GOTO-OBS / obs_goto

LSST stack interface for GOTO.
7 stars 3 forks source link

Hello GOTO #46

Open SimonKrughoff opened 5 years ago

SimonKrughoff commented 5 years ago

It's great to see how far this work has come. I was just looking at the talk from the LSST:UK All Hands. It looks like you are making good progress. I know there has been quite a bit of communication on https://community.lsst.org, but I'd like to offer to open up the lines of communication even a bit more. Is there anyone attending the LSST Project and Community Workshop in Tucson in August? If so, it would be great to carve out a little time to chat. If not, the Data Management group would welcome a chance to set up a virtual meeting to talk about your experiences.

I would also like to bring to your attention that we are planning on changing how obs_* packages are set up (hopefully for the better). See RFC-595 for discussion of this plan. We are also planning on moving to a new system for I/O abstraction. We understand these changes can be disruptive and would like to facilitate any changes you need to make to keep up to date.

If you have feedback that you would prefer to share privately, I'm very happy to communicate that way as well.

jrmullaney commented 5 years ago

Hi Simon,

Thank you so much for getting in touch; it's a really nice gesture. As you say, the community website has been a huge help, but we certainly welcome your offer of opening up the lines of communication even more.

Is there anyone attending the LSST Project and Community Workshop in Tucson in August?

I'm mulling this over. If we can get some funds from LSST:UK to attend, then I (and maybe Lydia) would very much like to come along. However, it's also possible that I may be on vacation at that time. I'll make some enquiries this week and hopefully I'll be able to attend. If I can't make it, I would very much like to set up a virtual meeting.

I would also like to bring to your attention that we are planning on changing how obs_* packages are set up (hopefully for the better). See RFC-595 https://jira.lsstcorp.org/browse/RFC-595 for discussion of this plan. We are also planning on moving to a new system for I/O abstraction. We understand these changes can be disruptive and would like to facilitate any changes you need to make to keep up to date.

Thanks for letting us know; it's good to be forewarned. At present, our main focus is ensuring that we can continue to make measurements using the stack on GOTO data, as Lydia needs data for her thesis. As such, we may end up continuing with the current obs* setup (and thereby sticking with the current version of the stack) for the purposes of her thesis (Lydia's got about 15 months of funding left). However, I must say I agree with your motivations for making these changes; I always felt the obs* packages contained a collection of files that, while necessary, didn't really sit cohesively together.

Despite the above caveat re: thesis, I expect to continue to develop the obsgoto package and will develop a new obs package for the new system once it is implemented (even if we delay full implementation for the sake of Lydia's thesis). I want to keep up-to-date as, beyond GOTO, I'm expecting to be closely involved in the UK's efforts for data fusion (which I'm presuming will require obs_ packages to be written for other cameras).

Quick question: Will the aforementioned changes make data fusion between LSST and other telescopes more straightforward? A major push for LSST:UK is to combine LSST data with other multiwavelength data, particularly using near-IR data as references for forced photometry (and vice-versa). However, it's not entirely clear to me how this would work, as I think we'd have to work across different obs_* packages.

If you have feedback that you would prefer to share privately, I'm very happy to communicate that way as well.

I've always found the community website broadly welcoming, and people on there respond really quickly. Unfortunately, I've been a little quiet recently as the Spring semester is very teaching-heavy for me. The teaching term's nearly finished now, and so I hope to delve back into development within the next few weeks. Expect to see a few requests for help from me on the community forum in due course (we seem to be having a problem with a memory leak which we've not pinned-down yet)! Along those lines, and while acknowledging my lack of expertise relative to the DM team, if there's anything you feel we could help with over the coming months, then do please let us know.

If I think of any more feedback during the coming days, I'll be sure to drop you a line,

Thanks again, and we're looking forward to hearing more from you and the rest of the DM team!

James

On Wed, 15 May 2019 at 22:49, Simon Krughoff notifications@github.com wrote:

It's great to see how far this work has come. I was just looking at the talk from the LSST:UK All Hands. It looks like you are making good progress. I know there has been quite a bit of communication on https://community.lsst.org, but I'd like to offer to open up the lines of communication even a bit more. Is there anyone attending the LSST Project and Community Workshop in Tucson in August? If so, it would be great to carve out a little time to chat. If not, the Data Management group would welcome a chance to set up a virtual meeting to talk about your experiences.

I would also like to bring to your attention that we are planning on changing how obs_* packages are set up (hopefully for the better). See RFC-595 https://jira.lsstcorp.org/browse/RFC-595 for discussion of this plan. We are also planning on moving to a new system for I/O abstraction. We understand these changes can be disruptive and would like to facilitate any changes you need to make to keep up to date.

If you have feedback that you would prefer to share privately, I'm very happy to communicate that way as well.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/GOTO-OBS/obs_goto/issues/46?email_source=notifications&email_token=AEWK7XPQ4ZN4CVPGVJV4GEDPVSAOXA5CNFSM4HNHIEP2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4GUA7STQ, or mute the thread https://github.com/notifications/unsubscribe-auth/AEWK7XICWVL5UZRLYZ66MJTPVSAOXANCNFSM4HNHIEPQ .

-- James Mullaney Lecturer in the Astronomy Group Department of Physics and Astronomy Hicks Building University of Sheffield S3 7RH, U.K. Phone: +44 (0) 114 222 3512 Skype: dph3jrm

SimonKrughoff commented 5 years ago

Regarding the following:

Quick question: Will the aforementioned changes make data fusion between LSST and other telescopes more straightforward? A major push for LSST:UK is to combine LSST data with other multiwavelength data, particularly using near-IR data as references for forced photometry (and vice-versa). However, it's not entirely clear to me how this would work, as I think we'd have to work across different obs_* packages.

There are many goals for the new I/O abstraction system, but a two key ones are:

  1. To make it much easier to relocate just a portion of a data repository
  2. It will be much easier to have data from multiple instruments in the same repository.

I think that will really help out your use case.