The holdings of the LBA and CAO were at very different levels of quality and therefore each had to be process separately in order to transfer the documents to a common repository. The holdings are not yet standardiz and have varying levels of metadata detail. The standardization and processing of the holdings will be carri out gradually via the new portal. [15]
[ 15 ]The metadata from the LBA’s old kleio database could only be export as a text document and had to be clean of spelling turkey phone number library errors, duplicates, and unnecessary spaces in a first pass. In particular, the identifiers of metadata fields often contain spelling errors or were duplicat (e.g., “accession number” and “accession numbers”). [16] The errors were automatically correct using Python scripts, and some metadata fields—such as “seal”—were further separat: The individual pieces of information were extract from “1 /VON URSPR.:/1/WS” (a wax seal from an original one preserv) and stor in individual metadata fields.
This allows for later filtering bas on
for example, the number of seals us or preserv, or even the type of seal. [17]
[ 16 ]Special attention was given to the fields relating to the place of issue and archives. These are stor in the document repository as separate entities to enable enrichment with authority data and avoid rundancies in data storage (for the entity model, see Section 3.1 ). Places of issue and archives were identifi semi-automatically and search engine optimization united states america enrich with GND authority data. A total of just under 370 archives and almost 1,500 locations were record (some duplicate removals are not yet complete).Description of data
The same procure is plann
for individuals (issuers, recipients, and co-seals), but has been postpon due to time constraints. To date, approximately 379 save time and money: by using artificial are record in the repository, who together issu approximately 37% of the documents issu by the LBA and CAO. [18]