% vim: tw=80
\load [doc, pdfextra]
\catcode`\.=11

% table notes (http://petr.olsak.net/optex/optex-tricks.html#tnote)
\newcount\tnotenum
\def\tnotelist{}
\def\tnote#1{\incr\tnotenum $^{\rm\_romannumeral\tnotenum}$\global\addto\tnotelist{{#1}}}
\def\tnoteprint{\typoscale[920/920]\par \tnotenum=0
   \ea\foreach\tnotelist
     \do{\advance\tnotenum by1 \par $^{\rm\_romannumeral\tnotenum}$##1 }\par
   \global\tnotenum=0 \gdef\tnotelist{}%
}

\let\_cslinkcolor\Blue

% allow hyperlinking of \OpTeX's control sequences (see doc.opm)
\let\_pdfextra_extdocaction\_pdfextra_urlaction
\let\_extdoclinkcolor\Blue
\def\_Xindex#1#2{\sdef{,#1}{}\slet{el:#1}{optexdoclink}}
\def\optexdoclink{\hlink[extdoc:\optexdocurl\#cs:\_tmpa]{\csstring\\\_tmpb}}
\def\optexdocurl{http://petr.olsak.net/ftp/olsak/optex/optex-doc.pdf}
\isfile{optex-doc.eref}\iftrue \input{optex-doc.eref}\fi

% allow hyperlinks to TeX's control sequences (TeX in a Nutshell)
\def\Xeref#1{\sdef{,#1}{}\slet{el:#1}{texdoclink}}
\def\texdoclink{\hlink[extdoc:\texdocurl\#cs:\_tmpa]{\csstring\\\_tmpb}}
\def\texdocurl{http://petr.olsak.net/ftp/olsak/optex/tex-nutshell.pdf}
\isfile{tex-nutshell.eref}\iftrue \input{tex-nutshell.eref}\fi


\insertoutline{PDF extra}
\tit PDF extra -- extra PDF features for \OpTeX/

\hfill Version \_pdfextra_version

\centerline{\it Michal Vlasák, 2021}
\bigskip

\noindent
`PDFextra` is a third party package for \OpTeX/. It aims to provide access to
more advanced PDF features, which are currently not supported in \OpTeX/ --
especially interactive and multimedia features. The development is hosted at
\url{https://github.com/vlasakm/pdfextra}.

In the spirit of \OpTeX/ you may use these macros in any form you want. Either
by installing this package and doing `\load[pdfextra]` in \OpTeX/, or just by
copying some useful parts of this package into your documents / packages.
\OpTeX/ namespacing is used, but it can be easily stripped, if you wish to
incorporate these macros into other macro packages. The code currently depends
on \LuaTeX/, but mostly uses only pdf\TeX/ primitives and a few simple macros
from \OpTeX/. Additionally, the package can  be used in plain LuaTeX and
LuaLaTeX (in a limited form), see section~\ref[user:plain+latex].

User documentation (`pdfextra-doc.tex`) and technical documentation interleaved
with source code (`pdfextra.opm`) are all typeset in this PDF file. Some
examples of usage are in the user documentation, but file
`pdfextra-example.tex` contains more examples.

\insertoutline{Contents}\outlines{0}
\notoc\nonum \sec Contents
\maketoc

\vfil\break

\chap User documentation

\sec Defining files

Many commands provided by this package require you to supply a file <name>. This
is because many commands either work directly (like inserting attachments or
multimedia) or can optionally use files (like inserting JavaScript).
The \"right" way to use <name> is to first define the <name> with:\nl\indent
\^`\filedef``/<type>[<name>]{<path or URL>}`.

