Automated Characterization of Health Information at Large-scale Longitudinal Evidence Systems (ACHILLES) - descriptive statistics about a OMOP CDM v4/v5 database
(Please review the Achilles Wiki for specific details for Linux)
-
Make sure you have your data in the OMOP CDM v4/v5 format (v4 link http://omop.org/cdm v5 link:http://www.ohdsi.org/web/wiki/doku.php?id=documentation:cdm).
-
Make sure that you have Java installed. If you don't have Java already intalled on your computed (on most computers it already is installed), go to java.com to get the latest version. (If you have trouble building with rJava below, be sure on Windows that your Path variable includes the path to jvm.dll (Windows Button --> type "path" --> Edit Environmental Variables --> Edit PATH variable, add to end ;C:/Program Files/Java/jre/bin/server) or wherever it is on your system.)
-
in R, use the following commands to install Achilles (if you have prior package installations of aony of these packages, you may need to first unistall them using the command remove.packages()).
install.packages("devtools")
library(devtools)
install_github("ohdsi/SqlRender")
install_github("ohdsi/DatabaseConnector")
install_github("ohdsi/Achilles")
- To run the Achilles analysis, use the following commands in R:
library(Achilles)
connectionDetails <- createConnectionDetails(dbms="redshift", server="server.com", user="secret",
password='secret', schema="cdm5_inst", port="5439")
achillesResults <- achilles(connectionDetails, cdmDatabaseSchema="cdm5_inst",
resultsDatabaseSchema="results", sourceName="My Source Name",
cdmVersion = "cdm version", vocabDatabaseSchema="vocabulary")
"cdm4_inst" cdmDatabaseSchema parmater, "results" resultsDatabaseSchema parameter, and "vocabulary" vocabDatabaseSchema are the names of the schemas holding the CDM data, targeted for result writing, and holding the Vocabulary data respectively. See the DatabaseConnector package for details on settings the connection details for your database, for example by typing
Execution of all Achilles pre-computations may take a long time. See notes.md file to find out how some analyses can be excluded to make the execution faster (excluding cost pre-computations)
?createConnectionDetails
Currently "sql server", "oracle", "postgresql", and "redshift" are supported as dbms. "cdmVersion" can be either 4 or 5.
- To use AchillesWeb to explore the Achilles statistics, you must first export the statistics to JSON files:
exportToJson(connectionDetails, cdmDatabaseSchema = "cdm4_inst", resultsDatabaseSchema = "results", outputPath = "c:/myPath/AchillesExport", cdmVersion = "cdm version", vocabDatabaseSchema = "vocabulary")
- To run only Achilles Heel (component of Achilles), use the following command:
achillesHeel(connectionDetails, cdmDatabaseSchema = "cdm4_inst", resultsDatabaseSchema = "results", cdmVersion = "cdm version", vocabDatabaseSchema = "vocabulary")
This is an alternative method for running Achilles that does not require R and Java installations, using a Docker container instead.
-
Install Docker and Docker Compose.
-
Clone this repository with git (
git clone https://github.com/OHDSI/Achilles.git
) and make it your working directory (cd Achilles
). -
Copy
env_vars.sample
toenv_vars
and fill in the variable definitions. TheACHILLES_DB_URI
should be formatted as<dbms>://<username>:<password>@<host>/<schema>
. -
Copy
docker-compose.yml.sample
todocker-compose.yml
and fill in the data output directory. -
Build the docker image with
docker-compose build
. -
Run Achilles in the background with
docker-compose run -d achilles
.
Alternatively, you can run it with one long command line, like in the following example:
docker run \
--rm \
--net=host \
-v "$(pwd)"/output:/opt/app/output \
-e ACHILLES_SOURCE=DEFAULT \
-e ACHILLES_DB_URI=postgresql://webapi:webapi@localhost:5432/ohdsi \
-e ACHILLES_CDM_SCHEMA=cdm5 \
-e ACHILLES_VOCAB_SCHEMA=cdm5 \
-e ACHILLES_RES_SCHEMA=webapi \
-e ACHILLES_CDM_VERSION=5 \
<image name>
Achilles is licensed under Apache License 2.0
Achilles is being developed in R Studio.
- This project is supported in part through the National Science Foundation grant IIS 1251151.