+Description of Core components for the Observation Data Model ( CoreObs or ObsCore)
++News /draft /reports
Last draft version is
WD-ObsCore-v1.0-20101207.pdf and can be found
here.
The current version of the original Tex file is attached below and currently updated following inputs during the Nara meeting.
++Goal
This model is meant to gather all queriable metadata that play a role in the discovery of observation data.
It is meant to be implemented using the Utypes derived from this model in a TAP/Schema implementation at various archives sites.
This exercise focuses on the capability to serve 80% of the data , with reduced / reasonable implementation costs at the archive side.
++ Uses cases
A list of Use- cases for data queries on all possible data types has been collected by David Schade , CVO and collaborators on behalf of the IVOA Uptake committee.
see the document
here.
++ UML Prototype Model
v0.1
This is the first version of the UML diagram describing the main components for this Observation DM.
The compressed archive file contains a bunch of webpages that describes the various packages and classes of the model.
see the archive compressed file in 7zip
here , in zip
there .
These are hyperlinked pages: to start with the top level , launch the index.html file and you will navigate in the various elements of the model.
v1.1 / sept 2010 to be uploaded soon
++ UTypes
From the UML classes and attributes names, one can derive Utypes strings as described in the current
UTYPES draft (
Utypes WD).
Utypes are built up accordingly for ObsDMCoreComponents and can be used as unique labels for each piece of information represented in the the data model. These Utype strings can be used in the TAP implementation of the model.
For each query mentionned in the Use-Case document
here there is an item in the model covering the corresponding concepts.
+ v0.11:
A set of Utypes has been identified by the Obs/TAP focus group as 'mandatory' for any service that supports an Obs/TAP discovery protocol. They are avaiblable in
Table 1.
Column-names will be mapped to these Utypes and used in
ADQL queries.
Other Utypes supported in the model and useful to identify the piece of metadata in the response of an Obs/Tap service have been listed in
Table 2.
+ v0.2:
Just one list of elements for the data model with corrected Utypes, Mandatory fields are shown in green.
see the full table
here
++ XML serialisation
The ObsCore Model may also be implemented using XML serialisation.
Here are some preliminary examples of XML documents and a first version of the XML Schema.
- Obs_4.xml xml document example
- Observation xml schema, an attempt of serializing the Observation data model in XML.
Click
here to see the validation of this example with Oxygen.
This requires upload of the updated version of the
new characterization schema which will be documented on the characterization page.
++ Registering ObsTAP services
Since ObsCore is just another table in an otherwise normal TAP service, the current notion is that support for ObsCore will be advertised via a flag in the TAP Capability extension. This is being discussed at
TAPRegExt.
++
ObsTap Discussion
Details on the model and its implementtaion have been discussed in various meetings and summarised in the notes attached below.
Summary in the
ObsTap page.
- Obs_4.xml: Xml example of observation metadata