Set Alt Text (via document the structure tree)

set_alt_text.py

# set_alt_text.py

# import utils to load required shared libraries
from utils import inputPath, outputPath
from pdfixsdk.Pdfix import *

# sets alt text on nested figures
def SetNestedAltText(struct_elem: PdsStructElement):
  for i in range(struct_elem.GetNumChildren()):
    # search nested child struct elements
    if struct_elem.GetChildType(i) == kPdsStructChildElement:
      child_obj = struct_elem.GetChildObject(i)
      child_elem = struct_elem.GetStructTree().GetStructElementFromObject(child_obj)
    
      if child_elem is None:
        raise Exception(pdfix.GetError())
      
      type = child_elem.GetType(True)
      
      if type == 'Figure':
        child_elem.SetAlt('This is a decorative image')
      
      table = SetNestedAltText(child_elem)
      if table is not None:
        return table
      
  return None
  
# starts with root document element and works down
def SetAltText(struct_tree: PdsStructTree):
  for i in range(struct_tree.GetNumChildren()):
    child_obj = struct_tree.GetChildObject(i)
    child_elem = struct_tree.GetStructElementFromObject(child_obj)

    paragraph = SetNestedAltText(child_elem)
    
    if paragraph is not None:
      return paragraph
    return None
            

pdfix  = GetPdfix()
if pdfix is None:
    raise Exception('Pdfix Initialization fail')

doc = pdfix.OpenDoc(inputPath + "/test.pdf", "")
if doc is None:
    raise Exception('Unable to open pdf : ' + pdfix.GetError())
  
# preflight document for better tagging structure (optional)
tmpl = doc.GetTemplate()
for i in range(0, doc.GetNumPages()):
    tmpl.AddPage(i, 0, None)
tmpl.Update(0, None)

# set the document title and language
title = "Supplemental Document"
lang = "en-US"

# run make accessibile
accessibleParams = PdfAccessibleParams()
if not doc.MakeAccessible(accessibleParams, title, lang, 0, None):
    raise Exception(pdfix.GetError())

struct_tree = doc.GetStructTree()
if not struct_tree:
  raise Exception('No struct tree found')

SetAltText(struct_tree)
    
if not doc.Save(outputPath + "/alt_text.pdf", kSaveFull):
  raise Exception(pdfix.GetError())

doc.Close()