Skip to content
This repository has been archived by the owner on Oct 12, 2023. It is now read-only.

Latest commit

 

History

History
169 lines (107 loc) · 5.91 KB

README.md

File metadata and controls

169 lines (107 loc) · 5.91 KB
Raw

REDI2 v1.4.1

REDCap Electronic Data - I (Ingester/Integrator/Importer) 2

REDI2 is a suite of tools designed to transform the CSV data and load it into the fields and events of a REDCap project. REDI2 excels at transforming large, longitudinal data sets and loading them into REDCap in a scripted, repeatable fashion. REDI2 is fast and has minimal effects on REDCap server load.

Meet the parts of redi2

data_flow

redi2 is really just a collection of tools and a few scripts which help you get set up.

Claw

Claw grabs files over sftp

Auditor

Auditor redacts, maps, cleans and basically does whatever to the output of claw which should be a CSV.

Optimus

Optimus takes your particular CSV and your project specific rules and transforms the data into a form which should be easy to make right with redcap. Here is when branching logic should be applied and any derived data which can be inferred, but is not explicitly present in your raw CSV should be made.

Lineman

Lineman is the first tool to talk to redcap. It grabs information related to the data you are passing in and validates that it can be imported. It will change the fields that it must in order to make sure that it will go in. For example, it can take dates and map them to the correct event in the redcap.

Pigeon

Pigeon carries the data to redcap. This is the last step and the only one where data will change on the redcap server. This tool takes a greedy approach, so it will attempt to push in all that it can at once. It can recover from errors and will keep trying to push the data in till it gives up and pushes in one record at a time.

Vagrant testing

RED-I2 can be tested on a vagrant. In order to do so the following steps need to be completed in order.

  • clone the repo
  • go into the redi2/vagrant/vagrant directory
  • get a copy of redcap version 6.16.8 zip file and place in the redi2/vagrant/vagrant directory with the name redcap6.16.8.zip
  • rename the file rename_to_dot_env to .env in the same directory $ cp rename_to_dot_env .env
  • run vagrant up. (Make sure to install the plugins in the plugins.txt file)
  • open your browser to http://redi2.dev/redcap/ -- go to Projects > HCV Target > API, take note of your token
  • run vagrant ssh from the vagrant/vagrant directory and go to ~/redi2/synthetic_data
  • edit the synth.pigeon.conf.yaml and insert the token from before
  • run bash prep_synthetic_data.sh if it fails then your install went wrong or more likely you dont have the right token

Now if you check your redcap you should see that there are four subjects entered. These are used in the full synthetic data run that we are about to do. Go back the terminal which is logged into the vagrant and do the following:

  • cd ~
  • cd redi2/synthetic_data
  • In the synth.optimus.conf.yaml, synth.lineman.conf.yaml change the token to the one in the synth.pigeon.conf.yaml
  • run bash synth_run_prepare.sh
  • cd ~/redi2/redi2/NEW_SITE
  • run bash synth.run.sh
  • go look at your redcap and see the data that has been added. Additionally check the NEW_SITE/data folder for the files generated during the run

Deploying RED-I 2

There are two strategies for deployment depending on the what the deploy target looks like. If you are able to talk to both github and PyPi on the hosting system then follow the first install. If your system is more locked down and you cannot talk to either github or PyPi then follow install 2.

Install 1

What follows are the requirements and steps to install redi2 with access to both github and PyPi

Requirements

  • git
  • python3
  • virtualenv
  • access to github
  • access to PyPi

Steps

Clone this repo and enter it.

Run bash install_normal.sh.

Edit the NEW_SITE directory in the redi2 directory to have the right name.

Build your configs in the configs directory

Set the run.sh script to run when you want it to.

Install 2

This install is for those who have a target system that is not able to call out to github or PyPi for whatever reason. The general strategy is to do as much work in an environment like the one to which you will be deploying and bringing all that stuff over.

Requirements

System 1: talks to outside world system
  • git
  • python3
  • virtualenv
  • tar
  • access to github
  • access to PyPi
System 2: target system
  • ssh access
  • python3
  • virtualenv

Vagrant Steps

By using vagrant up, one gets a copy of the redi2.tar.gz in repo/vagrant/redi2_deploy_tar that is used in the manual steps. It will be build on the debian jessie vagrant and bundled from there. If you want to skip the steps that take place on System 1 then this is what you should do.

Host machine

cd redi2/vagrant/vagrant

vagrant up

Now go to the "Host machine" section of the "Manual Steps" and continue from there using the tar in redi2/vagrant/redi2_deploy_tar

Manual Steps

System 1

Clone this repo and enter it on System 1.

cd redi2/build_scripts

Run bash package.sh.

Return to the directory where redi2 was cloned.

Host machine

Use scp to get the redi2.tar.gz directory to System 2. It will be in the directory above.

Use ssh to gain access to System 2.

System 2

Run tar -xzf redi2.tar.gz to extract the directory in the location that you want to install redi2

cd redi2/deploy_scripts

Run bash install_two_remote.sh

Edit the NEW_SITE directory in the redi2 directory to have the right name.

Build your configs in the configs directory

Add source ../correct_path.env to the line right below the virtualenv source line in the run.sh

Set the run.sh script to run when you want it to.

Putting it all together

Once one has built their configs all that remains is to run the tools in the right order. Pretty easy! Happy redcapping!