ORCID Authority allows you to link up DSpace metadata fields (added during the submission process) to a person's ORCID identifier. The main use case for this feature is to allow you to link author metadata fields to their ORCID identifier. This is a very basic ORCID integration that has existed since DSpace 5.x. |
The ORCID integration adds ORCID compatibility to the existing solutions for Authority control in DSpace. String names of authors are still being stored in DSpace metadata. The authority key field is leveraged to store a uniquely generated internal ID that links the author to more extended metadata, including the ORCID ID and alternative author names.
This extended metadata is stored and managed in a dedicated SOLR index, the DSpace authority cache.
The vision behind this project consists of the following two aspects:
Lowering the threshold to adopt ORCID for the members of the DSpace community
ORCID’s API has enabled developers across the globe to build points of integration between ORCID and third party applications. Up until today, this meant that members of the DSpace community were still required to implement front-end and back-end modifications to the DSpace source code in order to leverage these APIs. As DSpace aims to provide turnkey Institutional Repository functionality, the platform is expected to provide more functionality out of the box. Only an elite selection of members in the DSpace community has software development resources readily available to implement this kind of functionality. By contributing a solution directly to the core DSpace codebase, this threshold to adopt ORCID functionality in DSpace repositories is effectively lowered. The ultimate goal is to allow easy adoption of ORCID without customization of the DSpace software, by allowing repository administrators to enable or disable functionality by means of user friendly configuration.
Address generic use cases with appealing end user functionality
This proposal aims to provide user friendly features for both repository administrators as well as non- technical end users of the system. The addition of ORCID functionality to DSpace should not come at the cost of making the system more difficult for administrators and end users to use. Scope With this vision in mind, the project partners wanted to tackle the first phases for repository managers of existing DSpace repositories: ensuring that ORCIDs are properly associated with new works entering the system, as well as providing functionality to efficiently batch-update content already existing in the system, with unambiguous author identity information.
To enable ORCID authority control requires settings in both local.cfg and updating the registered beans in "orcid-authority-services.xml" as described below.
If you wish to enable this feature, some changes are required to the local.cfg file. The first step is to activate the authority as a valid option for authority control, this is done by adding/setting an additional plugin in the plugin.named.org.dspace.content.authority.ChoiceAuthority property. An example of this can be found below.
plugin.named.org.dspace.content.authority.ChoiceAuthority = \
org.dspace.content.authority.SolrAuthority = SolrAuthorAuthority |
The feature relies on the following configuration parameters in authority.cfg, solrauthority.cfg and orcid.cfg. To activate the default settings it suffices to remove the comment hashes ("#") for the following lines or copy them into your local.cfg. See the section at the bottom of this page what these parameters mean exactly and how you can tweak the configuration.
# This setting should already be specified in your solrauthority.cfg
solr.authority.server=${solr.server}/${solr.multicorePrefix}authority
# These settings can be found in your authority.cfg (or could be added to local.cfg)
choices.plugin.dc.contributor.author = SolrAuthorAuthority
choices.presentation.dc.contributor.author = authorLookup
authority.controlled.dc.contributor.author = true
authority.author.indexer.field.1=dc.contributor.author
|
Beginning with DSpace 7, you must specify which ORCID API you wish to use. A Client ID/Secret is also required, but can be obtained for free for the Public API: https://info.orcid.org/documentation/features/public-api/
If you are an ORCID Member Institution, you can use the Member API instead. The Member API is required for additional ORCID Integration features, but is NOT required for this basic ORCID Authority feature.
# These ORCID settings are now required for ORCID Authority # They can be found in your orcid.cfg (or can be added to local.cfg) orcid.domain-url = https://orcid.org # You can use either the Public API or Member API in this next setting orcid.api-url = https://pub.orcid.org/v3.0 # You do NOT need to pay for a Member API ID to use ORCID Authority. # Instead, you just need a Public API ID from a free ORCID account. orcid.application-client-id = MYID orcid.application-client-secret = MYSECRET |
The final part of configuration is to add the authority consumer in front of the list of event consumers (in dspace.cfg or local.cfg). Add "authority" in front of the list as displayed below.
event.dispatcher.default.consumers = authority, versioning, discovery, eperson |
You must also uncomment the ORCID beans in config/spring/api/orcid-authority-services.xml. These are commented out by default as they require setting the "orcid.*" settings described above.