Where <name> is the name you will use to refer to this file. It is currently
limited to ASCII only (as all \"<name>s" required by this package).
Interpretation of <path or URL>
depends on the type, which may be:

\begitems
 * `e`, \"embedded file". The file with path <path> will be embedded to the PDF
 file. A file that is embedded this once, can later be used many times in
 different contexts, e.g. you may use it to attach a video as an attachment but
 also have it play on page 1 and even other pages. This is the best way, because
 the resulting PDF file is self contained.
 * `x`, \"external file". <path> can only be a path to the current directory.
 To refer to the file only <path> is used, sort of like a reference. This way
 the file you want to refer to {\em has} to be present in the same directory as
 the PDF file when it is {\em viewed}!
 * `u`, \"URL file". <URL> is the URL of the file you want to refer to.
\enditems

All these create the same type of object, which ideally could be used
interchangibly everywhere a {\em file specification} is required in PDF. This is
sadly not always true. The limitations to only certain types of `\filedef`'s
will be mentioned in due sections. But as a rule of thumb, most of the time you
want to embed the files into PDF. The external/URL references are good for
refering to external files, although other methods are also possible there.

Because most of the times you want <name> to be embedded file, you may omit the
prior definition and instead use the <path> itself. The file will be
autoembedded.

Examples:

\begtt
\filedef/u[doc-internet]{http://petr.olsak.net/ftp/olsak/optex/optex-doc.pdf}
\filedef/x[doc-local]{optex-doc.pdf}
\filedef/e[doc-embedded]{optex-doc.pdf}
\endtt

\label[user:multimedia]
\sec Multimedia

It is possible to insert video, audio and 3D files for playback/display inside
a PDF file (on a page). There are several different PDF mechanisms for
inserting multimedia. For audio/video this package uses the so called
\"Renditions", which currently have the best support in PDF viewers (fully
works in Acrobat and Foxit, partly in Evince and Okular). Rich Media
annotations are used for 3D art (works only in Acrobat Reader), although it is
possible to also insert audio/video using this mechanism it is very restricting
and Renditions are recommended.

Use command \^`\render``[<name>][<optional key-value parameters>]{<appearance>}` for
inserting audio/video with Renditions or
\^`\RM``[<name>][<optional key-value parameters>]{<appearance>}` for inserting
audio/video/3D as Rich Media annnotation.
The result of this command is similiar to what `\inspic`
produces\fnote{But because annotations are involved, transformations using PDF
literals will not work as expected.}. Both commands expect <name> to be
\~`\filedef`'d name of the file to play/display. As usual, fallback for
interpreting <name> as path (and embedding it) is in place. It is recommended
to only use embedded files with both mechanisms (Rich Media requires it,
Renditions with other than embedded files do not work in Acrobat). Optional
key-value parameters may be used to customize default values. They may be left
out enitrely (including brackets). Last parameter, <appearance>, defines so
called \"normal apperance", which is shown before the annotation is activated
(audio/video starts playing or 3D scene is displayed). The dimensions of
resulting multimedia annotation will be taken from <appearance>. Most likely
you want to use a \"poster" picture (inserted with `\inspic`) as appearance --
the dimensions will be taken from it and e.g. aspect ratio will be nicely
preserved.

Customization of Renditions is possible using key-value parameters, but beware
that it mostly doesn't work in Evince and Okular (Acrobat and Foxit are fine in
this regard). The available parameters are in Table~\ref[tab:kv-renditions].
Most customizations of Rich Media concern 3D art. Available parameters are
listed in Table~\ref[tab:kv-rm].

\midinsert
\label[tab:kv-renditions]
\caption/t Key value parameters available for Renditions (\~`\render`)
\cskip
\centerline{\table{lllp{72mm\fL}}{
Key & Possible values & Default & Description \crll
`controls` & `true` or `false` & `false` & Whether to display audio/video player controls. \crl
`volume` & decimal betwen 0 and 100 & 100 & Audio volume. \crl
`repeat` & integer $\geq$ 0 & 1 & Number of repetitions (0 means loop forever). \crl
`background` & \OpTeX/ color & `\White` & Color used for part of the annotation not covered by video player (for wrong aspect ratios). \crl
`opacity` & decimal between 0 and 1 & 1 & Opacity of `background`. \crl
`aactions` & `\renditionautoplay` & (none) & Can be used for autoplay on page open. \crl
`name` & ascii string & `<name>` & Name for use with actions and scripts. \crll
}}
\endinsert

