Skip to content

New command to import ALTO files

Refs https://gitlab.com/arkindex/requests/-/issues/646

Goal: Be able to import https://git.litislab.fr/tconstum/popp-datasets/-/tree/main/

We need a new command arkindex upload alto that allows a user to import one or more ALTO XML files.

This tool will assume that images are already available on an IIIF server, and its base url will be provided as CLI arg.

To simplify, we'll only support creating elements in an existing folder, without support for any hierarchy: each alto file produces a page that is stored in the parent provided to the command.

Workflow should be:

  1. build full IIIF url using base url and fileName from header
  2. create the image on arkindex using IIIF endpoint
  3. iterate over Page element, and create page-type elements with mentionned coordinates
  4. iterate over TextBlock elements and create text-block-type elements with mentionned coordinates
  5. iterate over TextLine elements and create line-type elements with mentionned coordinates
  6. iterate over String and create transcription

CLI arguments:

  • --iiif-base-url required
  • --parent-id required
  • --page-type, optional, default to page
  • --line-type, optional, default to text_line
  • --text-block-type, optional, default to paragraph

This is not a generic solution, it most likely will not support a lot of ALTO files, but it's a start. We could later replace the --xxx-type configuration by dynamic configuration allowing to match each types from the XML with a type on Arkindex and thus reproduce the full hierarchy (or skip levels)