Metadata
Introduction
A PDF document can contain metadata in various locations. The main locations are the information
dictionary in the trailer (accessible via document.trailer.info
and the XMP
metadata stream in the document catalog (accessible via
document.catalog[:Metadata]
).
Among the supported metadata are the title and author(s) of the document, the application creating the original document from which the PDF was converted, the PDF processor creating the PDF itself, and the creation and modification dates.
Up to and including PDF 1.7 the information dictionary was as important as the main metadata stream. However, with PDF 2.0 all metadata keys except /CreationDate and /ModDate in the information dictionary have been deprecated in favor of the metadata stream.
The metadata stream is the more standard way of defining metadata since it can be read by any XMP processor. This is even possible if the XMP processor doesn’t understand the PDF file format since the metadata stream is always written without any PDF filters and may also be exempted from being encrypted.
HexaPDF fully supports the information dictionary as source and destination of metadata. Howver, it only supports the XMP metadata stream as destination, not as source.
Usage
There are two ways to access the metadata: manually or through a dedicated interface.
Manually accessing the information dictionary can be done via document.trailer.info
for the
information dictionary and document.catalog[:Metadata]
for the metadata stream. As mentioned
above, the metadata stream has to be parsed manually to extract the needed information.
require 'hexapdf'
doc = HexaPDF::Document.open(ARGV[0])
puts doc.trailer.info[:Title]
The dedicated interface is available via document.metadata
. Once
this method is invoked the metadata is read from the information dictionary and stored internally
where it can be read and changed. When writing the PDF document, the stored information is written
back to the information dictionary and the main metadata stream. Note that this will overwrite any
metadata manually changed in the information dictionary and a possibly existing metadata stream.
Whether the information dictionary and/or the XMP metadata stream are written can be controlled by using the appropriate methods.
The convenience interface supports all metadata from the information dictionary. Additionally, it supports adding custom XMP metadata, albeit not to the full extent possible with XMP. For this, the used XMP namespaces as well as property types need to be registered beforehand. That additional XMP metadata is only written to the metadata stream.
require 'hexapdf'
doc = HexaPDF::Document.open(ARGV[0])
puts doc.metadata.title
doc.metadata.author(['Author1', 'Author2'])
doc.metadata.register_property_type('dc', 'contributor', 'UnorderedArray')
doc.metadata.property('dc', 'contributor', ['Name1', 'Name2'])
doc.encrypt(owner_password: 'Some password here', encrypt_metadata: false)
doc.write('encrypted_with_XMP_metadata.pdf', optimize: true)