Metadata formats

During the submission process, users fill out metadata fields which are stored as a part of the record. We are able to disseminate the submission metadata in various formats including but not limited to CMDI and oai_dc. See the full list of supported formats but note that some of the formats might not be applicable to all items. The various formats help us promote the submitted content in number of aggregators (and/or search engines).

CMDI

See the CLARIN introduction to component metadata in order to get more information about this topic.

Our current submissions are adhering to the clarin.eu:cr1:p_1403526079380 profile/schema.

However, we are supporting submissions with arbitrary CMDI metadata files that are used in OAI-PMH when the CMDI metadata profile is requested.

oai_dc

oai_dc is the format required by OAI-PMH. See the mapping section in order to understand how we map our submission to this format.

Submitted metadata

Following list enumerates the fields we ask in the submission workflow (the list is subject to sporadic changes). The metadata are submitted in English. There are subtle differences depending on the type of the resource being submitted. Not all the fields are present in all the formats. There are fields that are automatically generated (eg. human readable language names acompanying the iso codes, identifiers, other dates).

Field name	Description	Status
Type	Type of the resource: "Corpus" refers to text, speech and multimodal corpora. "Lexical Conceptual Resource" includes lexica, ontologies, dictionaries, word lists etc. "language Description" covers language models and grammars. "Technology / Tool / Service" is used for tools, systems, system components etc.	required
Title	The main title of the item.	required
Project URL	URL of resource/project related to the submitted item (eg. project webpage). Regexp controlled (starts with http/https)	regexp controlled
Demo URL	Demonstration, samples or in case of tools sample output URL. Regexp controlled (starts with http/https)	regexp controlled
Referenced by	Link to original paper that references this dataset.	regexp controlled
Date issued	The date when the submission data were issued if any e.g., 2014-01-21 or at least the year.	required
Creators	Fullnames of the creators of the item. In case of collections (eg. corpora or other large database of text) you usually want to provide the names of people involved in compiling the collection, not the individual authors. A person name is stored as surname comma any other name (eg. "Smith, John Jr.").	required multiple
Publisher	Name of the organization/entity which published any previous instance of the item, or your home institution.	required multiple
Contact person	Person to contact in case of issues with the submission. Someone able to provide information about the resource, eg. one of the creators, or the submitter. Stored as structured string containing given name, surname, email and home organization.	required multiple
Funding	Sponsors and funding that supported work described by the submission. Stored as structured string containing project name, project code, the funding organization, the type of funds (national/eu/other) and OpenAIRE identifier (which is also stored in dc.relation). Note that you either have to fill out all or none of the fields.	multiple
Description	Textual description of the submission.	required
Language	The language(s) of the main contenten of the item. Stored as ISO 639-3 code. Required for corpora, lexical conceptual resources and language descriptions.	multiple type-bind required
Subject Keywords	Keywords or phrases related to the subject of the item.	multiple required
Size	Extent of the submitted data, eg. the number of token, or number of files.	multiple
Media type	Media type of the main content of the item, eg. text or audio. Dropdown selection, required for corpora, language descriptions and lexical conceptual resources.	dropdown selection type-bind required
Detailed type	Further classification of the resource type. Dropdown selection, required for tools, language descriptions and lexical conceptual resources.	dropdown selection type-bind required
Language Dependent	Boolean value indicating whether the described tool/service is language dependent or not. Required for tools.	type-bind required
Annotation type	The type of annotation of the corpus, e.g. tokenization, POS-tagging, lemmatization, chunking, segmentation, parsing, sematic role labelling, term hood scoring.	multiple

Metadata mapping

The following tables contains the Submission to OAI:DC mapping, it also lists some of the important automatically generated fields.

Submission field	Mapped DC metadata field
Type	dc.type
Title	dc.title
Project URL	dc.source
Demo URL	not mapped
Date Issued	dc.date
Creator	dc.creator
Publisher	dc.publisher
Contact person	not mapped
Funding	not mapped
Description	dc.description
Language	dc.language
Subject Keywords	dc.subject
Size	not mapped
Media Type	not mapped
Detailed Type	not mapped

Generated DC metadata field	Description
dc.identifier	PID (currently handle) of the resource.
dc.rights	Multiple field can contain the name of the license under which the resource is distributed, the URL to the full text of the license and so called label (PUB, ACA, RES)

The following table contains the CLARIN VLO facets (concepts) mapping to DSpace metadata:

VLO facet	Mapped DSpace metadata field
Organisation	dc.publisher
Description	dc.description
Availability	dc.rights.uri
Language	dc.language.iso
License	dc.rights.uri
Name	dc.title
National project	internal mapping
Country	not mapped
Modality	not mapped
Keywords	not mapped
Subject	not mapped
Format	not mapped
Genre	not mapped
Resource type	dc.type
temporalCoverage	not mapped
projectName	not mapped

Mappings are enabled by linked concepts (CLARIN CCR) in the CMDI metadata profile (and additionally by configured XPath mappings). Click here to see a full description of all VLO facet mappings