Skip to content

METS import for logical structure

Refs https://redmine.teklia.com/issues/2909

We need a limited support for METS file in order to import Gallica structure for newspaper. We are only interested by the toc.xml files, and their inner <structMap LABEL="Logical Structure" TYPE="logical"> description.

The cli must be extended with a new command arkindex upload mets <path/to/toc.xml> <corpus_id> --element-id=<ELEMENT_ID> which will parse:

  1. all the <file> in any <fileSec>, locate their inner <FLocat xlink:type="simple" LOCTYPE="URL"> and check that the href file exists
  2. for each of these files, try to find a matching JSON file (same name than xml file, but with json) to get the summaries built from #99 (closed)
    • when an XML or JSON file is missing, simply crash with explicit message
  3. parse all <structMap TYPE="logical"> recursively to build a list of trees (hierarchies to build on Arkindex).
    • each <div> with a type will generate a new arkindex element
    • only <div> with a <fptr> child will map to an element previously created
    • if LABEL attribute is set, it must be used as element name
  4. Check that every type used by the tree is available on the corpus
  5. build the hierarchy, browsing the hierarchy in a breadth-first fashion
    • if --element-id was created, top elements (newspaper most likely) must be created under this parent.
    • only CreateElement (without image) and CreateElementParent calls should be made