SolomonShorser-OICR / cli

A common pancancer command line interface

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Overview

The Pancancer CLI is a tool that allows you to provision and manage a fleet of virutal machines in a cloud environment (such as Amazon EC2), and execute Pancancer workflows on them.

Bootstrap Install

A user executes this on a Linux VM or host and this gets them setup with Docker, a Launcher container, and the various configurations needed to parameterize the Docker-based launcher.

$ wget -qO install_bootstrap https://raw.githubusercontent.com/ICGC-TCGA-PanCancer/cli/develop/scripts/install_bootstrap && bash install_bootstrap

QuickStart guide

The Quick Start guide can be found here.

Notes for Testing Pre-Release Branches

This feature branch makes changes to the bootstrap script to include:

WORKFLOW_LISTING_URL=${workflow_listing_url}

In the ~/.pancancer/pancancer.config file on the host VM. This is further converted to ~/.pancancer/simple_pancancer_config.json within the launcher. These files need to be updated in the launcher container to contain the workflow_listing_url parameter.

{
    "aws_key": "*****",
    "aws_secret_key": "*****",
    "max_fleet_size": "5",
    "name_of_key": "my-key",
    "path_to_key": "/home/ubuntu/.ssh/my-key.pem",
    "security_group": "launch-wizard-***",
    "workflow_listing_url": "https://raw.githubusercontent.com/ICGC-TCGA-PanCancer/cli/feature/solomon_workflowlist_from_url/config/workflowlist.json"
}

You also need to update the ~/arch3/cli to be whatever branch is currently being used, e.g.:

[LAUNCHER 3.1.7] ubuntu@490276e119cf:~/arch3/cli$ git checkout feature/solomon_workflowlist_from_url
Branch feature/solomon_workflowlist_from_url set up to track remote branch feature/solomon_workflowlist_from_url from origin.
Switched to a new branch 'feature/solomon_workflowlist_from_url'

Launcher PanCancer Command

See the specification at the CLI wiki page.

About

A common pancancer command line interface

License:Apache License 2.0


Languages

Language:Python 74.4%Language:Shell 17.2%Language:HTML 8.4%