PDF overview
This process for creating a PDF relies on Prince XML to transform the HTML content into PDF. Prince costs about $500 per license. That might seem like a lot, but if you’re creating a PDF, you’re probably working for a company that sells a product, so you likely have access to some resources.
The basic approach is to generate a list of all pages that need to be added to the PDF, and then add leverage Prince to package them up into a PDF.
It may seem like the setup is somewhat cumbersome, but it doesn’t take long. Once you set it up, building a pdf is just a matter of running a couple of commands.
Also, creating a PDF this way gives you a lot more control and customization capabilities than with other methods for creating PDFs. If you know CSS, you can entirely customize the output.
Demo
You can see an example of the finished product here:
1. Set up Prince
Download and install Prince.
You can install a fully functional trial version. The only difference is that the title page will have a small Prince PDF watermark.
2. Create a new configuration file for each of your PDF targets
The PDF configuration file will build on the settings in the regular configuration file but will some additional fields. Here’s the configuration file for the mydoc product within this theme. This configuration file is located in the pdfconfigs folder.
destination: _site/
url: "http://127.0.0.1:4010"
baseurl: "/mydoc-pdf"
port: 4010
output: pdf
product: mydoc
print_title: Jekyll theme for documentation — mydoc product
print_subtitle: version 5.0
output: pdf
defaults:
-
scope:
path: ""
type: "pages"
values:
layout: "page_print"
comments: true
search: true
Note that the default page layout specified by this configuration file is page_print
. This layout strips out all the sections that shouldn’t appear in the print PDF, such as the sidebar and top navigation bar.
Also note that there’s a output: pdf
toggle in case you want to make some of your content unique to PDF output. For example, you could add conditional logic that checks whether site.output
is pdf
or web
. If it’s pdf
, then include information only for the PDF, and so on. If you’re using nav tabs, you’ll definitely want to create an alternative experience in the PDF.
In the configuration file, customize the values for the print_title
and print_subtitle
that you want. These will appear on the title page of the PDF.
3. Make sure your sidebar_doc.yml file has a titlepage.html and tocpage.html
There are two template pages in the root directory that are critical to the PDF:
- titlepage.html
- tocpage.html
These pages should appear in your sidebar YML file (in this product, mydoc_sidebar.yml):
- title:
output: pdf
type: frontmatter
folderitems:
- title:
url: /titlepage/
output: pdf
type: frontmatter
- title:
url: /tocpage/
output: pdf
type: frontmatter
Leave these pages here in your sidebar. (The output: pdf
property means they won’t appear in your online TOC because the conditional logic of the sidebar.html checks whether web
is equal to pdf
or not before including the item in the web version of the content.)
The code in the tocpage.html is mostly identical to that of the sidebar.html page. This is essential for Prince to create the page numbers correctly with cross references.
There’s another file (in the root directory of the theme) that is critical to the PDF generation process: prince-list.txt. This file simply iterates through the items in your sidebar and creates a list of links. Prince will consume the list of links from prince-list.txt and create a running PDF that contains all of the pages listed, with appropriate cross references and styling for them all.
output: web
(rather than output: pdf
) in the list of attributes in your sidebar. The prince-list.txt file that loops through the mydoc_sidebar.yml file to grab the URLs of each page that should appear in the PDF will skip over any items that do not list output: pdf
in the item attributes. For example, you might not want your tag archives to appear in the PDF, but you probably will want to list them in the online help navigation.4. Customize your headers and footers
Open up the css/printstyles.css file and customize what you want for the headers and footers. At the very least, customize the email address (youremail@domain.com
) that appears in the bottom left.
Exactly how the print styling works here is pretty nifty. You don’t need to understand the rest of the content in this section unless you want to customize your PDFs to look different from what I’ve configured. But I’m adding this information here in case you want to understand how to customize the look and feel of the PDF output.
This style creates a page reference for a link:
a[href]::after {
content: " (page " target-counter(attr(href), page) ")"
}
You don’t want cross references for any link that doesn’t reference another page, so this style specifies that the content after should be blank:
a[href*="mailto"]::after, a[data-toggle="tooltip"]::after, a[href].noCrossRef::after {
content: "";
}
noCrossRef
as a class to the link to avoid having it say “page 0” in the cross reference.This style specifies that after links to web resources, the URL should be inserted instead of the page number:
a[href^="http:"]::after, a[href^="https:"]::after {
content: " (" attr(href) ")";
}
This style sets the page margins:
@page {
margin: 60pt 90pt 60pt 90pt;
font-family: sans-serif;
font-style:none;
color: gray;
}
To set a specific style property for a particular page, you have to name the page. This allows Prince to identify the page.
First you add frontmatter to the page that specifies the type. For the titlepage.html, here’s the frontmatter:
---
type: title
---
For the tocpage, here’s the frontmatter:
---
type: frontmatter
---
For the index.html page, we have this type tag (among others):
---
type: first_page
---
The default_print.html layout will change the class of the body
element based on the type value in the page’s frontmatter:
<body class="{% if page.type == "title"%}title{% elsif page.type == "frontmatter" %}frontmatter{% elsif page.type == "first_page" %}first_page{% endif %} print">
Now in the css/printstyles.css file, you can assign a page name based on a specific class:
body.title { page: title }
This means that for content inside of body class="title"
, we can style this page in our stylesheet using @page title
.
Here’s how that title page is styled:
@page title {
@top-left {
content: " ";
}
@top-right {
content: " "
}
@bottom-right {
content: " ";
}
@bottom-left {
content: " ";
}
}
As you can see, we don’t have any header or footer content, because it’s the title page.
For the tocpage.html, which has the type: frontmatter
, this is specified in the stylesheet:
body.frontmatter { page: frontmatter }
body.frontmatter {counter-reset: page 1}
@page frontmatter {
@top-left {
content: prince-script(guideName);
}
@top-right {
content: prince-script(datestamp);
}
@bottom-right {
content: counter(page, lower-roman);
}
@bottom-left {
content: "youremail@domain.com"; }
}
With counter(page, lower-roman)
, we reset the page count to 1 so that the title page doesn’t start the count. Then we also add some header and footer info. The page numbers start counting in lower-roman numerals.
Finally, for the first page (which doesn’t have a specific name), we restart the counting to 1 again and this time use regular numbers.
body.first_page {counter-reset: page 1}
h1 { string-set: doctitle content() }
@page {
@top-left {
content: string(doctitle);
font-size: 11px;
font-style: italic;
}
@top-right {
content: prince-script(datestamp);
font-size: 11px;
}
@bottom-right {
content: "Page " counter(page);
font-size: 11px;
}
@bottom-left {
content: prince-script(guideName);
font-size: 11px;
}
}
You’ll see some other items in there such as prince-script
. This means we’re using JavaScript to run some functions to dynamically generate that content. These JavaScript functions are located in the _includes/head_print.html:
<script>
Prince.addScriptFunc("datestamp", function() {
return "PDF last generated: June 13, 2019";
});
</script>
<script>
Prince.addScriptFunc("guideName", function() {
return " User Guide";
});
</script>
There are a couple of Prince functions that are default functions from Prince. This gets the heading title of the page:
content: string(doctitle);
This gets the current page:
content: "Page " counter(page);
Because the theme uses JavaScript in the CSS, you have to add the --javascript
tag in the Prince command (detailed later on this page).
5. Customize the PDF script
Duplicate the pdf-mydocf.sh file in the root directory and customize it for your specific configuration files.
echo 'Killing all Jekyll instances'
kill -9 $(ps aux | grep '[j]ekyll' | awk '{print $2}')
clear
echo "Building PDF-friendly HTML site for Mydoc ...";
jekyll serve --detach --config _config.yml,pdfconfigs/config_mydoc_pdf.yml;
echo "done";
echo "Building the PDF ...";
prince --javascript --input-list=_site/pdfconfigs/prince-list.txt -o _pdf/mydoc.pdf;
echo "done";
Note that the first part kills all Jekyll instances. This way you won’t try to serve Jekyll at a port that is already occupied.
The jekyll serve
command serves up the HTML-friendly PDF configurations for our two projects. This web version is where Prince will go to get its content.
The prince script issues a command to the Prince utility. JavaScript is enabled (--javascript
), and we tell it exactly where to find the list of files (--input-list
) — just point to the prince-list.txt file. Then we tell it where and what to output (-o
).
Make sure that the path to the prince-list.txt is correct. For the output directory, I like to output the PDF file into my project’s source (into the files folder). Then when I build the web output, the PDF is included and something I can refer to.
6. Add conditions for your new builds in the sidebarconfigs.html file
In the _includes/custom/sidebarconfigs.html file, there’s a section that looks like this:
{% if site.product == "mydoc" %}
{% assign sidebar_pdf = site.data.sidebars.mydoc_sidebar.entries %}
{% endif %}
{% if site.product == "product1" %}
{% assign sidebar_pdf = site.data.sidebars.product1_sidebar.entries %}
{% endif %}
{% if site.product == "product2" %}
{% assign sidebar_pdf = site.data.sidebars.product2_sidebar.entries %}
{% endif %}
Add your own condition here that points to your sidebar.
What this does is allow the prince-list.txt and toc.html files to use a variable for the sidebar (called sidebar_pdf
) when iterating through the sidebar. Otherwise, you would need to create a unique prince-list.txt and toc.html file for each separate PDF output you have.
7. Add a download button for the PDF
You can add a download button for your PDF using some Bootstrap button code:
<a target="_blank" class="noCrossRef" href="/pdf/mydoc.pdf"><button type="button" class="btn btn-default" aria-label="Left Align"><span class="glyphicon glyphicon-download-alt" aria-hidden="true"></span> PDF Download</button></a>
Here’s what that looks like:
<a target=”_blank” class=”noCrossRef” href=pdf/mydoc.pdf”></a>
JavaScript conflicts
If you have JavaScript on any of your pages, Prince will note errors in Terminal like this:
error: TypeError: value is not an object
However, the PDF will still build.
You need to conditionalize out any JavaScript from your PDF web output before building your PDFs. Make sure that the PDF configuration files have the output: pdf
property.
Then surround the JavaScript with conditional tags like this:
{% unless site.output == "pdf" %}
javascript content here ...
{% endunless %}
For more detail about using unless
in conditional logic, see [Conditional logic][mydoc_conditional_logic]. What this code means is “run this code unless this value is the case.”
Overriding Bootstrap Print Styles
The theme relies on Bootstrap’s CSS for styling. However, for print media, Bootstrap applies the following style:
@media print{*,:after,:before{color:#000!important;text-shadow:none!important;background:0 0!important;-webkit-box-shadow:none!important;box-shadow:none!important}
This is minified, but basically the *
(asterisk) means select all, and applied the color #000 (black). As a result, the Bootstrap style strips out all color from the PDF (for Bootstrap elements).
This is problematic for code snippets that have syntax highlighting. I decided to remove this de-coloring from the print output. I commented out the Bootstrap style:
@media print{*,:after,:before{/*color:#000!important;*/text-shadow:none!important;/*background:0 0!important*/;-webkit-box-shadow:none!important;box-shadow:none!important}
If you update Bootrap, make sure you make this edit. (Sorry, admittedly I couldn’t figure out how to simply overwrite the *
selector with a later style.)
I did, however, remove the color from the alerts and lighten the background shading for pre
elements. The printstyles.css has this setting.
pra_ngx_lua_continue_after_eof
pra_ngx_lua_how_one_instance_time
pra_ngx_lua_lua-variable-scope
pra_ngx_lua_resolve_the_domain_name
pra_redis_dynamic_redis_module_method
pra_postgres_not_support_transaction
pra_nginx_pitfalls_and_common_mistakes
candylab_common_sites_log_system
candylab_base_on_openresty_waf
mydoc_hyperlinks.html#automatedlinks