the LexEVS command line, load the NCI Thesaurus with options:
Change to the LexEVS home directory.

Code Block
cd {LEXEVS_HOME}/admin

For Windows installation use the following command:

Code Block
LoadOWL.bat –in "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_10.10d.owl" -mf "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_MF.xml"
-lp "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_prefs.xml"

For Linux installation use the following command:

Code Block
LoadOWL.sh –in "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_10.10d.owl" -mf "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_MF.xml"
-lp "file:///{NCI_THESAURUS_DIRECTORY}/Thesaurus_prefs.xml"

...

Note
titleNote

NCI Metathesaurus contains many individual vocabularies some of which are large vocabularies in and of themselves. It requires many hours to load and index. It can require 36 hours on a multiprocessor machine with 6 GB or more of memory. The total time to load the NCI MetaThesaurus will vary depending on machine, memory, and disk speed. Because this loader uses a batch loading strategy it is less dependent on memory, but some users will see 3 or 4 day load times with average multiprocessor processing power. While the batch loading process itself is not heavily memory dependent, the creating the index following the load will require at least 3 GB of memory.

Take a look at Best Practices for recommendations for loading large terminologies like NCI Metathesaurus in optimal way.

Step

Action

1

Using a web or ftp client go to the URL: http://ncicb.nci.nih.gov/download/evsportal.jsp
Note that a valid UMLS license is required to download the NCI Metathesaurus due to the UMLS content inside. You will find out how to manage this on the site above if you do not have a license.
screenshot of download

2

Select the version of NCI Metathesaurus RRF you wish to download. There may only be one. Save the file to a directory on your machine.

3

Extract the files from the ZIP download and save to a directory on your machine. This directory will be referred to as NCI_METATHESAURUS_DIRECTORY. RELASE_INFO.RRF is required to be present for the load utility to work.

4

Check that you are able to open a large number of files before starting the load.

Code Block
ulimit -Hn

Usually having around 10,000 available open files is sufficient. If your limit is set to low this will need to be raised.

5

Using the LexEVS utilities load the NCI Thesaurus:

Code Block
{LEXEVS_HOME}/admin

For Windows installation use the following command:

Code Block
LoadMetaBatch.bat –in "file:///{NCI_METATHESAURUS_DIRECTORY}/"

For Linux installation use the following command:

Code Block
LoadMetaBatch.sh –in "file:///{NCI_THESAURUS_DIRECTORY}/"

...

Step

Action

1

Open the lbGUI

2

Find in the table the terminology with the broken or stopped load.
This should have a status of pending

3

Highlight and double click the terminology row in the table.

4

On the resulting window note for the following command line execution:
The URI and the version.

5

Using the LexEVS utilities restart the load of the NCI Metathesaurus:

Code Block
{LEXEVS_HOME}/admin

For Windows installation use the following command:

Code Block
ResumeMetaBatch.bat –in "file:///{NCI_METATHESAURUS_DIRECTORY}/" -s "NCI Metathesaurus" -uri "urn:oid:2.16.840.1.113883.3.26.1.2" -version "200601"

For Linux installation use the following command:

Code Block
ResumeMetaBatch.sh –in "file:///{NCI_THESAURUS_DIRECTORY}/" -s "NCI Metathesaurus" -uri "urn:oid:2.16.840.1.113883.3.26.1.2" -version "200601"

...