STOA Script Tracking for Observational Astronomy Peter Hague - - - PowerPoint PPT Presentation

stoa script tracking
SMART_READER_LITE
LIVE PREVIEW

STOA Script Tracking for Observational Astronomy Peter Hague - - - PowerPoint PPT Presentation

STOA Script Tracking for Observational Astronomy Peter Hague - University of Cambridge H2020-Astronomy ESFRI and Research Infrastructure Cluster (Grant Agreement number: 653477). H2020-Astronomy ESFRI and Research Infrastructure Cluster


slide-1
SLIDE 1

STOA – Script Tracking for Observational Astronomy

Peter Hague - University of Cambridge

1

H2020-Astronomy ESFRI and Research Infrastructure Cluster (Grant Agreement Number: 653477).

H2020-Astronomy ESFRI and Research Infrastructure Cluster (Grant Agreement number: 653477).

slide-2
SLIDE 2

My OBELICS Goals

  • WP 3.3 D-INT (Data integration) - Development of

STOA to provide workflow system for current and future projects

  • WP 3.4 D-ANA (Data analysis) – Next generation

source finding and characterisation for radio astronomy - BaSC

2

slide-3
SLIDE 3

Progress

Source matching Source finding

BaSC STOA

Paper

VO service?

Now

Lea eads s to

slide-4
SLIDE 4

BaSC - Background

4

https://www.astromatic.net/pubsvn/software/sextractor/trunk/doc/sextractor.pdf

From the SExtractor manual:

slide-5
SLIDE 5

BaSC - Background

5

Each pair of antennas gives a component

  • f the Fourier transform of the sky brightness

Use many antennas to get good coverage… And invert it to get a map of the sky

slide-6
SLIDE 6

BaSC - Background

  • Missing information causes a

complicated points spread function (PSF)

  • Two possible approaches – deal

with it as is, or try to reprocess the image to make it look friendlier

  • The latter approach exemplified

by the CLEAN algorithm

slide-7
SLIDE 7

BaSC

  • BaSC is a Bayesian method of finding the sources in

a dirty map.

  • Uses a likelihood function proven correct for the

visibilities (even though for speed it works on the map)

  • An efficient MCMC process with variable model size
  • Available at

http://www.github.com/petehague/BASC

  • Contact me if you need help!
slide-8
SLIDE 8

BaSC - Comparison

8

Dirty CLEANed

Should not be there!

slide-9
SLIDE 9

BaSC - Comparison

9

  • vs. CLEAN+SExtractor

40x difference in flux

Hague et al in preparation

Point source discrimination

slide-10
SLIDE 10

STOA - Background

  • BaSC and other projects required batch operations
  • n ALMA archive
  • STOA emerged from my efforts to streamline and

automate these operations

  • Now a standalone web application:

https://github.com/petehague/stoa

slide-11
SLIDE 11

STOA - Workflows

products.py kernel.py sourcefind.py cutouts.py CWL A B B requires

  • utput of A
slide-12
SLIDE 12

STOA - Worktables

  • Each row is an execution of the

workflow

  • Changing inputs automatically

triggers a minimal recomputation of the outputs

  • Software tracks status (e.g. if

currently shown outputs correspond to inputs or if new

  • utputs are pending)
  • Can connect worktables in

relational style; automatically triggering recomputation

* Hidden fields – tracking processing status etc. Inputs Outputs *

slide-13
SLIDE 13

STOA - Worktables

  • STOA allows access to a worktable as Owner,

Collaborator, and Reader.

  • Owner can initiate computation, collaborator can

flag rows and add comments, and reader sees worktable as a simple read only table

  • Objective is to use worktables to generate data to

be served through VO

  • Currently provides facilities for web access,

downloading as in fits format, or SAMP bridging to push to TOPCAT (a bit flaky)

slide-14
SLIDE 14

STOA - Provenance

Provenance recorded by keeping workflow with data Existing provenance standards

?

slide-15
SLIDE 15

Summary

BaSC Advanced source detection for interferometers https://www.github.com/petehague/BASC STOA Workflows + tables = worktables https://www.github.com/petehague/STOA Contact: prh44@cam.ac.uk