*IVOA DAL WG Running Meeting #8 *Wednesday 24 February 2021 - 20:00 UTC - vconf Participants: (11) Marco Molinaro, Laurent Michel, Pat Dowler, François Bonnarel, James Dempsey, Markus Demleitner, Brent Miszalski, Brian Major, Grégory Mantelet, Dave Morris, Tom Donaldson Agenda: 1. SODA progress 2. dal_usecases repository 3. ADQL, UDF Catalogue & DALI path to REC 4. (AAO Data Central) Optical Data Center SSA service *Notes: *1. SODA progress PD: cfitsio-style pixel cutout, implemented in CADC, moving it to start standardisation. Not sure if it's sufficiently general for non-FITS files. Can request range of pixels, negative to flip an axis. Multi-extension FITS is one of the concerns, may not be applicable to other file structures. MD: interoperable pixel cutout should follow DALI pattern, the addition of cfitsio pattern is harmless but pixel ranges should be enough. MD: the cfitsio solution could be part of a useful Note. FB: show it up, the community (e.g. Vera Rubin) might be interested. JD: Will add the CASDA syntax for pixel cutouts to the issue FB: introducing MOC in SODA. Noticed that current MOC misses ASCII syntax for time axis. MD: how would cutout using MOCs will look like? FB: there could be a response with empty values outside the selected region. MD: Sounds complex - have there been requests for his? FB: Some disucssion in last interop, also interest in ST-MOC PD: DALI-1.2 adding a moc xtype was mainly to have spatial MOCs only with ASCII serialisation. The idea is to have a different xtype and serialisation for MOC-2.0. [we need to ask MOC authors about this] PD: If need to have a binary serialisation of the ST-MOC, would just need we have to have a type of doubles rather than char. Problematic if they pick something that cannot be represented in a VO table cell though. FB: HiPS-to-FITS inclusion in SODA (includes rebinning and/or reprojecting). Force parameters to match results with requested resolution. Same for rotation and projection dedicated parameters, via WCS-like parameters or WCS posted directly. Issue #4 and associated PR. FB: Problems with the bibliography in builds. PD: The bibliography needs to be fixed - intent was that FB would address this. Once done we will have a working build for SODA 1.0 FB: issue on datatype change in the response (e.g. a cutout of a cube could be an image) [CASDA, JD] has two SODA on this: extracting cube and extracting a spectrum. Might be (PD) something like a rebinning in producing the spectrum, not exactly a cutout. [issue #7] JD: WIll add CASDA syntax and behaviour to the issue for information TD: cutouts around central points, given a number of pixels. (e.g. 10x10 pixels around some point) PD: central point in sky or pixel coordinates? TD: sky coords (taken, e.g., from a catalogue) PD: could bring to different radiuses from a single pixel "distance". TD: idea is to avoid remapping, using the natural orientation of the image. PD: this works roughly with circles (due to simmetry) *2. dal_usecases repository https://github.com/ivoa/dal-usecases Following the lead of DM and dm_usecases repo, trying to fill in the need for examples about DataLink templating. Would be used for DAL protocols and parts of them to provide examples. Currently available the use case about templating in DataLink. FB:(tech question) do we ned another level in the hierarchy: repo/DataLink/templating? LM: having multiple levels could make it less readable/more confusing. Use the README to specify the content. JD: would grouping files from same examples will be better? MD: in favour of a less nested solution. DM: think about how to refer to these from the standards when using them or solving them. MD: better put them inline in the standard rather than link a github repo part. DM: agree about the standards, but in conversation about the standards it would be useful to link to these. MM: Should ADQL examples be merged into this new repository? DM: They are a bit different - could be picked up and moved forward but they are probably fine where they are. *3. ADQL, UDF, DALI quick recap on chances to move to RFC MM: Should we open these up for a DAL RFC before going forward with a formal RFC to the the wider TCG. MM: GM had some small remaining changes to be done for ADQL. MM: Mainly up to the editors DM: +1 for the DAL review to check cross references between them PD: Noticed some todos in the ADQL text for xtypes - volunteered to resolve these (still a TBD under interval, I will think about and propose some text) PD: Would be nice to aim to get these ready in time for May interop if feasible GM: One major issue to resolve - casting datatypes - working to finish this. Planning to merge propsed fix in the next couple of days. Then will need to reread in detail to check readiness before starting RFC. DM did some work on the datatype sections - this may have resolved the TODOs that Pat was talking about MM: Lots of work has been done in the standards in github - would be good to raise the profile of this so the community is aware of the activity. MM: Will push for UDF catalogue to be endorsed at the next TCG meeting GM: Will look at ADQL types GM: Might put aside some types such as CLOB, BLOB and interval as they aren't currently used in ADQL MD: +1 *4. Optical Data Center SSA service (could be a quick intro in view of a more complete overview in March's running meeting) BM: Links https://docs.datacentral.org.au/reference/services/simple-spectral-access-ssa-service/ https://docs.datacentral.org.au/help-center/examples/simple-spectral-access-ssa-examples/ BM: New SSA service to allow access to the AAT spectra in the optical data centre BM: Implemented in python and django BM: Formats are complex, so have converters using specutils to convert to a standard spectra format as well as links to the original formats BM: Applied a 200k spectra per query limit BM: Can be used with pyvo so can pull out lots of spectra, useful for the time series spectra BM: Have been using the HIPS2FITS and Data Central SIA2 service to produce plots featuring image cutouts BM: Have Data Central API to query GALAH, so could build an interactive tool to query spectra for a source BM: Have converters to pull out spectra details including ObsCore parameters