Simply uncomment these settings as-is & restart Tomcat. They will pull their configs from orcid.cfg or your local.cfg.
<!-- This bean & alias are commented out by default. Simply uncomment them -->
<alias name="OrcidSource" alias="AuthoritySource"/>
<bean name="OrcidSource" class="org.dspace.authority.orcid.Orcidv3SolrAuthorityImpl">
<property name="clientId" value="${orcid.application-client-id}" />
<property name="clientSecret" value="${orcid.application-client-secret}" />
<property name="OAUTHUrl" value="${orcid.token-url}" />
<property name="orcidRestConnector" ref="orcidRestConnector"/>
</bean>
<!-- Also uncomment Orcidv3AuthorityValue in the list of supported types below.
The other settings in this AuthorityTypes bean can be left as-is. -->
<bean name="AuthorityTypes" class="org.dspace.authority.AuthorityTypes">
<property name="types">
<list>
<bean class="org.dspace.authority.orcid.Orcidv3AuthorityValue"/>
<bean class="org.dspace.authority.PersonAuthorityValue"/>
</list>
</property>
...
</bean> |
When first enabled the authority index will be empty, to populate the authority index run the following script:
[dspace]/bin/dspace index-authority |
This will iterate over every metadata under authority control and create records of them in the authority index. The metadata without an authority key will each be updated with an auto generated authority key. These will not be matched in any way with other existing records. The metadata with an authority key that does not already exist in the index will be indexed with those authority keys. The metadata with an authority key that already exist in the index will be re-indexed the same way. These records remain unchanged.
“Luyten, Bram” is present in the metadata without any authority key.
GOAL: “Luyten, Bram” gets added in the cache ONCE
All occurences of “Luyten, Bram” in the DSpace item metadata will become linked with the same generated uid.
“Snyers, Antoine” is present with authority key “u12345”
The old authority key needs to be preserved in the item metadata and duplicated in the cache.
“u12345” will be copied to the authority cache and used as the authority key there.
Item metadata already contains an author with name “Haak, Danielle” and a uid in the authority field 3dda2571-6be8-4102-a47b-5748531ae286
This uid is preserved and no new record is being created in the authority index.
Running this script again will update the index and keep the index clean. For example if an author occurs in a single item and that item is deleted the script will need to be run again to remove it from the index. When run again it will remove all records that no longer have a link to existing authors in the database.
When ORCID Authority is enabled, the Author field can be used to search entries in ORCID. Simply type in an Author name to search your locally indexed authors and authors in ORCID.
Select an author entry from the list to add that Author. The List of authors is updated as you type.
Authors that already appear somewhere in the repository are differentiated from the authors that have been retrieved from ORCID.
Not yet available in DSpace 7.x. The below screenshots are from 6.x |
In the edit metadata page, under the values for the dc.contributor.author fields, an extra line shows the author ID together with a lock icon and a Lookup button. The author ID cannot be changed manually. However the Lookup button will help you change the author name and ID at the same time.
Clicking the Lookup button brings back the Lookup User Interface. This works just the same way as in the submission forms.
Instructions on how to use the Batch CSV Editing are found on the Batch Metadata Editing documentation page.
ORCID Integration is provided through the Batch CSV Editing feature with an extra available headers "ORCID:dc.contributor.author". The usual CSV headers only contain the metadata fields: e.g. "dc.contributor.author". In addition to the traditional header, another dc.contributor.author header can be added with the "ORCID:" prefix. The values in this column are supposed to be ORCIDs.
For each of the ORCID authors a lookup will be done and their names will be added to the metadata. All the non-ORCID authors will be added as well. The authority keys and solr records are added when the reported changes are applied.
ORCID authorities not only link a digital identifier to a name. It regroups a load of metadata going from alternative names and email addresses to keywords about their works and much more. The metadata is obtained by querying the ORCID web services. In order to avoid querying the ORCID web services every time, all these related metadata is gathered in a "metadata authority cache" that DSpace can access directly.
In practice the cache is provided by an apache solr server. When a look-up is made and an author is chosen that is not yet in the cache, a record is created from an ORCID profile and added to the cache with the list of related metadata. The value of the Dublin Core metadata is based on the first and last name as they are set in the ORCID profile. The authority key for this value links directly to the solr document's id. DSpace does not provide a way to edit these records manually.
