Download - Iugonet 20120710 poster
Metadata Database for Upper Atmosphere by using DSpace
Yukinobu KOYAMA*1, Shuji ABE*2, Tomoaki HORI*3, Norio UMEMURA*3, Hiroo HAYASHI*1, Yoshimasa TANAKA*4, Atsuki SHINBORI*1, Satoru UeNo*1, Yuka SATO*4, Manabu YAGI*5, Akiyo YATAGAI*1
*1 Kyoto Univ., *2 Kyushu Univ, *3 Nagoya Univ., *4 NIPR, *5 Tohoku Univ.
The 7th International Conference on Open Repositories Edinburgh, UK
July 9th – 13th, 2012
IUGONET provide the analysis software.
• Built the Metadata DB for upper atmosphere by using DSpace.• Metadata files in XML is stored as ‘content’ of DSpace.• OpenSearch has been prepared and we implemented the
procedure to connect from the data analysis software to the Metadata DB via OpenSearch interface.
Conclusion
Issues & Solution of upper atmospheric research
The integration analysis by using various kinds of observational data is necessity for investigating the mechanism of long-term variations in the upper atmosphere. However, there are no way to cross-search their DBs which are distributed over many research institutes in Japan. For resolving this inconvenience, we build the Metadata DB in order to cross-search their DBs.
References[1] SPASE 2.0: a standard data model for space physics, Todd King, James Thieman and D. Aaron Roberts, Earth Science Informatics, 1865-0473
• Metadata which includes location information (e.g., URI) of observational data is sent to the analysis software via OpenSearch.
•The analysis software downloads observational data by using the location information.
Collaboration by using OpenSearch
The OpenSearch interface is opened also to the other applications.
Metadata DB
We defined the IUGONET common metadata format which is based on the SPASE Data Model [1]. Basically the metadata includes URI of observational data, start & end time of observation, spatial coverage of observation and so forth.
Metadata Format & Customization Points
Ideally, the metadata DB system which can deal with arbitrary metadata is the best but
• default setting of DSpace is for Dublin Core metadata format,• IUGONET Common metadata format has deeper tree structure
than Dublin Core metadata format.
Example of the records in the metadatafieldregistry table
AccessInformation
Format
NumericalData
2177
SourceURLGranule2147
qualifierelementmetadata_schema_id
metadata_field_id
<Spase><NumericalData>
<AccessInformation><Format/>
</AccessInformation></NumericalData>
</Spase>
<Spase><Granule>
<Source><URL/>
</Source></Granule>
</Spase>
http://www.iugonet.org/
Customized DSpace-1.7.0
http://search.iugonet.org/iugonet/
Combine DBs on metadata level
Search of various observational data at 1 query
http://search.iugonet.org/open-search/
(1a) Query
(2a) Load
(3) Plot
(1b) Metadata
(2b) Wget
(2c) Data
Link to the Observational Data
Issues:
•Observational DBs are distributed.
•To find out & collect many kinds of data is difficult.
Hurdle of the interdisciplinary study
Global Observation Networks
Analysis Soft.
Metadata DB
Observational DBs
Metadata DB
Observational Data
Therefore we gave up to handle complete shape of our metadata. Then we customized it according to following two rules.
1.Some search targets are chosen from the metadata, and are inserted into “metadatafieldregistry” table.
2.Complete shape of our metadata is stored as ‘content’ of DSpace.