2016_08_26 visualization meeting


(Jon Thaler) #1

Bluejeans: 107541143

Agenda:

  • Report from BNL meeting (maybe)
  • Firefly Javascript API.
  • Near-term UIUC plans (see notes).

Notes:

  • Some comments from the BNL meeting:

  • BNL reception was generally positive.

  • BNL has full raft electronics, but not enough CCDs, yet.

  • FE needs to tell BE which image is being processed. This is being implemented. The image ID will be included as a command parameter. THis allowes the user to specify which image to process, if more than one is being displayed.

  • Need to pass displayed pixel coordinates as well. In case of multiple coordinate systems (rotations, etc.) the BE needs to know about displayed pixels.

  • For functions on small regions, it may not be worth reading in the entire image.

  • @tony_johnson will work on the client/server architecture for the image feed.

  • Action: We’ll want to display with and w/o overscan. This was implemented by @weiren2 when he was building the diaplayable image. Now, it’s a job for @tony_johnson.

  • SLAC installation is in progress.

  • A coordinate system complication. In focal plane coordinates, x is the vertical direction. In order to avoid excessive confusion (It’s already unpleasant …), we should probably use the same convention for CCD and raft displays. A topic for further discussion.

  • JS API documentation will be available in a week or so.

  • The IPAC/SLAC meeting will be late September (October didn’t work).

  • Near-term UIUC development plans:

  • Action: @weiren2: Modify JSON file to accommodate gain and offset scaling by segment.

  • Action: Histograms (e.g., pixel distribution along a line).
    @jbpagliuco: Front end.
    @weiren2: Back end.

  • Action: @jbpagliuco: Commands driven by cursor location.

  • Action: @ywang: Image math (sum, difference, ratio) with display of result.

  • @jjt will be gone next week (September 2) – no meeting.


(Tony Johnson) #2

Hi, I am afraid I will not be able to join the meeting this week, as I will be driving to Newark at that time with Mike Huffer. We did succeed in getting firefly installed at BNL. It took most of the week to get the test stand itself working, so we did not get as much time to demo firefly and get feedback as we had hoped, but we did succeed in using CCD emulators to generate a full raft image through camera electronics, DAQ, CCS and assembled into a mosaic which we could view with firefly (see attached). We will now work on setting up the same thing at SLAC and getting more feedback from the BNL team.

Some issues which came up during the install:

  • There was some issue with getting firefly to work with tomcat 8, to do with duplicate classes and class loading order. With help from the IPAC team we were able to work around it, although a permanent fix would still be good.
  • We found some places where there were hardwired paths when attempting to use the python code. I think Stuart knows more of the details. We need to make sure that all paths are configurable, or decided at runtime.
  • I think the IP address where the camera code is looking for the visualization server may also be hardwired (or I do not know how to configure it)

I am sure we will get much more feedback over the next few weeks as more people start to use it.


(Xiuqin Wu) #3

I would like us to start planning to migrate the camera visualization code to use the new JS/React/Redux based Firefly. I am willing to support the effort with some IPAC people working closely with UIUC team. What do you all think?


(Jon Thaler) #4

This is a good idea. I will be gone until September 8. Let’s plan to discuss this at the September 9 meeting.


(Tatiana Goldina) #5

I have put the new Firefly API links into a separate topic. Everyone is welcome to check it out. Forgot to mention, that master branch matches the release version.