1 of 21

25/08/2022

Tech Talk: Making AuScope Portal Play FAIR

Vincent Fazio

AuScope AVRESenior Engineer

CSIRO

vincent.fazio@csiro.au

ARDC is enabled by NCRIS.

2 of 21

  1. What is FAIR?

2

Guiding principles that aid in knowledge discovery

�Helping users of scientific data, both humans and machines (e.g. workflows, algorithms)

Improving discovery, access, integration and analysis

Four facets of FAIR:

  • Findable
  • Accessible
  • Interoperable
  • Reusable

3 of 21

2. Why be FAIR?

3

It is very abstract, but necessary in a practical sense, answering questions such as:

  1. How do I cite this dataset in my research paper?
  2. Who funded this research?
  3. Am I permitted to use this dataset in my project?
  4. How was this dataset created?

Accurate academic citation is vital for most researchers and must be supported by infrastructure

Project sponsors also need to be recognised

4 of 21

3. AuScope Discovery Portal

4

AuScope – Australia’s premier provider of research tools, data, analytics and support to Australia’s geoscience community

AuScope Discovery Portal is part of AuScope’s AVRE (Australian Virtual Research Environment)

Explore, display and download Australian geological data

Data sourced from:

  • Australian universities
  • Geoscience Australia
  • State Geological Surveys

Types of data:

  • Gravity, magnetism, seismic
  • Borehole mineralogy and rock properties
  • Satellite
  • Links to published reports
  • And lots more …

5 of 21

4. Is AuScope Portal FAIR Enough?

5

Challenge

FINDABLE:

  • Has DOIs for datasets in all metadata records for a dataset?
  • Rich metadata. e.g. Sponsorship information?
  • Is publicly findable?

ACCESSIBLE:

  • Is it publicly accessible?
  • Uses standardised, open, free comm protocols?

INTEROPERABLE?

  • Is metadata in a shareable, accessible and broadly understandable format?
  • Uses vocabularies that are also FAIR?

REUSABLE:

  • Does it have a copyright/licence describing usage rights?
  • Does it have provenance information?

6 of 21

5. AuScope Portal Metadata Architecture

6

NCI Geonetwork4

Geoscience Australia Geonetwork4

AuScope Portal API

server

AuScope portal in browser

AuScope

Geonetwork4

iso19139

Public datasets

Harvested via Geonetwork OGCWxS Harvester (OGC WMS GetCapabilities)

Builds a cache of catalogue dataset records (OCG CSW GetRecords)

Display geospatial

datasets and metadata

Metadata catalogues

OCG WMS server*

Custom datasets via OCG WMS GetCapabilities

* - Any of: Geoserver3, ESRI GIS Server2, NCI GSKY1

OCG WMS/WFS server*

OCG WMS/WFS server*

OCG WMS/WFS server*

OCG WMS/WFS server*

7 of 21

6. AuScope Portal modifications

7

Response

Requires:

  1. Extract metadata from GetCapabilities response for user-defined datasets
  2. Extract metadata from GetRecords response for featured datasets
  3. Display metadata for licensing, citation etc.

* - Any of: Geoserver3, ESRI GIS Server2, NCI GSKY1

8 of 21

7. Geonetwork: pros of metadata harvesting

8

Response

Harvesting metadata via Geonetwork OCGWxS harvester has some advantages:

  1. Relieves us of the burden of maintaining hundreds of metadata records by hand�
  2. Less error-prone�
  3. Most dataset changes can be quickly reflected in the catalogue

9 of 21

8. Geonetwork: cons of metadata harvesting

9

Response

Harvesting has a couple of drawbacks:

  1. These OCG web services are designed to describe data, not metadata. Apart from a reference via external URL, only a limited amount of metadata can be included in the response.�
  2. Paucity of metadata. Often the WMS/WFS servers are configured with a minimum amount of metadata. ��Some owners do respond to encouragement.

10 of 21

9. Deficiencies to be rectified

10

Challenge

This means that our metadata may be missing things such as:�

  1. Licensing, copyright, classification
  2. Provenance
  3. Project sponsorship
  4. DOIs for citation

?

11 of 21

10. One solution - Templates

11

Response

During harvesting Geonetwork can use templates to insert metadata

Some elements are copied directly from “GetCapabilities” response

TIP: Many other elements require a stub to be present in the template to trigger copying

e.g. WMS URLs

ISO19139 template

GetCapabilities

New ISO19139 record

Merge

12 of 21

11. How to use Geonetwork’s templates

12

Response

Procedure:

1. Load template

2. Edit template, add licensing details, stubs etc.

3. Select template in OCGWxS harvester configuration

13 of 21

12. Another problem

13

ISO19139 is an older standard from 2007

Challenge

Being older there are many things it cannot easily represent

e.g. project sponsor, end-user license, distributor license …

By default Geonetwork creates harvested records in ISO19139 standard

14 of 21

13. Solution: upgrade harvesting to ISO19115-3:2018

14

Response

Creation of harvested records in ISO19115-3 XML format is supported by Geonetwork’s ISO19115-3 plugin

If the template is ISO19115-3 then the created records will be ISO19115-3

ISO19115-3 template

GetCapabilities

New ISO19115-3 record

Merge

15 of 21

14. How to upgrade to ISO19115-3:2018?

15

Response

Create and configure new ISO19115-3 template before harvesting

Select new template in OCGWxS harvest config

ISO19115-3 plugin is built into Geonetwork v3.8+

16 of 21

15. Dealing with hand curated records – STEP 1

16

Challenge

How to upgrade hand curated ISO19139 records to ISO19115-3?

STEP1: Export files as XML/ZIP to local filesystem and then import them back again using XSLT

17 of 21

16. Upgrading hand curated records – STEP 2

17

Response

How to upgrade hand curated ISO19139 records to ISO19115-3?

STEP2: Re-import XML/ZIP file, use provided XSLT to convert to ISO19115-3

TIP: For Geonetwork v3.12.5/v4.2.0 and earlier the record will still have “ISO19139” in the metadata field�Update this metadata field to “ISO19115-3:2018”

18 of 21

17. Lessons Learned.

18

Response

In terms of knowing a record’s metadata standard the Geonetwork UI can be deceptive.

Check the raw XML or use the tick boxes in the editor board to confirm your record’s standard.

If you have metadata-poor sources harvesting will not allow you to completely bypass the burdens of curation.

You will still need to maintain templates.

Geonetwork is complex. It requires experimentation to find the most suitable configuration.

19 of 21

Questions from the Audience?

19

20 of 21

References.

20

OGC Server Implementations�

  1. NCI GSKY https://gsky.nci.org.au/
  2. ESRI GIS Server https://enterprise.arcgis.com/en/server/
  3. Geoserver https://geoserver.org/
  4. Geonetwork https://www.geonetwork-opensource.org/

21 of 21

Thank you!

Acknowledgements:�

* AuScope Ltd

* AuScope AVRE MAINTAIN Team

ARDC is enabled by NCRIS.