The information in the authority cache can be updated by running the following command line operation:
| Command used: | |
| Arguments | description |
| -i | update specific solr records with the given internal ids (comma-separated) |
| -h | prints this help message |
This will iterate over every solr record currently in use (unless the -i argument is provided), query the ORCID web service for the latest data and update the information in the cache. If configured, the script will also update the metadata of the items in the repository where applicable.
The configuration property can be set in config/modules/solrauthority.cfg, or overridden in your local.cfg (see Configuration Reference).
solrauthority.auto-update-items = false | true |
When set to true and this is script is run, if an authority record's information is updated the whole repository will be scanned for this authority. Every metadata field with this authority key will be updated with the value of the updated authority record.
In the Enabling the ORCID authority control section, you have been told to add this block of configuration.
For all of the configuration options described below, you can use either dspace.cfg or local.cfg. Either will work. It is possible that, when you compile your code with Maven, and you have tests enabled, your build will fail. DSpace unit tests utilize parts of dspace.cfg, and the configuration options you will utilize below are known to cause unit test errors. The easiest way to avoid this situation is to use the local.cfg file. |
solr.authority.server=${solr.server}/authority
choices.plugin.dc.contributor.author = SolrAuthorAuthority
choices.presentation.dc.contributor.author = authorLookup
authority.controlled.dc.contributor.author = true
authority.author.indexer.field.1=dc.contributor.author
# These ORCID settings are now required for ORCID Authority
orcid.domain-url = https://orcid.org
# You can use either the Public API or Member API
orcid.api-url = https://pub.orcid.org/v3.0
# You do NOT need to pay for a Member API ID to use ORCID Authority.
# Instead, you just need a Public API ID from a free ORCID account.
# https://info.orcid.org/documentation/features/public-api/
orcid.application-client-id = MYID
orcid.application-client-secret = MYSECRET |
The ORCID Integration feature is an extension on the authority control in DSpace. Most of these properties are extensively explained on the Authority Control of Metadata Values documentation page. These will be revisited but first we cover the properties that have been newly added.
solr.authority.server is the url to the solr core. Usually this would be on the solr.server next to the oai, search and statistics cores.authority.author.indexer.field.1 and the subsequent increments configure which fields will be indexed in the authority cache. However before adding extra fields into the solr cache, please read the section about Adding additional fields under ORCID.That's it for the novelties. Moving on to the generic authority control properties:
authority.controlled property every metadata field that needs to be authority controlled is configured. This involves every type of authority control, not only the fields for ORCID integration.choices.plugin should be configured for each metadata field under authority control. Setting the value on SolrAuthorAuthority tells DSpace to use the solr authority cache for this metadatafield, cfr. Storage of related metadata.choices.presention should be configured for each metadata field as well. The traditional values for this property are select|suggest|lookup. A new value, authorLookup, has been added to be used in combination with the SolrAuthorAuthority choices plugin. While the other values can still be used, the authorLookup provides a richer user interface in the form of a popup on the submission page.webui.browse.index.2 = author:metadata:dc.contributor.*,dc.creator:text should become webui.browse.index.2 = author:metadataAuthority:dc.contributor.author:authoritychoices.closed , authority.required,
authority.minconfidence .For the cache update script, one property can be set in config/modules/solrauthority.cfg:
auto-update-items = false | true |
The default value for when the property is missing is false.
The final part of configuration is to add the authority consumer in front of the list of event consumers. Add "authority" in front of the list as displayed below.
event.dispatcher.default.consumers = authority, versioning, discovery, eperson, harvester |
Without the consumer there is no automatic indexing of the authority cache and the metadata will not even have authority keys.
Changes to the configuration always require a server restart before they're in effect.
Other metadata fields besides "dc.contributor.author" can benefit from the ORCID authority control at the same time. Here is an example of how to get the same ORCID functionality for the "dc.contributor.editor" metadata field assuming that "dc.contributor.author" is already configured correctly. It can be achieved by modifying configuration files only.
