Yay to more frequent downloads... now we just need to be ready to act on them locally. :)
The efforts to make the files more legible are helpful too.
Along with more frequent downloads, are you considering adding some/all of the smaller vocabularies as downloads? I don't know if others have a need to have these as downloads, or if because we're trying to provide normalized lookup services for these datasets (and others) via QA (https://github.com/samvera/questioning_authority) our needs are unique.
Thanks for progressing id.loc.gov,
On 9/18/19, 12:19 PM, "LC Linked Data Service Discussion List on behalf of Matt Miller" <[log in to unmask] on behalf of [log in to unmask]> wrote:
We are testing a new bulk export process for LCSH and would like to hear any feedback from anyone who uses the bulk downloads. The new bulk files can be found http://id.loc.gov/download/ with the titles LC Subject Headings (LCSH) *NEW Pilot*
- New compacted JSON-LD serialization
- The JSON-LD and XML files are now newline delimited meaning each line in the file is a completely self-contained record
- There are void files for each download with the date the export was created, title, description and MD5 hash of the unzipped download.
- The N-Triple file has records separators now as comments, each group of triples start with “# Start of sh12345678”
- Increased updated frequency, should be updated when new LCSH updates are released monthly.
- The new LCSH export will contain the same data as before, including broader and narrower relationship but is slightly more verbose.
- It is available in MADSRDF, SKOS and both combined MADSRDF and SKOS in all serializations.
Thinking of removing:
- The current XML dump is one large XML file, the new XML is each record as RDF XML on its own individual line. The current XML file could be used for bulk loading into a triple store, but the current and future NT file could be used in the same way. Is anyone using the current XML dump file for bulk loading?
- The Turtle serialization
The first 10 records for MADSRDF & SKOS in all serializations:
Thanks for any feedback,