Participants: 11

Laurent Michel, Jesus Salgado, Marco Molinaro, Francois Bonnarel, Janet Evans, Gerard Lemson,

Mark Cresitello-Dittmar, Tom Donaldson, Paul Harrison, Mireille Louys, Mathieu Servillat

Agenda

• Ongoing work status

• ADASS BOF

LM description of the ADASS BoF on TAP and DMS

Different areas covered from presentation, annotation, browsing hierarchical data, stroting complex objects in TAP and query languages issues

Two propotypes presented

All the discussion will be described into the ADASS proceedings

GL: Effort developed on this area could be also interesting on mapping simulations on TAP

LM: Yes, this is similar to the provenance mapping presented by FB

GL: There are java classes and it is a relational approach

PH: I am progressing along these lines on code generation. If there is a VO/DML repository, a plugin could be written. Current work is in the line of mapping data to models

LM: the conclusion is that we need to use TAP. Discussion to continue in the list

• PhotDM

Mireille presenting on the photDM 1.1

Review the VO/DML work in Modelio

And the pages with the documentation

ML showing the xml file and the html (human readable) file

New model is almost fully compatible with current serializations

GL: Some formating issues on the html with the css and the generated png not present

ML: If you provide support we could solve it

GL: Also there is a concept of preamble that could be used to be included automatically

• Meas/Coords

Mark CD: Presentation of DM use cases implementations and also some Jupyter notebooks that make use of this models

Once this is approved, no major extra steps to start standardization

JE: Is this model ready for the next step?

LM: In principle yes as this is reused. One thing to test if the model is ready for interoperability.

TD: This is something that is something to be done and doable. Problem about versioning of mapping on data model. Probably not needed to have full sync but it is sometimes confusing to find the versioning link

FB: Is there a future for transform DM?

Mark CD: It should be the next to be worked out

• Mapping Syntax, Mapping Library and Model Validation

LM: Presentation on the work done on the mapping syntax

Work documented into the ModelInstanceInVot into the IVOA github

Presentation of the new VOTable strucutre including the Resource Meta that contains pure XML with the VODML container

That includes: Report (on the mapping process), model, globals and templates

Three group of elements:

- Data Mapping

- Data Referencing

- Data Filtering

Work done based on Unit tests

One problem peding on element dm-mapping: tag. Support from schema experts requested

Python Client code being developed

From a XML DM serialization, a JSON serialization, Meas/Coords instances. PhotDM instance and other lib instances can be generated

Example with an astropy parser is able to get SkyCoords and other magnitudes from it

Live demostration of the jupyter notebook creating a model viewer and extracting metadata like Sky Coord

TD: This is fantastic. Volunteer to help on the astropy workflow to include it

MM: This is a way to prove validation and also to test inteoperability

Mark DM: This only proves validation partially, based on the point that astropy can be used but not in all the areas

LM: But this implies consistency on the model