The FASTDB workshop will be held at UIUC from Wednesday, July 16 through Saturday, July 19. The workshop will be held in-person, as we expect this to be as much if not more working session than presentation. There will be a subset of sessions that will be hybrid where we want some other people to be able to connect remotely. The agenda below will have more information as it becomes available.
Below are the goals as written when the workshop was proposed. Updated and more specific goals are below in "Agenda".
We will work out the full agenda for the workshop on the first day of the workshop, based on the interests of and input from everybody present.
Important things to cover include the following. They will be parceled out into times in the agenda after a discussion on the first day.
Basic functionality demonstration: From a loaded database, demonstrate that we can extract bulk lightcurves. Also demostrate that we can ingest and keep up with alerts during a simulated survey. Both of these things have already to some degree been done.
Spectroscopic schema and policies: Whereas lightcurves are going to come entirely from LSST, and so are well-defined, we anticipate heterogeneous spectroscopic information. What do we want to keep and track? How are we going to handle multiple different information? How are we going to track data quality? What is "the" redsdhift of a transient? Expect to implement basically none of this, but we should understand by the end of the workshop what it is we need to track, and how we're going to make deciisions.
Spectroscopic connections: spectroscopic facilities need to find out what transient objects and transient hosts we have. This includes 4MOST/TiDES, but also others. In the other direction, we need to make it as easy as possible for spectroscopic facilites to import classifications, redshifts, and spectra. Some of this exists already. At this workshop, we should make sure we understand the requirements and spec out as much of it as possible. As a stretch goal, implement and test as much as possible. Chris Frohmaier will be of particular importance for this.
Spectroscopic priorities: This is related to the previous one. The connection between FASTDB and RESSPECT. A prototype exists in the DESC TOM, which Amanda Wasserman has been using. Related to this is actually implementing RESSPECT into services that could run continuously (it currently exists only in interactive prototype form). Again, the primary goal is to make sure it's all specced out, but the stretch goal is implementation and testing.
Broker connections: a version of this already exists, but we need a number of questions answered:
Ideally, at the workshop, we'd figure out the answers to these questions, so that the work we do in the coming months will be well-grounded. A stretch goal would be to update and/or adapt the implementation that currently exists, and perhaps even to test integrating it with a sample version of a real broker (e.g. Pitt/Google, given that Chris Hernandez is present) as opposed to the currently-existing fake broker used in tests.
Skyportal/TOM connections: FASTDB itself is not integrated directly with a TOM such as the LCO TOM toolkit or Skyportal, as a direct integration was too cumbersome for the specific needs of FASTDB. However, DESC may well want to maintain something like Skyportal for tracking targets of interest. Do we want to do this? What do we want to use it for? How much should get exported from FAST to Skyportal (or whatever)? We shouldn't expect to do any implementation of this at the workshop, but it would be good to achieve clarity on the plans.
Lensed Sources: How do the FASTDB schema need to change for lensed sources? Understanding what's needed is more important at the workshop than actually doing anything.
API endpoints: An API endpoint exists for doing generic SQL queries to the Postgres backend. There are a handful of other API endpoints for pulling lightcurves of individual objects, getting lightcurves of currently active transients, and asking for, getting, and reporting on desired spectroscopic targets. What other API endpoints should exist? Defining as much as possible would be good; a stretch goal at sprint sessions would be to implement and test some of these.
Web UI: A web UI right now exists that lets you get the lightcurve of an individual object, and to search for objects based on RA/Dec. How much else do we want? What kind of search criteria do we want to have implemeted on the Web UI? Again, working out what we want is the primary goal, and implementing and testing is a secondary goal.
All times are local, i.e. CST (UTC-5h).
| 9:00AM | Welcome, logistics, (re)introductions |
|---|---|
| 9:30AM | FASTDB Overview (Rob) |
| 10:00AM | FASTDB Technical details (Rob) |
| 11:00AM | Planning the agenda for the rest of the workshop |
| 12:00PM | Lunch |
| 1:30PM | FASTDB hands-on workshop: using the client on existing installations, installing your own local development instance. Open-ended. |
| 5:30PM | End of Day |
All times are local, i.e. CST (UTC-5h).
| 9:00AM | Services using FASTDB (RESSPECT, etc) |
|---|---|
| 10:00AM | Talk to broker teams |
| 12:00PM | Lunch |
| 1:00PM | Spectral and metadata formats |
| 3:00PM | Use cases |
| 5:30PM | End of Day |
TBD.
TBD.
The closest airport to UIUC is the Willard Airport (CMI). This is a very small airport, and it may not be possible to find reasonably-priced flights into it. You may also fly into ORD in Chicago, and take a Peoria shuttle to and from UIUC. (The round trip on the shuttle should be ~$90.)
For the FASTDB workshop, it will be most convenient to stay in the same hotel as for the DESC Meeting the following week; see the logistics page for the meeting.
Note that the hotels may have different blocks for the fastdb workshop and the DESC meeting, so you may need to make two different reservations. At least for the Hampton Inn, here is the link for the fastdb event.
If you are one of the participants whose travel will be funded by DESC, you should have received (or will soon receive) an email invitation from Heather Shaughnessy. You will need this information before incurring any expenses. You will also need to submit a SLAC site access form (information about which should be in the email from Heather Shaughnessy).