EISCAT_3D use case of data reuse Ingemar Hggstrm & EISCAT_3D CC - - PowerPoint PPT Presentation

eiscat 3d use case of data reuse ingemar h ggstr m eiscat
SMART_READER_LITE
LIVE PREVIEW

EISCAT_3D use case of data reuse Ingemar Hggstrm & EISCAT_3D CC - - PowerPoint PPT Presentation

EISCAT_3D use case of data reuse Ingemar Hggstrm & EISCAT_3D CC team 09/05/2019 1 EISCAT_3D project status Antenna, Receiver, Transmitter units ordered Land acquired (Norway ground preps, Fin/Swe survey) Transmitter control in


slide-1
SLIDE 1

1 09/05/2019

EISCAT_3D use case of data reuse Ingemar Häggström & EISCAT_3D CC team

slide-2
SLIDE 2

EISCAT_3D project status Antenna, Receiver, Transmitter units ordered Land acquired (Norway ground preps, Fin/Swe survey) Transmitter control in negotiation Site computing clusters in study (acc: GPU/FPGA)

slide-3
SLIDE 3

3

deploy and integrate necessary tools, services and infrastructures

Data management and processing

DIRAC interware

integration component

a single access point towards e-Infrastructures

EUDAT's B2 services

unify the data management

discovery system across difgerent storages

storage access management

EGI and INDIGO services

deploying the software stack on HPC/HTC systems including release management

provide secondary services for production operation

user authentication and access control

09/05/2019

EISCAT_3D Competence Centre

slide-4
SLIDE 4

4

User authentication development

EGI Checkin

QQ service for CN users in progress

Job submission

Deployed on cPouta cloud, CSC

Docker containers

Existing EISCAT user software (Octave, C)

09/05/2019

EISCAT_3D CC status

Job defjnitjon fjle

[ Executable = "run_rtg_docker.sh"; Arguments = ""; JobName = "my_job_name"; Site = "Cloud.CSC.fi"; CPUTime = 86400; InputSandbox = { "run_rtg_docker.sh", "LFN:/eiscat.se/archive/2016/... } OutputSandbox = {"output/∗"}; ]

slide-5
SLIDE 5

5

Dirac fjle catalogue

L1 & L2 data

B2safe

Processed data

L3 data (now only ionospheric data)

B2share

Search and select data

Processes

Dockers/VMs

  • GUISDAP General analysis L1/L2 → L3
  • RTG General tool for display of L2 data (L3)

09/05/2019

E3D Services

slide-6
SLIDE 6

6

Search & Select data Search & Select data

Dirac/B2 fjle catalogue

Here: eiscat portal

Select process

GUISDAP

  • Need matlab in VM
  • gfd window allows

analysis with other parameters

09/05/2019

E3D Process 1

slide-7
SLIDE 7

7

Search & Select data Select process

RTG

  • Octave/Matlab
  • Setup window

09/05/2019

E3D Process 2

slide-8
SLIDE 8

8

Plot/Reorder L1/L2 data Sets of data plot windows (movies in batch mode

09/05/2019

E3D RTG Process

EISCAT Tromsø VHF Real Time Graph https://www.eiscat.se/raw/rtg/rtg.cgi?V 1 of 4 24/08/18 09:42

slide-9
SLIDE 9

9 09/05/2019

E3D RTG Process +

Use RTG semi-fjnished data products (2 entries)

RTG only as Reader/Integrator

  • Provide own defjnition fjle (Can analyse L1 data)

Use Data in plot windows (prepared L2 spectral data)

  • Select window
  • Provide analysis script (ex Non ionospheric)
slide-10
SLIDE 10

10

specfjt.m

Needs development

May contain several fjles

Result is datafjles (or graphics)

09/05/2019

E3D RTG Process Example

%global Ana M0 M1 M2 [dum,d]=strtok(d_ExpInfo); File=fullfile(tempdir,[strtok(d) '_specfit.mat']) if exist(File) load(File) else Results=[]; Time=[]; end GG=3:66; [S,F]=acf2spec(selax.sacf(:,GG),selax.lag'); [NF,NG]=size(S); H=(selax.r0+GG*selax.dt)*.15e6; F=F'; Dl=mean(diff(selax.lag)); Df=mean((diff(F))); %moments M0=sum(S); M1=sum((F*ones(1,NG)).*S); M2=sum(((F*ones(1,NG)-ones(NF,1)*M1).^2).*S); M=[M0' M1' M2']; if 0 Model='fit_gauss'; % Model='fit_lorentz'; M0(find(M0)<0)=eps; M2(find(M2)<0)=eps; X0=[log(M0)' asinh(M1/2)' log(M2)'] O=optimset('Display','off'); end Ana=[]; for i=1:NG if 0 X=fminsearch(Model,X0(i,:),O,selax.sacf(:,GG(1)+ %% Err=norm(feval(Model,X,selax.sacf(:,GG(1)+i-1),F X([1 3])=exp(X([1 3])); X(2)=2*sinh(X(2)); Pars=[X];

slide-11
SLIDE 11

11

Setup an EGI Notebook

Using the RTG docker SW

Add a small set of data

wget from eiscat portal

Insert analysis code

First versions

Run RTG interactively

Develop/Refjne code

Store analysis script(s)

.m/.zip/.tgz

Run batched RTG process

CC docker

Big data set

09/05/2019

Use Case

slide-12
SLIDE 12

12

EGI Marketplace

Order octave

Python/Julia/Octave/R

T erminal → git, wget,...

09/05/2019

JupyterLab - Octave

slide-13
SLIDE 13

13

Setup an EGI Notebook

Using the RTG docker

Add a small set of data

wget from eiscat portal

Insert analysis script(s)

First versions

Run RTG interactively

Refjne script(s)

Store analysis script(s)

.m/.zip/.tgz

Run batched RTG process

Big data set

09/05/2019

JupyterLab - Octave

slide-14
SLIDE 14

Thank you

Any questions?