Skip to content

2nd sprint

Carsten Fortmann-Grote edited this page Oct 2, 2020 · 2 revisions

Retrospective

Mads

    • A lot got done
    • understand role of libpyvinyl now
    • much work was done in parallel
    • releases took much longer than anticipated
  • -> should communicate expectations far ahead of the sprint

Juncheng

    • many things got done
    • many bugs fixed in simex although at slower pace
    • thanks to dan for docker images
    • nothing particular -> work better together, involve ELI people better

Carsten

    • finished almost all tasks
    • very dedicated, good team spirit
    • better team collaboration than last sprint
    • did not involve ELI people
    • did not advance the simex tutorial
  • -> define tasks better fitted to personal profiles

Shervin

not much to contribute since out of office most of the time

  • now working on MPI compatibility of openpmd (which was on the board)

Dan

    • good experience to be on the sprint
    • guidance went well
    • inspiring, good team spirit
    • mcstas docker image improvements expensive, so leave it for now.
    • have not yet had any feedback on simex image (latest)

Demo session

Aljosa:

  • new releases for wiser and
  • demo of oasys workflow in oasys served in VUO
  • new features in oasys:
    • find best focus
    • true focus scan (scan z)
  • collaborators can get accounts
  • openpmd validator and widget almost done

Juncheng

  • simex release for D5.2
    • branch cleanup
    • tests: fixed 9 failures, 33 errors to 4 errors and 5 skips (can tag as allowed failure)
  • demo for MS5.2
    • crystallography notebook
      • paths should be in place if conda env is selected in kernels
      • rotation: uniform vs. random
      • harmonized simulation interface
      • open simulation results in hdf viewer from wp4 ?
      • numerous analysis options directly in simex, but due to hdf format (cxi format) also wp6 analysis tools usable
    • use cases: users come from EuXFEL, estimate number of patterns to get good reconstruction. can get user testimonials
    • documentation updated, not online yet.
    • will go for versioned documentations
    • travis not working for now -> register as scientific organization with github, fw details to andy

Dan

  • docker images
    • simex: not all cmakelists have find cmake

Mads

  • demo of mcstas script for mcstas and mcxtrace on ess jhub

Carsten

  • libpyvinyl now available on pypi, tested on travis-CI and documentation on rtd.

2020-09-25

Carsten

  • almost done with basic API documentation for libpyvinyl

Mads

  • straightened out missing reqs in mcstas docker image
  • will test the new image today
  • discovered pygments needs to be in 2.6.1

Juncheng

  • simex release: v0.5.0-alpha is tagged
  • documentation needs to be on gh-pages, for now have to build manually and push to gh-pages branch
  • demo notebook from 18M review updated
  • new demo for spi

Dan

  • still problems with MKL
  • will get some working version today

Aljosa

  • will not finish validator widget
  • IO widgets are finished
  • integrated into vuo version of oasys
  • will fill in details into MS5.2 report
  • improve oasys documentation

2020-09-24

Carsten

  • reviewed raytracing extension
  • more discussion needed, shervin, mads, andrea, aljosa will schedule a meeting
  • libpyvinyl now available on pypi
  • will add documentation for libpyvinyl
  • will review MS5.2 report draft

Juncheng

  • simex prerelease 0.5
  • dan helped build a new docker image
  • make one base layer image and on top of that customized images for specific (hpc) systems
  • will try to add documentatian to libpyvinyl
  • will prepare demo notebook
  • should include jupyter for testing images

Dan

  • will assist Juncheng to get docker and demo notebooks up and running

Mads

  • cntd documentation

Shervin

  • tried to pull docker image
  • updated instructions for local installations
  • works well with conda but not without (mpi4py)
  • changes in readme should go into simex release

2020-09-23

Carsten

  • meeting with aljosa
  • fixes to pyvinyl
  • will rename pyvinyl to libpyvinyl
  • will review x-raytracing openpmd extension
  • make pypi package

Juncheng

  • fixed sdf issue in TNSACalculator
  • merged all fixes to develop
  • will fix sdf cmake file in forked sdf repo

Mads

  • updated mcstasscript to support mcxtrace
  • needs more documentation
  • ILL-ESS user meeting: andy will give talk today, link will be
  • will update milestone report
  • consider pyvinyl as possible interface to mcstas

Aljosa

  • will finish two widgets today
  • will start on validator widget that displays beam properties
  • Heard that RIs ran out of funds.
  • ESRF works on hdf5 viewer, same as h5nuvola.