First add the same configuration fields that have been added for the "dc.contributor.author"
choices.plugin.dc.contributor.editor = SolrAuthorAuthority choices.presentation.dc.contributor.editor = authorLookup authority.controlled.dc.contributor.editor = true authority.author.indexer.field.1=dc.contributor.author authority.author.indexer.field.2=dc.contributor.editor |
This is enough to get the look-up interface on the submission page and on the edit metadata page. The authority keys will be added and indexed with the information from orcid just as it happens with the Authors.
But you're not completely done yet, There is one more configuration step. Because now when adding new editors in the metadata that are not retrieved through the external look-up, their first and last name will not be displayed in the look-up interface next time you look for them.
To fix this, open the file at config/spring/api/orcid-authority-services.xml and find this spring bean:
<bean name="AuthorityTypes" class="org.dspace.authority.AuthorityTypes">
<property name="types">
<list>
<bean class="org.dspace.authority.orcid.Orcidv3AuthorityValue"/>
<bean class="org.dspace.authority.PersonAuthorityValue"/>
</list>
</property>
<property name="fieldDefaults">
<map>
<entry key="dc_contributor_author">
<bean class="org.dspace.authority.PersonAuthorityValue"/>
</entry>
</map>
</property>
</bean> |
The map inside the "fieldDefaults" property needs an additional entry for the editor field:
<entry key="dc_contributor_editor">
<bean class="org.dspace.authority.PersonAuthorityValue"/>
</entry> |
With this last change everything is set up to work correctly. The rest of this configuration file is meant for JAVA developers that wish to provide integration with other systems beside ORCID. Developers that wish to display other fields than first and last name can also have a look in that section.
Note: Each metadata field has a separate set of authority records. Authority keys are not shared between different metadata fields. E. g. multiple dc.contributor.author can have the same authority key and point to the same authority record in the cache. But when an ORCID is chosen for a dc.contributor.editor field, a separate record is made in the cache. Both records are updated from the same source and will contain the same information. The difference is that when performing a look-up of a person that has been introduced as an authority for an author field but not yet as an editor, it will show as record that is not yet present in the repository cache.
The authority cache and look-up functionality can be extended to use other sources than ORCID or to show more information in the look-up interface. However some JAVA development is necessary for this. Specific instructions can be found in the readme file of the org.dspace.authority package.
Here is a breakdown of the fields stored in the solr cache.
The system/dspace related fields are: id, field, value, deleted, creation_date, last_modified_date, authority_type.
The fields with data coming directly from ORCID are: first_name, last_name, name_variant, orcid_id, label_researcher_url, label_keyword, label_external_identifier, label_biography, label_country. The field all_labels contains all the values from the other fields starting with "label_".
There is currently no configuration to control which fields are indexed. The only way to achieve this is to modify the source code.
List of the files at work for this job: config/spring/api/orcid-authority-services.xml: OrcidSource contains the URL for orcid's REST API.
org.dspace.authority.orcid.Orcid makes the REST call
+ org.dspace.authority.orcid.xml.XMLtoBio converts the received XML to a java object (Bio).
+ org.dspace.authority.orcid.model.Bio
+ org.dspace.authority.orcid.OrcidAuthorityValue#create(org.dspace.authority.orcid.model.Bio) inserts all the values from Bio into the AuthorityValue subclass.
+ org.dspace.authority.orcid.OrcidAuthorityValue#getSolrInputDocument defines what's included in solr.
The files preceded with a '+' would be necessary to modify to add more info into the cache.
The look-up UI is currently the only place this information is sent to. However just a limited number of fields are sent. The place in the source code to modify to get more fields there is org.dspace.authority.orcid.OrcidAuthorityValue#choiceSelectMap. This is also documented in the readme of the org.dspace.authority package.
Make the same configuration step as for adding additional fields under ORCID. Currently the ORCID suggestions cannot be turned off for specific fields, that would require custom code.
As long as the metadata does not get indexed, there will be no changes. However every time any metadata of an item is modified, the metadata under authority control for that item will be re-indexed. When that happens a record will be inserted in the solr cache. That record's ID will be the authority key of the metadata. This can be done for all metadata at once with the index-authority script.
In short: authority keys that exist prior to enabling the solrauthority are kept. They just won't show in the look-up until they are indexed.
It is found in the config/spring/api/orcid-authority-services.xml configuration file. Look for the <bean name="OrcidSource">, which is initialized with a URL of http://pub.orcid.org