PDF metadata

[SimonSapin]

Suggested mapping of HTML metadata to PDF metadata:

<title> → Title
<meta name=author> → Author
Maybe? <meta name=description> → Subject
<meta name=keywords> → Keywords
<meta name=generator> → Creator
"WeasyPrint vX.Y" → Producer
<meta name=dcterms.created> → CreationDate
<meta name=dcterms.modified> → ModDate

Relevant specs:
http://www.whatwg.org/specs/web-apps/current-work/multipage/semantics.html#the-title-element
http://www.whatwg.org/specs/web-apps/current-work/multipage/semantics.html#standard-metadata-names
http://wiki.whatwg.org/wiki/MetaExtensions

[SimonSapin]

Is it useful to override these from CSS or Python?

[SimonSapin]

One suggestion was to add the date of PDF generation in the metadata (ie. when WeasyPrint rendered this document) but:

• That does not really match the definition of either CreationDate or ModeDate, the only two date fields in PDF
• I don’t find this date really significant or useful: if you re-render a PDF file from HTML at a later date, it’s still the same document.

My current plan is to expose this metadata (read/write) on Document objects. (In Python, not in CSS.) This will allow people who really want the generation date to add in with Python.

[liZe]

It may be added to cairo: https://lists.cairographics.org/archives/cairo/2016-September/027755.html