You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After iterating with the MMODA team, we have evaluated the following use-case:
Demo from Sky to dirty-image on their backend renkulab, with the according MMODA template for input and output parameters. Configurable must be the most important parameters like sky-sources, observation-time, phase-center, which telescope, field of view. To what extend they must be configurable is unclear atm (e.g. sky-sources). The following things need to be considered:
Freedom of choice for configurable parameters?
How can we ensure that these configurations don't have memory, resource and time-issues?
MMODA interface of the defined parameters?
How do we provide and ensure input parameter dependencies (E.g. frequency-channels on a survey)?
What about reproducibility and maintenance? Different Karabo-versions (breaking API), different dependency-resolving (rebuilding of environment), how much do we have to care for CSCS object storage backwards-compatibility?
That's my understanding of how we proceed from our last meeting with MMODA on Feb22. @lmachadopolettivalle and @anawas were also part of the meeting. Feel free to add or change something.
The text was updated successfully, but these errors were encountered:
Looks correct to me, thank you for the write-up @Lukas113. We agreed on starting with a MVP that has one or a few default sky models, fields to input key observation parameters (pointing phasecentre, frequency channels, observation length...), and a dropdown with a few telescope configurations (SKA-MID, SKA-LOW, MeerKAT). The demo should go from the sky model to visibilities and dirty images, and output the dirty images generated by the pipeline. We will meet again with the MMODA team in 2 weeks to reassess our progress.
@sfiruch I think we should create an i4ds (or Karabo?) group on renkulab for repository-management.
Yeah, that sounds reasonable. Can you create one for Karabo? Ideally the project wouldn't be I4DS/FHNW-specific. (Somewhat tangential: I also thought that we should move all Karabo repos into a Karabo GH org)
The initial setup is done on karabo-dirty-image-sim repository. I wasn't able to a Karabo or i4ds group, because the MMODA-bot needs the repository to live in the astronomy/mmoda group. So you have to request access to the repository if you want to contribute. The most important information about how MMODA is working can be found in their dev-guide.
After iterating with the MMODA team, we have evaluated the following use-case:
Demo from Sky to dirty-image on their backend renkulab, with the according MMODA template for input and output parameters. Configurable must be the most important parameters like sky-sources, observation-time, phase-center, which telescope, field of view. To what extend they must be configurable is unclear atm (e.g. sky-sources). The following things need to be considered:
That's my understanding of how we proceed from our last meeting with MMODA on Feb22. @lmachadopolettivalle and @anawas were also part of the meeting. Feel free to add or change something.
The text was updated successfully, but these errors were encountered: