symbiote / silverstripe-pdfrendition
A module that makes use of the Flying Saucer XHTML renderer project to create PDFs from XHTML pages.
Installs: 4 883
Dependents: 0
Suggesters: 0
Security: 0
Stars: 17
Watchers: 3
Forks: 10
Open Issues: 6
Type:silverstripe-vendormodule
Requires
- silverstripe/framework: ~4.1
Replaces
README
Maintainer Contact
- Marcus Nyeholt marcus@symbiote.com.au
- Nathan Glasl nathan@symbiote.com.au
Requirements
- SilverStripe 4.12+ || 5+
- Tidy (preferably the built in PHP tidy module, otherwise the commandline binary)
- Java 1.8 (the latest version this has been tested against)
- Important: earlier 1.7 versions cause the PDF to not load Cloudflare specific assets such as images and CSS
- Important: regardless of the version you end up using, make sure the PDF output is correct on the production server prior to go-live (please see the known issues section below)
Documentation
This module allows users to easily create complex PDF renditions of content by utilising HTML and CSS3 to define page layouts for printing. It provides a simple extension that adds a simple action for automatically generating PDF renditions of a page, and an API for developers to generate more specific PDF renditions.
Installation Instructions
composer require symbiote/silverstripe-pdfrendition
Usage Overview
- Add
Symbiote\PdfRendition\Extension\PdfControllerExtension
as an extension to PageController - Add $PdfLink in your template to insert a link to the PDF version of the page
- To customise the PDF layout, create a 'pdfrendition.css' file in your theme directory, link to it with your preferred method (@import, requirements, etc.) and add styles specifically for your pdf using the @print media query. See the github wiki for some examples of how to do some common PDF based things.
Known Issues / Troubleshooting
- Using HTTPS without a valid certificate can cause the PDF to not render correctly.
- Cloudflare can cause the PDF to not render correctly.
- This may be due to Java attempting to reference the assets (images and CSS), and being listed as a "bad browser".
- To resolve this, Cloudflare
Page Rules
need to be added for those specific assets (or a general/*
blanket rule) withBrowser Integrity Check
set toOff
.
- To resolve this, Cloudflare
- This may also be due to the Java version (please see above).
- The PDF output can be tested by intercepting the render process locally and holding onto the
xhtml
file generated prior to render. - Using this, replace the asset URLs with those you want to test and confirm (production for example).
- The Java Flying Saucer utility will retrieve assets and external sources via links contained in the
xhtml
source. Ensure these links are fully formed, and are able to be retrieved from within the production server (i.e. outbound firewall restrictions or localised DNS/host definitions could cause issues).
- The Java Flying Saucer utility will retrieve assets and external sources via links contained in the
- Using this
/tmp/xhtml
, the below should give you a correctly rendered PDF when run from within the production server. - If not, Cloudflare and/or Java are likely the issue.
- This may be due to Java attempting to reference the assets (images and CSS), and being listed as a "bad browser".
java -classpath '{project}/pdfrendition/thirdparty/xhtmlrenderer/flying-saucer-core-9.0.7.jar:{project}/pdfrendition/thirdparty/xhtmlrenderer/flying-saucer-pdf-9.0.7.jar:{project}/pdfrendition/thirdparty/xhtmlrenderer/itext-4.2.1.jar' org.xhtmlrenderer.simple.PDFRenderer '/tmp/xhtml' '/tmp/output.pdf'
- Make sure you don't define @font-face inside @media print.
Occasionally a page won't correctly render, throwing some kind of junk back to the browser as the PDF rendition process fails. Typically, this is caused by malformed XML being sent to the renderer; for this reason everything is first passed through HTML Tidy, however in some rare cases this can still not correctly convert the raw content.
In these cases, errors will be sent through to your error log files; it will indicate the temporary files that were created, so you should first check these for XML errors. If that does not work, you can also attempt to manually perform the conversion using commandline tidy and the commandline for the PDF rendition to see if there are more verbose errors available for debugging the problem.