This document is meant to guide someone in running the weekly Rt estimation pipeline from within the VAP (Virtual Analytic Platform). The main command for running the weekly pipeline (found in the Makefile) is make run-batch
. Running this will utilize a configuration file (.json) specified from within the associated blob storage account and will produce outputs in the rt-epinow2-output
Azure blob storage account.
-
VAP environment & Account
- git (
sudo apt-get install git
) - docker CLI
- gh CLI (
sudo apt-get install gh
)
- git (
-
cfa-epinow2-pipeline repository in VAP
- Navigate to where you would like to clone the repository code
- Clone the repository (
git clone https://www.github.com/cdcgov/cfa-epinow2-pipeline
ORgh auth login
and thengh repo clone cdcgov/cfa-epinow2-pipeline
)
-
Authentication to Azure To authenticate to the requisite Azure resources provide a
.env
file containing the secrets necessary for authentication.- Request access to necessary Azure credential file (.env) from any of the admins listed in the README.md
- decrypt the file (
gpg --decrypt .env
) - Place the decrypted file in your
cfa-epinow2-pipeline
directory
make config
Running this command runs code located in the CDCgov/cfa-config-generator repository. This command creates a configuration file and saves it into the appropriate azure blob storage account. If you receive an error that you do not have the necessary permissions to run this command please reach out Agastya Mondal ([email protected]) for assistance
- The following command will test your setup for using the
CFAEpiNow2Pipeline
package as well as your connection to the azure resourcesmake run CONFIG=test/test.json
This command will run the pipeline for a single state and disease locally (using the computing power of your VAP account). This will take approximately 2 minutes.
If you have succesfully setup the pre-requisites and are able to run make config
and make run CONGIF=test/test.json
you are ready to run the entire pipeline in production make run-batch
. This command will connect to Azure Batch and setup approximately 100 unique tasks that Azure Batch will run. This command is intended to close after initializing the jobs in Azure Batch. Please open Azure Batch Explorer to view the progress of these tasks.
The default container management software is setup to utilize docker. For users that are currently using podman, please adjust the variable CNTR_MGR
within the makefile prior to running any commands. Further, it will be necessary to authenticate to azure resources through podman (podman login
).