Metadata structures for films and film images

The primary purpose of the data structures described below is to allow for publication of selected articles, after their intellectual property status had be checked.

Films

• File checked.yaml in a film directory lists sections of the film which can published, because every image had been checked for intellectual property issues
  • contains (possibly multiple) sections, per section:
  • title_en, title_de: short title for the section (optional, free format)
  • start, end: first and last image checked for copyright issues (mandatory)
  • checked_by, checked_date: who has done the checking, and when (mandatory)
  • example
• for each article with a named text and/or image author, create lock and/or meta files for each page

Film images

• File {image_name}.locked.txt indicates that the image contains an article with a named text or image author, who is either not identified, whose death date is unknown, or whose death was less than 70 years ago. Each page (image) with such an article has to be locked. The file is created manually, it cannot be replaced automatically.
• File {image_name}.meta.yaml contains metadata fields about the image
  • date: publication date
  • author: in case of an identified author, QID of the author’s Wikidata item. If the item has a death date, that could be evaluated in order to remove the lock file when the IPR has expired
  • title: title of the article
  • publication: in case of an identified publication QID of the publication’s Wikidata item
  • pos: position on the page L=left, R=right
  • example 1 (image), example 2 (locked) (image)
  • the file may describe more than one article, and one article may stretch across multiple pages
  • for meta files of follow-up pages of an article, symlinks may be used
  • see also structure of annotations for articles

Input files and worksflow

Small sections (working ad-hoc workflow)

• Create or edit a checked.yaml the film directory

    cd {film_directory}
    cp ../../../checked.yaml .

• Fill in checked.yaml (set title for the section, identify first and last image, etc.)
• Check images for intellectual property rights and create a lock file for every dubious image file

    touch {image_name}.locked.txt

• Optionally create {image_name}.meta.yaml files

    cp ../../meta.yaml {image_name}.meta.yaml

• Execute bin/public_film_sections.sh

Massive ammounts (draft)

• Files
  • per film, create a csv or yaml file for checked_yaml, with fields film_id, section_no, title_de, title_en, checked_by, checked_date
  • create a csv file with, for each image of a film_id, image_id (as link), is_locked (defaults true), pos, author, date, publication, title
• Workflow
  • user sets the fields for checked.yaml
  • in csv, user fills in section number(s) and optionally removes all lines not needed for sections (optional)
  • user checks ipr status for every image, removes is_locked
  • optional: user fills in author and other fields (make pos required in that case)
• Verification
  • valid and matching film_ids
  • matching and ascending section numbers
  • pos set for locked and annotated images
• Alternatively, set annotations completely with Hypothes.is