\midinsert
\label[tab:kv-rm]
\caption/t Key value parameters available for Rich Media (\~`\RM`)
\cskip
\centerline{\table{lp{31mm\fL}lp{75mm\fL}}{
Key & Possible values & Default & Description \crll
`activation` & `explicit` or `auto` & `explicit` & Whether to automatically activate the annotation on page open or display normal apperance until user clicks. \crl
`deactivation` & `explicit` or `auto` & `explicit` & Whether to automatically deactivate the annotation on page close or require explicit deactivation by user (from right click menu). \crl
`toolbar` & `true` or `false` & `true` & Whether to show 3D toolbar (with view and other options). \crl
`views` & comma separated list of view <name>s & `<name>` & List of names of 3D views to be used. The default is to try a view of same <name>. Beware that unknown views are silently ignored. \crl
`scripts` & comma separated list of script <name>s & (none) & List of names of JavaScript script file <name>s to be used. \crl
`name` & ascii string & `<name>` & Name for use with actions and scripts. \crll
}}
\endinsert

The weird `name` key is only required if one media file is used
more than once and control using actions or JavaScript scripts is needed.

Examples of video insertion:

\begtt
% embed file under name "video"
\filedef/e[video]{example-movie.mp4}
% insert video into page using Renditions mechanism with controls and autoplay
\render[video][
  name=bigvideo,
  controls=true,
  aactions=\renditionautoplay,
]{\picwidth=\hsize \inspic{example-image.pdf}}

% render the same file again, but with different dimensions, no controls
% and explicit activation
\render[video]{\inspic{example-image.pdf}}
\endtt

When displaying 3D there are more things involved. First, only U3D and PRC can
be included in PDF files. The simplest way to show the 3D scene on page is without
any optional parameters:
\par\nobreak\begtt
\RM[part.prc]{\picwidth=\hsize \inspic{part.png}}
\endtt

The resulting view will be what is defined in the 3D file. But it is possible
to customize it, by creating custom view. Or even more of them -- they will be
available in the user interface for easy switching, first one is considered
default. Parameters not defined in a custom view often take what is in the 3D
file as default value. \^`\DDDview``[<view name>][<key-value parameters>]` is
the command for defining 3D views. The brackets surrounding key-value
parameters have to be included even if no key-value parameters are used. The
available parameters are explained in Table~\ref[tab:kv-3dview].

\midinsert
\label[tab:kv-3dview]
\caption/t Key value parameters available for 3D views (\~`\DDDview`)
\cskip
\centerline{\table{lp{38mm\fL}lp{55mm\fL}}{
Key & Possible values & Default & Description \crll
`projection` & `perspective` or `ortho` & `perspective` & Projection type to use (perspective distorts the view, e.g. parallel lines are not shown as such, but is more natural to human eye, orthogonal projection is what is generally used with technical parts). \crl
`scale` & decimal number & 1 & Scaling to use when orthogonal projection is used. \crl
`FOV` & number between 0 and 180 & 30 & Field of view for use with perspective projection. \crl
`background` & \OpTeX/ color & `\White` & Color used as background in the 3D scene. \crl
`rendermode` & {\typosize[9/10]`Solid`,\hfil\break`SolidWireframe`,\hfil\break`Transparent`,\hfil\break`TransparentWireframe`,\hfil\break`BoundingBox`,\hfil\break`TransparentBoundingBox`,\hfil\break`Wireframe`,\hfil\break`ShadedWireframe`,\hfil\break`Vertices`,\hfil\break`ShadedVertices`,\hfil\break`Illustration`,\hfil\break`SolidOutline`,\hfil\break`ShadedIllustration`} & (taken from 3D file) & Used rendering mode. See PDF standard\tnote{\url{https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf\#G12.2358303}} for more details. \crl
`lighting` & `White`, `Day`, `Night`, `Hard`, `Primary`, `Blue`, `Red`, `Cube`, `CAD`, `Headlamp` & (taken from 3D file) & Used lighting scheme. See PDF standard\tnote{\url{https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf\#G12.2358356}} for more details.\crl
`method` & `media9`, `manual`, `u3d` & `media9` & Method used for defining 3D camera position and orientation. \crll
}}
\tnoteprint
\endinsert

