Transcript Interoperability of astronomy data bases Françoise Genova, CDS

Interoperability of
astronomy data bases
Françoise Genova, CDS
F. Genova, VO as a Data Grid, 2003/06/30
1
Astronomy
• A small discipline
• Few commercial constraints
• A long term partnership to define exchange
standards and links
FITS: early data exchange format
on-line information in astronomy:
from observations to results
F. Genova, VO as a Data Grid, 2003/06/30
2
Object name
>> position
An early example
of interoprerability:
Name resolver
F. Genova, VO as a Data Grid, 2003/06/30
3
Evolving world
- Web Services
- Very Large
Surveys
Simbad Name
Resolver
>>
Sesame Web
Service
F. Genova, VO as a Data Grid, 2003/06/30
4
Web Services
as building
blocks
F. Genova, VO as a Data Grid, 2003/06/30
5
The astronomy bibliographic network
Links between
–
–
–
–
electronic journals
the ADS bibliographic database
on-line services (SIMBAD, NED)
… archival data/published results
bibcode
e.g. 1999A&A...351.1003G
F. Genova, VO as a Data Grid, 2003/06/30
6
References in an
on-line article
 links to ADS
 quality check
F. Genova, VO as a Data Grid, 2003/06/30
7
ADS
- Abstracts
- Scanned articles
- Links to original
on-line paper
- Links to other,
distributed information
e.g. original
observations
- Also-read articles
F. Genova, VO as a Data Grid, 2003/06/30
8
ADS
From bibliography
To the original
observations
Search an
author’s
publications
F. Genova, VO as a Data Grid, 2003/06/30
9
F. Genova, VO as a Data Grid, 2003/06/30
10
F. Genova, VO as a Data Grid, 2003/06/30
11
HST archive
From observation
To publication
F. Genova, VO as a Data Grid, 2003/06/30
12
A&A on line
From publication
to SIMBAD
and more
F. Genova, VO as a Data Grid, 2003/06/30
13
Simbad
Information
About the object
F. Genova, VO as a Data Grid, 2003/06/30
14
Links
To high energy
observations in
HEASARC,
To Catalogues
F. Genova, VO as a Data Grid, 2003/06/30
15
Links from
object names
in journals:
an early example:
IBVS
F. Genova, VO as a Data Grid, 2003/06/30
16
Links to
images
and data
F. Genova, VO as a Data Grid, 2003/06/30
17
GCVS
F. Genova, VO as a Data Grid, 2003/06/30
18
Lessons learnt
• De facto standard
SIMBAD/NED, ADS, journals, archives
• Cooperation between all the actors
+ snowball effect
Journals, ADS, data centres, archive centres
• The community is trained (everyday tools)
• Easy-to-build link
but contents / validation are fundamental
the role of experts remains fundamental for
building value-added services
F. Genova, VO as a Data Grid, 2003/06/30
19
Data federation
Tabular data in astronomy
A common description for tabular data
–
–
–
–
Reference catalogues
Published tables
Surveys
Catalogues of observations in archives
ReadMe
physical organization
F. Genova, VO as a Data Grid, 2003/06/30

contents
20
ReadMe
F. Genova, VO as a Data Grid, 2003/06/30
21
Tables published
in articles are
usable data!
F. Genova, VO as a Data Grid, 2003/06/30
22
Links to observational
data (images, spectra,
time series),
often distributed
in observatory
archives
F. Genova, VO as a Data Grid, 2003/06/30
23
An homogenized view of
heterogeneous information
With links to data
F. Genova, VO as a Data Grid, 2003/06/30
24
Towards data integration
• XML
• VOTable > Roy’s talk
F. Genova, VO as a Data Grid, 2003/06/30
25
HEARSARC
Browse
Astrores
in action!
F. Genova, VO as a Data Grid, 2003/06/30
26
With a touch of GLU
Générateur de Liens Uniformes
• Resource dictionary (shared, distributed,
hierarchical name space, clone
management…), knows the query syntax
• Resolver: symbolic name > URL
• An early registry prototype with many of
the required functionalities (conversions,
failure tests, …)
F. Genova, VO as a Data Grid, 2003/06/30
27
Observations
from European
archives
Inclusive:
NASA missions
HST, Chandra,
surveys, …
F. Genova, VO as a Data Grid, 2003/06/30
28
Glu Tag
Resolver
Symbolic name
>>
URL
F. Genova, VO as a Data Grid, 2003/06/30
29
Data mining: the Uniform Content
Descriptors
• A set of UCDs has been first developed in the
frame of the ESO/CDS Data Mining project to
describe VizieR catalogue columns (100.000
columns, 1.300 UCDs)
• Name for concept
• Assigned semi-automatically using column label,
description and unit
• Used e.g. to check the coherence of information in
tables
F. Genova, VO as a Data Grid, 2003/06/30
30
UCD browser
F. Genova, VO as a Data Grid, 2003/06/30
31
UCD+units: conversion, selection
Select catalogues & target
Join on UCDs :
F. Genova, VO as a Data Grid, 2003/06/30
32
Colour excess
characterizes
carbon stars
F. Genova, VO as a Data Grid, 2003/06/30
33
F. Genova, VO as a Data Grid, 2003/06/30
34
http://cdsweb.u-strasbg.fr/UCD/
F. Genova, VO as a Data Grid, 2003/06/30
35
Extension of UCDs
• SDSS: 1.500 columns
– Manual verification needed
– Very few additional UCDs
• On-going:
– UCDs for VOX
– UCDs for IDHA data model
Add ‘Metadata’ branch
… FITS keywords are often not accurate
F. Genova, VO as a Data Grid, 2003/06/30
36
Validation, maintenance
• A new structure
• Prune the existing UCD tree
• UCD steering group
– Proposal in October 2003
– UCD V1.0
– Evolution mechanism
F. Genova, VO as a Data Grid, 2003/06/30
37
F. Genova, VO as a Data Grid, 2003/06/30
38
The future
• Other knowledge basis
– List of object names
– Journal keywords
– Thesaurus (built by librarians)…
Astronomy described with different points
of view
Converging towards an ontology of
astronomy??
F. Genova, VO as a Data Grid, 2003/06/30
39