ISAD Descriptions of Archival Records With Entity Annotation

In order to obtain structured metadata that constitute an added value for the semantic migration, elements of ISAD(G) were considered. To this end, a study was carried out on the description of existing records in Digitarq, which complies with ISAD(G) and ISAAR(CPF) archival description standards. The annotated fields are those in which the information tends to be less structured:
Context zone: Metadata about the origin and custody of the unit of description are recorded. Within this zone there are fields such as administrative/biographical history, custodial and archival history, and immediate source of acquisition or transfer.
Content and structure zone: Metadata relating to the organization and content of the unit of description are described. Here are exposed the scope and content.
Notes area: Only one element called notes where metadata can be added.
An initial sample was defined for the annotation, which included 16 documents with different subjects among them, imported from the Digitarq database. These documents are entitled:
• Mesa da Consciência e Ordens
• Provedoria de Leiria
• Coleção de forais
• Maçonaria
• Miscelâneas Manuscritas do Convento da Graça
• Confraria do Menino Deus
• Cabido da Sé de Coimbra
• Armário Jesuítico e Cartório dos Jesuítas
• Empresa Pública Jornal O Século
• Casa de Santa Iria
• Tribunal Judicial da Comarca de Oeiras
• Mosteiro de São Simão da Junqueira
• Cartório Notarial de Colares – 1.o ofício
• Paróquia de Paderne
• Conservatória do Registo Civil de Loures – Delegação de Moscavide
• Henry Burnay
• Tribunal Judicial da Comarca de Torres Vedras

Entity List:
• Temporal – Identifies a temporal element relative to the description in question, be it a concrete date or a time span. Includes eras, periods.
• Pessoa Singular – The citizen identified by a proper name. Within the scope of this work, this entity identifies an individual, by name or by the concrete appearance of the person in the text.
• Título – This entity must be noted exclusively in a relationship with a Natural Person entity, as it works as additional information for the latter. Records the person's job title, occupation, or title.
• Pessoa Coletiva – Organization or group of persons identified by a proper name and which acts, or may act, as an entity. Identifies elements in the text that refer to a set of people in its base.
• Evento – Identifies concrete occurrences such as marriages, deaths, wars, etc.
• Documento – Information recorded on a medium, regardless of its characteristics. Identifies a concrete document in the description.
• Tipologia Documental – Set of formal elements that characterize a document according to the functions for which it is intended. In the context of electronic documents, the functionally related characteristics are associated with the specific computer nature of the document. Identifies document types.
• Lugar – Identifies elements that are described as having a permanent physical presence. Identifies a place, city, country.
•  Tipo de Lugar – This entity must be annotated exclusively with the Place entity, which specifies what type of place is annotated.
• Cota – Set of letters and numbers used to classify books or publications. This type of entity only appears in one register due to its scarcity, but it was considered important to record it in a specific register.

Properties List:
• Annotatable – Document starting matrix. Indicates which parts of the document can be annotated identified by an id (s - section or subsection; h - heading; p - paragraph; f - figure)
• Anncomplete – Boolean value that indicates whether the document annotations are complete. The user marks this value through the interface.
• Sources – Array of font collections that can be linked.
• Entities – Array of annotated entities in the documents (Class id identifies the type of entity. Offsets are relative to the part of the text that is annotated and defined by its starting position in the start text and the annotated text size text.size).
• Relations – Identifies related entities where classid identifies the type of relationship. Format: partid1e_id1start,end. ConnectigWords indicates whether there is evidence to support the argument that entities can be related. It can be “null” if it doesn't exist.
• Confidence State - it can be empty (manual annotation) or have the following values: pre-added (automatic selection) or pre-removed. Who designates who took the notes. ML: (automatic annotation or pre-annotation) or user: (automatic annotation).