Due to the ongoing COVID-19 situation, this meeting will be virtual. Please do not attempt to travel to Seattle — or to anywhere else — to participate.
Is this document acceptable to the DMLT? What are the remaining open questions? How will we resolve them?
DMTN-148 is almost there suggest 2 weeks review by DMLT.
John to setup feedback system with Chris.
This should be baselined (change controlled)
Robert asks when we will start "acting on this" - e..g when could it be used for LATISS on the mountain. On going work from Andres and Merlin - where is the ingest and validate..
KT last stage getting from production system via oods to summit to be used for ISR on summit. Certified and transferred to where its needed.
Jim - Good to separate operations concerns (how its used on the mountain) from about the code and how we implement. DMTN-111 could have the summit details. Tim - no agreement on every curated calibration had class somewhere, one end - other is the certification
Headline: The DMLT agrees that the story we tell the community is that our data model is effectively two tables, and users will need to join them themselves.
General agreement about using PyVO and Pandas.
Are DIAForcedSources included?
The same considerations apply mapping DIAObject to DIAForcedSource.
Our feature computation may be based on DIASources or DIAForcedSources; a recommendation from Eric will be forthcoming.
Adding support for e.g. non-detection upper limits in feature computation is a possible, and may make the inputs to feature computation more complex. However, this should not be unmanageable.
How tightly coupled is the AP pipeline with the database? Is this a technical risk?
Reconstructing data structures from the AP pipelines based on VO interfaces would be challenging.
The details of feature computation are well abstracted and testable; they are not tightly coupled.
Plugins are implemented for feature computation below the task level; the master task takes a Pandas data frame as input.
None of these proposals are changes to previous promises made to the community.
In terms of announcements to the community, we suggest that this should be rolled into discussion of capabilities available for DP0.
Some discussion of a PST-SciCollab talk if necessary.
Eric Bellm — update time-series technote to contain a discussion of the way in which data will be presented to users.
Gregory Dubois-Felsmann — update the Science Platform design documentation to reflect that data access services should be tested with PyVO.
Kian-Tat Lim Convene a meeting with Colin, Tim, Robert, Yusra to resolve graph generation with per-dataset quantities (likely based on Consolidated DB work).
Everybody's talking about it, but what does it mean? Who will have to do what when?
Can we use this opportunity to get ahead of whatever Victor/Kevin/etc will ask for, and make sure DM comes out of the rebaselining process in good shape?