CodeMapper assists in the creation of code sets from case definitions, for several coding systems simultaneously while keeping a record of the complete mapping process. Its workflow is structured in three steps:
A presentation describes the concepts, and shows the user interface with a walk-through. More details about the background, implementation, and effectiveness of the approach are documented in our article:
Becker BFH, Avillach P, Romio S, van Mulligen EM, Weibel D, Sturkenboom MCJM, Kors J: CodeMapper: Semi-automatic coding of case definitions. A contribution from the ADVANCE project. Pharmacoepidemiology and Drug Safety 2017. doi:10.1002/pds.4245 (link)
The installation of CodeMapper requires two Java web applications which can be deployed on a java web application server like Tomcat (version 7 is required by CodeMapper). Note that the use of Peregrine requires at least 5GB RAM.
The war file and the ontology file can be requested from the authors. Peregrine is
configured by its parameters in the file WEB-INF/web.xml
. The path of the ontology file
is read from parameter ontology.file
, and the path of the properties file of the LVG
installation (part of the UMLS) is read from parameter lvg.properties.filename
.
Peregrine is only required for the automatic indexing of case definitions in the first tab of the application. Concepts can be added in the second tab without Peregrine.
Peregrine can deployed with jetty to control its memory consumptions and to separate its runtime from the other webapps by using the following command:
JETTY_VERSION=7.6.9.v20130131
UMLS_VERSION=2014AB
PEREGRINE_VERSION=2.0.2
java -Xmx5000m -jar jetty-runner-${JETTY_VERSION}.jar --port 8081 --log yyyy_mm_dd-requests.log --out yyyy_mm_dd-output.log --path UMLS${UMLS_VERSION}_ADVANCE UMLS${UMLS_VERSION}_ADVANCE\#\#${PEREGRINE_VERSION}.war
CodeMapper reads its configuration from the file
src/main/resources/code-mapper.properties (in
WEB-INF/classes
in the war file). Please edit after installation of Peregrine and
creation of the databases accordingly.
Building CodeMapper requires maven3
. Then just run
mvn package
and the war-file will be build in directory target
.
All properties required to initialise database connections and external APIs are read from a properties file.
There are three profiles, which determine the final name of the .war
file and
the properties file:
Profile | Final name | Properties | |
---|---|---|---|
dev (default) | codemappper-dev | codemapper-dev.properties | |
testing | codemapper-testing | codemapper-production.properties | |
production | codemapper | codemapper-production.properties |
While running, CodeMapper is configured to print debugging information to
${sys:catalina.base}/logs/code-mapper.log
, i.e.,
/var/log/tomcat*/code-mapper.log
.
The CodeMapper web application uses Java servlet version 3.0.1, which requires Tomcat7.
CodeMapper requires to run tomcat using Java 8, otherwise it will fail to authenticate with the UMLS API due to unsupported certificate protocols.
CodeMapper uses three databases:
- Download the UMLS from https://uts.nlm.nih.gov using your UMLS license.
- Use Metamorphosis (included in the download) to generate a subset of the UMLS.
- Load the UMLS data into the database using the generated MySQL scripts.
CodeMapper requires a database for storing the mappings, workspaces and users. The databases can be created using the script src/main/resources/user-tables.sql.
The database UMLS-ext
contains data complementary to the UMLS, such as the mapping from
Read CTv3 (RCD
) to Read-v2 (RCD2
). The original mapping files
(nhs_datamigration_22.0.0_20161001000001
) are distributed by the Health and Social Care
Information Centre at Digital NHS TRUD.
-
Table
ctv3rctmap_uk
for mapping Read-CTv3 to Read-v2 from filenhs_datamigration_22.0.0_20161001000001/Mapping Tables/Updated/Clinically Assured
CREATE TABLE `ctv3rctmap_uk_YYYMMDD` (`MAPID` varchar(38), `V2_CONCEPTID` varchar(5), `V2_TERMID` varchar(5), `CTV3_TERMID` varchar(5), `CTV3_TERMTYP` varchar(1), `CTV3_CONCEPTID` varchar(5), `USE_CTV3_TERMID` varchar(5), `STAT` varchar(1), `MAPTYP` varchar(3), `MAPSTATUS` int(11), `EFFECTIVEDATE` int(11), `ISASSURED` int(11)) ENGINE=InnoDB DEFAULT CHARSET=latin1; database=umls_ext file=ctv3rctmap_uk_YYYMMDD.txt mysqlimport --delete \ --columns=$(head -n1 "$file"|sed "s/$(printf '\t')/,/g") \ --ignore-lines=1 --lines-terminated-by='\r\n' \ --local "$database" "$file"
-
Table
Corev2
for Read-v2 labels from fileNHS-READv2-20.0.0/V2/Unified/Corev2.all
CREATE TABLE `Corev2` (`CODE` VARCHAR(5), `DESCRIPTION` VARCHAR(100), `C3` VARCHAR(100), `C4` VARCHAR(250), `C5` VARCHAR(20), `C6` INTEGER, `C7` VARCHAR(20), `C8` VARCHAR(2), `C9` VARCHAR(20), `C10` VARCHAR(2), `C11` VARCHAR(10), `C12` VARCHAR(2), `C13` VARCHAR(2)); mysqlimport --delete \ --columns=CODE,DESCRIPTION,C3,C4,C5,C6,C7,C8,C9,C10,C11,C12,C13 \ --fields-enclosed-by='"' --fields-terminated-by=',' --lines-terminated-by='\r\n' \ --local "$database" Corev2.all
Access to the external service Snowstorm and UTS are configured through the
.properties
file.
CodeMapper includes a Django web application to administer CodeMapper projects and users (see folder codemapper_manager/
).
The Python script src/main/tools/manage.py is used for managing a running CodeMapper instance. It allows to
- Create users
- Create projects
- Assign users to projects
- Copd and move mappings
Use ./manage.py --help
for a full description of functionality and options.
Several other scripts are available for post-processing mappings. Please refer to the src/main/tools/README.md for details.
Using httpie
:
NAME="..."
PASSWORD="..."
POST="--session /tmp/session --form POST"
HEADER="Content-Type:application/x-www-form-urlencoded"
eval http $POST http://advance/CodeMapper/rest/authentification/login $HEADER username==$NAME password==$PASSWORD
eval http $POST http://advance/CodeMapper/rest/code-mapper/related-concepts $HEADER cuis==C0087031 codingSystems==ICD9CM relations==CHD relations==RN