Date
Attendees
ipac: Serge Monkewitz, Tatiana Goldina, Unknown User (xiuqin), Unknown User (jar), Gregory Dubois-Felsmann, Unknown User (zhang), Loi Ly, Trey Roby
- slac: Andy Salnikov, Unknown User (danielw), Fabrice Jammes, Fritz Mueller, John Gates, Jacek Becla
- observers: Ashley
Discussion items
S15 planning
- DB plans affecting SUI: ddl, restful API, metadata+db+imageCutout all integrated
- #1 feature needed by SUI team: integrated db+set of images
- also, SUI team cares which format will be supported first, ipac table format would be good
- although it is inefficient (lots of empty spaces)
- need to discuss early how to pass metadata
- would be best to have something (even if very basic) working early in the S15 cycle
- action item for IPAC team: document queries they are planning to run
Tatiana's tests of Qserv
- 2 problems reported some time ago: DM-1822 (fixed, in review), and DM-1905 (started looking into that during the meeting)
- need to discuss at DM-AHM how to improve turn around time
Async requests
- "POST /db/v0/query?type=async" and "GET /db/v0/query/status?queryId=<qId>" look reasonable
- No need to return any estimates from POST, SUI will just call periodically "explain"
- Need to add "explain" to the API page
Terminology
- need K-T for this discussion, postponing till next time
VO cutout service
- http://www.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/caom2ops/examples
- the CADC sources are here https://code.google.com/p/opencadc/source/browse/#git%253Fstate%253Dclosed
- IPAC team familiar, but not with details, will read up about it
- also, Serge wrote image cutout code for Ursa. Discuss all that at the DM-AHM
Next meeting
- No 11:00 am hangout next week, because of AHM
-----
Keeping this for future discussion about terminology:
- dataset / repo etc
In IPAC archives, dataset always refers to a collection of data.
For WISE, each data release is a dataset: like preliminary release, allsky release, AllWISE release, ...
For Spitzer, each legacy or exploration team produce a set of data, they are re leased with each team name or the program name
Sometimes a dataset includes both catalogs and images; sometimes only catalogs or images in a dataset.
Inside a dataset, they are further separated into smaller units, sometimes by type (catlog, image, spectrum).
Images can be further separated by level of data products (level1-single frame image, leve2-coadded image), and by wavelength (wise has 4 bands).
For each image, there are also ancillary data associated with it, like artifacts, uncertainty, coverage, PSF file, mask, ...
- organization of data will affect the REST API when specifying resource. What's repo definition?