Remove all metadata and links matching a URL pattern from a PDF document

remove_metadata_and_links.py

import pikepdf
import re

def remove_metadata(pdf):
    '''remove all metadata'''
    del pdf.Root.Metadata
    print('XMP metadata removed')
    del pdf.docinfo
    print('document info metadata removed')
    return pdf

def remove_link(pdf, url):
    '''remove all links matching url pattern'''
    removed = 0
    for p, page in enumerate(pdf.pages):
        # TODO: use /Parent property instead of manual indexing
        for a, annot in enumerate(page['/Annots']):
            if annot['/A']['/S']=='/URI':
                if re.search(url, str(annot['/A']['/URI'])):
                    del page['/Annots'][a]
                    removed = removed + 1

    print(str(removed) + ' links removed')
    return pdf

def remove_text(pdf, txt):
    '''remove all text matching txt pattern'''
    removed = 0
    for page in pdf.pages:
        commands = []
        for operands, operator in pikepdf.parse_content_stream(page):
            commands.append([operands, operator])
        
        for command in commands:
            operands = command[0]
            if operands and isinstance(operands[0], pikepdf.String):
                if re.search(txt, str(operands[0])):
                    operands.clear()
                    removed = removed + 1
        new_content_stream = pikepdf.unparse_content_stream(commands)
        page.Contents = pdf.make_stream(new_content_stream)
    print(str(removed) + ' text elements removed')
    return pdf

# the actual link which is clickable but invisible
url = r'^http://www.foobar.com/$'
# the text element which is overlayed on the link
txt = r'^www.foobar.com$'

input_pdf = 'input.pdf'
output_pdf = 'output.pdf'

pdf = pikepdf.Pdf.open(input_pdf)
pdf = remove_metadata(pdf)
pdf = remove_link(pdf, url)
pdf = remove_text(pdf, txt)

print('saving to ' + output_pdf)
pdf.save(output_pdf)