increase coverage

[bertsky]

• Confidence (unfortunately, this conflates Coords and Text @conf)
• TextType (HANDWRITING → @production=handwritten-printscript|handwritten-cursive, PRINTED → @production=printed)
• support tables:
  • top-level TableRegion for TABLE block
  • recursive TextRegion for CELL block (i.e. ColumnIndex → Roles/TableCellRole/@columnIndex, RowIndex → Roles/TableCellRole/@rowIndex)
  • recursive TextRegion for MERGED_CELL block (i.e. ColumnSpan → Roles/TableCellRole/@colSpan, RowSpan→Roles/TableCellRole/@rowSpan) – diverging recursion between Textract and PAGE?
  • recursive TextRegion for TABLE_TITLE and TABLE_FOOTER block (i.e. Roles/TableCellRole/@header... or via ReadingOrder)
  • EntityTypes – STRUCTURED_TABLE|SEMI_STRUCTURED_TABLE (unclear how to represent in PAGE), TABLE_TITLE|TABLE_SECTION_TITLE|TABLE_FOOTER|TABLE_SUMMARY|COLUMN_HEADER (unclear how this looks and compares with the actual recursive BlockType)?
  • also via ordered groups in ReadingOrder?
  • unclear: LineItemGroup and LineItems
• PageClassification/PageType (unclear, but probably Page/@type)
• support forms
  • BlockType=KEY_VALUE_SET and EntityTypes=KEY|VALUE → unclear how to represent: TableRegion or recursive TextRegion? Labels/Label?
  • register KEY_VALUE_SET
  • represent in page
• support checkboxes within tables or forms
  • BlockType=SELECTION_ELEMENT and SelectionStatus=SELECTED|NOT_SELECTED → unclear how to represent
  • register SELECTION_ELEMENT
  • represent in page
• ignore query type