Skip to content

2022-01-11.md

Latest commit

 

History

History
77 lines (55 loc) · 2.7 KB

2022-01-11.md

File metadata and controls

77 lines (55 loc) · 2.7 KB

2022/01/11 - Microscope development meeting

Attendees

Attending:

  • David Miguel Susano Pinto
  • Ian Dobbie
  • Julio Mateos-Langerak
  • Nicholas Hall
  • Thomas Fish

Apologies:

  • Danail Stoychev

Notes

  • First time with Tom in the meeting so introductions from everyone.

  • Thorlabs stage (python-microscope/microscope#141)

    David discussed the thorlabs stage dll updates which prevent the crashes. New version fixes the problem but Thorlabs has not yet released it.

  • Cockpit OpenGL invalid context error (cockpit#778)

    Julio has Linux installation problems. David says #778 is duplicated of #739 and that Cockpit on Debian/Ubuntu does not work with wxPython-4.1.1 so Julio needs to install 4.1.0. No available fix yet, issue is likely on wxWidgets.

  • Tom has issues with mosaic (cockpit#787)

    Discussing and debugging the issue with Tom disappearing tiles in the mosaic, we found that what was happening was the rescaling. When zomming out, the tiles disappear. These were the megatiles not being displayed. Turns out that the problem was that the display scaling meant they were displayed as white tiles (same colour as the background). Tom needs to find out why this is happening.

    There was also an OpenGL crash which David suggested could be related to MicronOxford/cockpit@c974d9a3ae441eca391acc337

    Graphics card is Nvidia Quadro P2200

    There are no issues in an older version of Cockpit. Suggested that Tom bisects git history until he finds the commit that introduced the problem.

  • automatic data processing after acquisition

    Julio wants to automate the file save, covert to ome-tif and then start a deconvolution process automatically allow online data processing.

    David suggests, outside cockpit, to monitor for new files events and trigger as script when a new file is added to the results directory. This might have issues because we open and write to the file during acquisition. So maybe write file to temp place then move to the fonal directory to trigger the file conversion and deconv. Alternatively, doing this from Cockpit, subclass the experiment and add extra stuff to the cleanup methods.

  • Hardware recommendation (cockpit#783)

    Julio wants hardware recommendations to run Cockpit. He reports issues saturating the 10Gb link when acquiring images with sCMOS at very high frame rates and plans to use ethernet bonding. Ian suggested instead to use separate ethernet interfaces, one per camera.

    Hardware recommendation for Cockpit in general is very application and device specific so something we can't do general. Suggested that we write on documentation tips and tricks (such as one camera per ethernet port) for the different cases.