A prototype for online DAL examples and model snipets
Laurent Michel
laurent.michel at astro.unistra.fr
Fri Feb 5 17:51:06 CET 2016
Markus,
Thank you for your comments.
The scope of our prototype is to provide examples in a way independent of any specific service or implementation.
It is more a cookbook meant to give a straightforward access to individual examples or snipets in many domains, DALIs services
but also DM, ADQL, or anything else in the VO.
That fits very well with your suggestion to use it as a "continuation" of DALI examples.
Another use case is to help users to tackle with some VO aspects by providing live pointers at the end of documents. We plan to
do so in the annex of Obscore1.1.
Finally, giving a very simple way (nothing neither to parse nor to process) for software clients to get HELP input could be
valuable. That is why we have specific endpoint for each example section: you want a query, just get it as a string.
It is important to note that the service content can be extended without touching any service description.
About the implementation:
-----------------------
I implemented 2 endpoints (/show and /vosi - dali indeed) because of a very restrictive interpretation of the DALI spec which
requires that example documents "... should be viewable with 'common web browsers'. Javascript or CSS must not be necessary to
find and interpret the elements specified". I just put 2 skins on the same data, one with CSS/JS and one without. The DALI-like
pages have no references to external resources.
As you can guess, this is just a proof of concept, apart the poor layout the DALI-like output has not been validated at all.
I thank you anyway for having investigated this point.
Cheers
Laurent
Le 05/02/2016 10:20, Markus Demleitner a écrit :
> Laurent,
>
> On Tue, Feb 02, 2016 at 05:46:14PM +0100, Laurent Michel wrote:
>> In order to give some quick references to DAL query examples, we
>> have setup a service prototype available here:
>> http://saada.unistra.fr/voexamples
>>
>> This service provides a set of examples for various protocols (currently
>> Obscore and VODML). Each example (or snipet) has 4 sections (in plain text
>> right now):
>> - title
>> - description
>> - query (ADQL, XML....)
>> - usage (how to run the example).
>>
>> We identified 2 use cases:
>> 1) Human users can browse the Web site to access individual examples. In this case, all sections are gathered on a single page.
>> A keyword search may be implemented later.
>>
>> 2) For software clients, each of these items can be consumed to build up a contextual help facility.
>> Below is an example of an URL returning the "query" section of the "Composition" feature in "VODML".
>> http://saada.unistra.fr/voexamples/vosi/vodml/Composition/query
>> Whereas the following URL
>> http://saada.unistra.fr/voexamples/vosi/obscore/usecase1.1
>> returns a whole example as an XHTML page compliant with the VOSI examples specification.
>>
>> Please, try it out and contribute with your comments and feedback.
>
> Well, the original plan for DALI-examples was that the examples all
> reside on one page (or at least on very few pages, linked pages), as
> in, for instance, http://dc.g-vo.org/tap/examples. This lets a
> client (TOPCAT already does this) simply pull the whole thing in and
> build the service-provided examples (try a recent TOPCAT on GAVO's
> TAP service and check out the examples menu).
>
> Having a central page with obscore examples would be particularly
> useful with DALI's continuation property. Sites having obscore
> could then, on their examples endpoint, just include
>
> <p>See also
> <a href="http://saada.unistra.fr/voexamples/vosi/obscore"
> property="continuation">Generic examples for ivoa.obscore</a></p>
>
> and thus pull in all your examples.
>
> The intention also was that there should be no need for having
> separate files for DALI examples and "human" reading. What made you
> provide separate URLs?
>
> As to the actual examples, for instance,
> http://saada.unistra.fr/voexamples/vosi/obscore/usecase1.5 -- I'm
> missing the RDFa annotation as per DALI -- is that an oversight?
> Also, just as a matter of taste, I'd much prefer using proper markup
> rather than pre in the explanatory text. So, for that example I'd
> write something like:
>
> <body vocab="ivo://ivoa.net/std/DALI#examples">
> <div id="usecase1.5" resource="#usecase1.5" typeof="example">
> <h2 property="name">Searching Quasars using Obscore</h2>
> <p>To locate science-ready datasets having spatially resolved
> spectra of high-redshift quasars, one could constrain matches in
> <span property="table">ivoa.obscore</span> as follows:</p>
>
> <ul>
> <li>data type = IFU</li>
> <li>data quality: fully calibrated</li>
> <li>object class: quasar</li>
> <li>redshift > 3</li>
> <li>radio flux > 1 mJy</li>
> </li>
>
> <p>In a first step, you would get a set of appropriate positions
> from an appropriate service. A good source, obviously, is SIMBAD
> (TAP access URL
> <code>http://simbad.u-strasbg.fr:80/simbad/sim-tap</code>).
> There, you could execute
> <pre>
> SELECT target_name AS name,coordsys,ra,dec FROM SomeObjectCatalog
> WHERE target_class = 'quasar'
> AND redshift >= 3.0
> AND flux_radio > 1.0</pre>
>
> <p>On this obscore service, you would then run:</p>
>
> <pre property="query">
> SELECT q.Name,q.ra_RAJ2000,q.dec_DEJ2000,o.*
> FROM TAP_UPLOAD.t1 AS q JOIN ivoa.ObsCore AS o
> ON CONTAINS(POINT('ICRS',q.ra_RAJ2000, q.dec_DEJ2000), o.s_region) = 1
> WHERE o.dataproduct_type = 'cube'
> AND o.calib_level >= 2
> AND (o.em_max - o.em_min)/o.em_res_power > 100
> </pre>
>
> <p>-- where you will have to adapt "TAP_UPLOAD.t1" to the conventions
> and the state of your TAP client.</p>
>
> </div>
> </body>
>
> I'd maintain that's much more human- *and* machine readable. The
> "table" and "query" properties used here come for the TAP
> implementation note and will be documented in TAP 1.1 (Pat -- is that
> true?).
>
> A good way to see if your document is machine-readable is to let a
> machine read it; you can use TOPCAT for that for TAP examples.
> Generically, the W3C RDFa validator is a good choice:
>
> http://www.w3.org/2012/pyRdfa/Validator.html
>
> If you run it on http://dc.g-vo.org/tap/examples you'll see the
> machine can actually figure out the various examples given there.
>
> Incidentally, don't worry about the warning about "Unusual URI
> scheme" that that validator will give for DALI examples documents --
> I'm afraid it'll be some time until ivo: won't be considered an
> unusual URI scheme any more :-)
>
> Cheers,
>
> Markus
>
--
jesuischarlie
Laurent Michel
SSC XMM-Newton
Tél : +33 (0)3 68 85 24 37
Fax : +33 (0)3 )3 68 85 24 32
laurent.michel at astro.unistra.fr <mailto:laurent.michel at astro.unistra.fr>
Université de Strasbourg <http://www.unistra.fr>
Observatoire Astronomique
11 Rue de l'Université
F - 67200 Strasbourg
http://amwdb.u-strasbg.fr/HighEnergy/spip.php?rubrique34
More information about the dal
mailing list