Obviously PD4ML has no info about URL base of the document and it cannot resolve the relative path to an absolute one.

A possible reason – you render the HTML to PDF with

[language=java:3s3k01xg]render(java.io.StringReader isr, java.io.OutputStream os)[/language:3s3k01xg]
and the document has no tag.

A solution is to use

[language=java:3s3k01xg]render(java.io.StringReader isr, java.io.OutputStream os, java.net.URL base)[/language:3s3k01xg]


If the explanation does not match your PD4ML usage scenario, please provide more info.