| layout | title | tagline |
|---|---|---|
page |
Create a Custom Application |
Clone an example app |
Here, we will take the first steps in developing our own Agave app. It is assumed that you already have an executable in mind (i.e. a program that you would like to run), and you have an idea of how you would like to run it (i.e. what are the inputs and outputs, what flags might be used, etc.).
#### Develop an app by example
Prerequisite: git should be installed on your local machine
On the SD2E GitHub pages,
you will find examples of several apps that have already
been built. Clone the entire repo using git and navigate into the reactors folder:
% git clone https://github.com/SD2E/reactors-etl
Cloning into 'reactors-etl'...
remote: Counting objects: 602, done.
remote: Compressing objects: 100% (46/46), done.
remote: Total 602 (delta 18), reused 41 (delta 12), pack-reused 541
Receiving objects: 100% (602/602), 168.23 MiB | 853.00 KiB/s, done.
Resolving deltas: 100% (253/253), done.
Checking out files: 100% (159/159), done.
% cd reactors-etl/reactors/
% ls
fcs-tasbe/ hello-container/ lcms/ sailfish/
fcs-tasbe-tacc/ kallisto/ msf/
As of this writing, there are seven example apps including hello-container,
which is both a simple example app and a template for creating new apps. Spend
some time navigating through the folders to get an idea of what components are
part of an app bundle.
#### Structure of an app
The basic structure of an app looks as follows:
hello-container/
├── Dockerfile
├── build.sh
├── deploy.sh
├── hello-container-0.1.0
│ ├── _util
│ │ └── container_exec.sh
│ ├── app.json
│ ├── app.yml
│ ├── job.json
│ ├── runner-template.sh
│ └── tester.sh
└── src
└── hello.sh
At the top level, you will find a Dockerfile and build.sh script. These are
necessary for building the container reproducibly, and having them will facilitate
future app updates. The deploy.sh script is the final script of the app building
process which helps the user add their app to the Agave SD2E tenant.
The major subdirectory is named hello-container-0.1.0/ - a concatenation of the
name of the app and the version of the executable. This directory contains both
runner-template.sh, which is a script directing how the app should be run upon
invocation, and tester.sh, which is a clone of runner-template.sh for testing
locally. Also included are template files app.json and job.json which contai
Agave-based descriptions of the app itself, and how the app should be run on an
Agave system.
Finally, the other subdirectory is a src/ directory. It contains the source
code and runtime assets (reference files, libraries, etc.) required for your app.
In this example app, the only source file is the simple hello.sh shell script.
#### Set up to create your own app
Here, we will create an app for fastq quality control using the publicly available
FastQC tool.
Copy the hello-container template app and change the name to the app you would
like to develop:
% cp -r hello-container/ fastqc/
% cd fastqc/
% mv hello-container-0.1.0 fastqc-0.11.5/
The contents of the fastqc/ app bundle should now appear as:
fastqc/
├── Dockerfile
├── build.sh
├── deploy.sh
├── fastqc-0.11.5
│ ├── _util
│ │ └── container_exec.sh
│ ├── app.json
│ ├── app.yml
│ ├── job.json
│ ├── runner-template.sh
│ └── tester.sh
└── src
└── hello.sh
From this point forward, we will refer to this location on our local machine
as ~/fastqc/. Next, we will edit the files necessary to containerize the FastQC
executable.
Proceed to Containerize your executable
Go back to Create Custom Applications
Return to the API Documentation Overview