Add ability to include DC mapping row in metadata CSV
|Google Code Legacy ID:||Pull Request:|
Currently when a metadata.csv is included with a transfer package (for upload to access systems such as CONTENTdm), if the user wishes to include custom lables for better usability for collection patrons, then information must be duplicated in the CSV to follow best practice and map any available fields to Dublin Core metadata elements. As such, fields such as title, description, subject, rights, publisher, and so on are often duplicated across the CSV, slowing the workflow considerably and introducing the potential for error.
This feature would allow for a user to include a row beneath the first label row with mappings to related DC fields. The /Parts column is left blank, and the system should be able to use this blank column to determine that what follows is a mapping row, and not metadata for the transfer objects (as all CSV metadata rows require a /part column to be populated, indicating to which objects the metadata relates).
This way a user does not have to duplicate information in the CSV, but any fields that can be crosswalked from the custom labels used will be included in a recognized metadata exchange standard (i.e. DC).
See attached sample for illustration.
In future releases, it should be possible to add multiple mapping rows to other standards, such as MODS and EAD -- this would require a logic to parse the metadata elements, so that an EAD element such as <scopecontent> might be written in the CSV as: ead.scopecontent - thereby allowing the system to recognize the standard and then mapp accordingly.
Assigning to Evelyn for triage.