Skip to content

Latest commit

 

History

History
84 lines (66 loc) · 3.29 KB

2022-03-09.md

File metadata and controls

84 lines (66 loc) · 3.29 KB
Raw

2022/03/09 - Microscope development meeting

Attendees

Attending:

  • Danail Stoychev
  • David Miguel Susano Pinto
  • Francesco Reina
  • Ian Dobbie
  • Jacopo Abramo
  • Julio Mateos-Langerak
  • Nadia
  • Thomas Fish

Apologies:

Notes

  • Project mailing list. Ian has requested John Hopkins to create a mailing list as he often forgets someone to include someone when emailing. David mentioned that we already have a tag in image.sc and prefers to have a single place. Julio prefers image.sc. Others also prefer image.sc. It is possible to receive emails from it, follow specific tags with RSS, and most present already have an account on it. It was agreed that we would use image.sc instead of the mailing list (but use github issues for more code issues and development).

  • Jacopo and Francesco, PhD student and Post-doc in Christian Eggeling in Jena have been integrating Microscope into ImSwitch for an iSCAT setup. They have a working proof of concept that needs to be solidified.

    Lasers are working but are not able to transfer some configuration parameters to laser from ImSwitch.

    They have started with adding cameras but hitting trouble with ROIs. They are using an industrial model from Ximea. The width and height of the ROI are interdependent. They probably need to be checked after change. Jacopo will open an issue on Python Microscope to be followed so others can try on other Ximea cameras.

    Jacopo also wants to use freerun mode of the Ximea camera that is not currently supported. David said it probably works if they access the internals directly. Jacopo will on image.sc to discuss how.

    Jacopo also has a stage from MadCityLabs which they will try to add next.

  • When adding support for more devices, David suggested not using python bindings but directly wrapping the vendor DLL's with ctypes as this often reduces bugs and implementation issues.

    What to interface DLLs with - Ctypes - CFFI - Swig. Swig horrid. CFFI has advantge that uses real headers which might solve documentation issues, but requires distribution of headers which can lead to licensing problems. So prefer ctypes as long as it's manageable.

  • Tom reports things are working fine on his side.

  • Ian reports much less crashing on exit after Tom changes.

  • DLL load path in Windows since Python 3.8 has changed (python-microscope/microscope#235) and causes issues.

  • Julio asked about SLM in Python-Microscope (python-microscope/microscope#67, python-microscope/microscope#136 python-microscope/microscope#198, and python-microscope/microscope#199). Ian says that SLM was the prototype for Microscope device server but never ported. David suggests that we should implement minimal base device class for SLM based on it and that the SIM based functions should be a separate package. This is the same approach with microscope-aotools where Microscope only provides a simple Deformable Mirror class.

  • Danny reports issues with hardware triggers the Alpao deformable mirror. The device triggeres on every change, rasing or falling. Additionally every 131 pattern is repeated. They have a work around that trigger twice for these patterns getting the expected results. Not ideal but seems to work, as time is limited Danny was not going to raise the issue with Alpao.