2020-09-22

Carsten

  • PR to ocelot to fix observed test failure in simex tests
  • successfully tested xmdyn and xcsit calculators

Juncheng

  • observed some problems in reading geom files in
  • merge simex forks into simex dev
  • will try to fix TNSA Calculator (sdf)

Mads

  • renaming mcstas script
  • testing mcxtrace
  • setup pyvynil for pypi and add as dependency
  • name suggestion for pypi package?
  • should take shervin into account

Aljosa

  • started on openpmd widget but needs to talk to manuel: py3.8 has problems with shadow3
  • modified existing widgets but not tested

AoB

  • informal "virtual coffee meeting" at 3pm in the same zoom

2020-09-21

Carsten

  • patch genesis/ocelot
  • work on pyvinyl this week and if time permits on the source |+ prop notebook

Juncheng

  • cleanup and demo notebooks

Mads

  • continue work on MS5.2: added sections on pyvinyl
  • will try to make a pyvinyl dependend version of mcstasscript

aljosa

  • needs some discussion with luca rebuffi
  • some tweaks needed to automatically update docker images from pypi
  • will start on oasys as a service
    • openpmd io update
    • integration of services currently only possible in VUO
    • meet carsten at 8pm in our zoom room

2020-09-18 (Day 5 of the 2nd sprint)

Carsten

  • fixing tests for non-public backengines

juncheng

  • works on several remaining simex issues to be fixed

mads

  • draft for ms5.2 report
  • juncheng will move the draft to overleaf

dan

aljosa

  • finished the two features
  • docker images for oasys are available
  • new image with new features available next week

Plan for next week

  • prepare demos, if possible on the jupyter hub instance with new docker images.

2020-09-17 (Day 4 of the 2nd sprint)

Carsten

  • fixed tests, will continue with missing backengines

juncheng

  • fixed mpi issues
  • distributed remaining issues between carsten and juncheng (in issue #43)

dan

  • reduced mcstas docker image size by ~1GB
  • has a openshift instance at desy now on egi test cloud

mads

  • made demo for jupyter hub, video. data not published yet
  • wants to read up on jupyter hub deployment
  • will draft M5.2 report

2020-09-16 (Day 3 of the 2nd sprint)

shervin

  • installed simex
  • some changes to build system in PR:
    • downgrade cmake to 3.11
    • gapd not installed by default
    • will look into making a docker image, dan can help

carsten

  • crystfel updated
  • will implement one more feature in gauss photon source

juncheng

  • cleaned up all branches
  • will fix tests (locally bec. travis above size limit)
  • clean up in readme and install notes

aljosa

  • will finish two major wiser features today (focus find, focus sweep)
  • will have release ready by friday
  • pack wiser with oasys and rafec into docker container -> can then deploy all services through docker!!!

mads

  • looking into docker recent developments: howto combining docker images
  • can make template of combined docker images

2020-09-15 (Day 2 of the 2nd sprint)

Carsten

  • removed branches from simex
  • recommends to use develop branch for test installations (e.g. shervin)

Shervin

  • fights with installing simex, will try again with develop branch
  • waits for mcstas script to be released

Juncheng

  • cleaned up branches
  • will not merge gapd branch, still in development
  • plans to test example scripts
  • will assist Shervin

Mads

  • prepares mcstas script release
  • serialization (string representation) needed more work
  • Carsten recommends using a standardized data format (eg json) for storing the serialized version.
  • release is on pip
  • will demo on ess jupyter hub

2020-09-14 (2nd sprint starts)

  • We start with the tasks on D5.2
  • Shervin can start working on the metarepo
  • Dan will help getting it to work on a cloud instance.