Resolve workshop 2022

Group photo

Topics

Material

Schedule

This is what we have planned but we will adapt this plan to the interests and needs of the participants as we go.

October 24 (Monday)

October 25 (Tuesday)

When you first enter MPIfR on Tuesday, sign up as a guest and then go to room 3.25.

October 26 (Wednesday)

Venue: room 0.02

October 27 (Thursday)

Venue: room 3.25

Prerequisits

Please bring a laptop (or access to a remote machine) with python3 (>=3.7) installed. We are familiar with Linux and MacOS. Windows should work as well but you will be on your own with Windows-specific problems.

Data

You will have time to try to apply resolve to your own data. Please think about a suitable data set that you can bring to the workshop. It makes sense to first choose a data set that you are well familiar with and that does not contain strange not understood effects. If you bring an interferometric non-VLBI data set from telescopes like VLA, ALMA or MeerKAT, please not that in this case the application of resolve can quickly become computationally expensive. So you have to choose a data set that is small (approx 10 Mio visibilities) such that you can work on it interactively. Additionally, it should be well-calibrated and especially well-flagged.

Software

If you like, you can already install nifty, resolve and the ehtimaging package by running e.g.:

pip3 install --upgrade ift-resolve ehtim

Note that parts of resolve are written in C++. Therefore during the installation some code needs to be compiled and this takes a minute or so. If the installation fails, please doublecheck that you have a C++17-capable compiler installed on your system.

In order to check that the installation was successful, you may run the following script.

import nifty8 as ift
import numpy as np
dom = ift.RGSpace([500, 500], harmonic=True)
HT = ift.HarmonicTransformOperator(dom)
pdom = ift.PowerSpace(dom)
PD = ift.PowerDistributor(dom, pdom)
S = ift.DiagonalOperator(PD(ift.PS_field(pdom, lambda k: 100./(20. + k**3))), sampling_dtype=float)
s = S.draw_sample()
ift.single_plot(HT(s), name="nifty_test.png")

If the resulting file nifty_test.png looks like the following, NIFTy is correctly installed.

Output of test script

Venue

Max-Planck Institute for Radio astronomy (MPIfR)

Lecturers

Registration

We have slots for contributed talks on other Bayesian imaging methods.

Please send your registration via email to web (at) philipp (minus) arras (dot) de. You will recieve a confirmation with a couple of days.

Name:
Institute:
I accept to be publicly listed as workshop participant on https://philipp-arras.de/2022bonn.html: yes/no

Talk title: (optional)
Talk abstract: (optional)

Indico website

Back to main page