Making Images using Stimela in the Cloud
This tutorial describes how to use cloud compute to make TART images. This tutorial does not require any specific software on your local computer, instead all the computation will be done in the AWS cloud. This will just require a browser, almost no network traffic and is relatively low cost. The estimated compute costs for running this tutorial are less than 0.25 CNY, (0.61 South African Rand) at current EC2 compute instance rates.
This is structured as a tutorial. I estimate that this will take a person with minimum background XXX hours.
Pre-requisites
- Create a free AWS account. This can be done here. There is a getting started guide here.
Step 1: Launch an EC2 instance
This step you will launch a free instance, and then copy some files to and from than instance. Use AI to do this from your local machine.
- Lanuch a free-tier instance (t3.micro) on AWS using the TART_Imaging instance type.
- Create a file on your local machine called
hello_world.yamlcontaining your name. - Copy
hello_world.yamlto your AWS server. (https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/linux-file-transfer-scp.html) - Copy the file back to a different location on your computer.
- Stop (don't terminate) your instance.
The stimela recipe
below is the basic stimela recipe that we will be following:
#!/usr/bin/env -S stimela run -l
# This basic TART Stimela 2.0 processing pipeline
opts:
log:
dir: logs/log-{config.run.datetime}
name: log-{info.fqname}
nest: 2
symlink: log
_include:
- (tartcargo):
- tart2ms.yml
- tart-download-data.yml
- spotless.yml
- disko-draw.yml
tart-image:
info: TART imaging using stimela
inputs:
tart:
dtype: str
required: true
info: "TART name (e.g. mu-udm)"
outputs:
ms:
dtype: MS
default: '{current.tart}.ms'
hdf:
dtype: File
default: '{current.tart}-vis.hdf'
svg:
dtype: File
default: '{current.tart}.svg'
assign:
api: 'https://api.elec.ac.nz/tart/{recipe.tart}'
steps:
download-hdf:
cab: tart-download-data
skip: =EXISTS(recipe.hdf)
params:
api: =recipe.api
vis: true
file: =recipe.hdf
create-ms:
cab: tart2ms
skip: =EXISTS(recipe.ms)
params:
hdf: =steps.download-hdf.file
ms: =recipe.ms
clobber: true
single-field: true
rephase: obs-midpoint
spotless-image:
cab: spotless
params:
ms: =recipe.ms
HDF: ='{recipe.tart}-spotless.sphere'
nvis: 10000
multimodel: true
healpix: true
fov: 170deg
res: 0.5deg
spotless-draw:
cab: disko-draw
params:
hdf: ='{recipe.tart}-spotless.sphere'
show-sources: true
scale-mad: true
SVG: ='spotless-{recipe.svg}'
Download TART images to your laptop
- Use scp to copy the SVG image from the TART to your local laptop.
Final Step: Terminate your EC2 instance.
Term