The value of `method` key influences what other key-value
parameters are available. For details about the manual method see the technical
documentation~(\ref[mm-3dviews]). The `u3d` method works only with U3D files
(not with PRC files) and requires you to know the internal \"path" of the view
contained in the file and is hence not always useful (especially when the paths
are weirdly constructed by exporting applications). But if you know the path
you can use \"`method=u3d, u3dpath=<path>`". All Unicode characters are allowed in
<path>.

The most useful method of specifying the camera position/orientation is
`method=media9`. The same method is also used in a number of other packages
(movie15, rmannot, Con\TeX/t). Its input are key-value parameters, listed in
Table~\ref[tab:kv-media9]. For illustration of the parameters check the
documentation\fnote{\url{https://mirrors.ctan.org/macros/latex/contrib/media9/doc/media9.pdf\#figure.8}}
of media9.

\midinsert
\label[tab:kv-media9]
\caption/t Key value parameters available for media9 method of 3D views (\^`\DDDview`)
\cskip
\centerline{\table{lp{38mm\fL}lp{50mm\fL}}{
Key & Possible values & Default & Description \crll
`coo` & three space separated (decimal) numbers & `0 0 0` & \"Center of orbit". Coordinates of the point camera is supposed to look at. \crl
`roo` & (decimal) number & `0` & Distance of camera from `coo`. \crl
`c2c` & three space separated (decimal) numbers & `0 -1 0` & \"Center of orbit to camera" vector. A directional vector (i.e. length doesn't matter). The direction the camera will be looking at is opposite of this vector. Default is view towards positive $y$.\crll
}}
\endinsert

You can construct simple views only by using these few parameters. When talking
about meanings/names of different views it is needed to know the real placement
of the 3D object in the 3D world. But for a reasonably orientated object in the
center (`coo=0 0 0`) constructing simple views is very easy (keep in mind
`method=media9` is default):

\begtt
\DDDview[front][
  projection=ortho,
  roo=400,
]
\DDDview[left][
  projection=ortho,
  roo=400,
  c2c=-1 0 0,
]
\DDDview[top][
  projection=ortho,
  roo=400,
  c2c=0 0 1,
]
\DDDview[isometric][
  projection=ortho,
  roo=500,
  c2c=-1 -1 1,
]
\endtt

We can then perhaps use these views in \^`\RM`:

\begtt
% Auto activated 3D Rich Media annotation. White background just ensures
% the right dimensions. Comma in `views` is escaped using "{}".
\RM[part.prc][
  activation=auto,
  views={front, left},
]{{\White\vrule width\hsize height\vsize}}

\RM[part.prc][ % the same 3D file, now with different views
  name=part2,
  activation=auto,
  views={top, isometric},
]{{\White\vrule width\hsize height\vsize}}
\endtt

If you want to deduce the view parameters automatically it is possible. You can
just not specify any view, but include the `3Dmenu.js` script from media9
package. It enables you to right click the annotation and select \"Get Current
View" (or even \"Generate Default View" which finds the view all by itself). A
window with generated parameters will show up. You can then copy the ones this
package understands (`c2c`, `coo`, `roo`) and use them. You don't have to be
excessive with precision, because after calculation everything gets rounded to
6 decimal places anyways.

\begtt
% using 3Dmenu.js to generate 3D view parameters automatically
\RM[part.prc][
  name=part3,
  activation=auto,
  scripts=3Dmenu.js,
]{{\White\vrule width\hsize height\vsize}}
\endtt

The use of `scripts` isn't limited to this though, there are many other
possibilities. First, you can use as many scripts as you want
(`scripts={script1.js, script2.js, ...}`), but be careful that 3D JavaScript is
kind of special, and has to come from embedded files (here we
were using the auto-embedding, but we could have used e.g.
\~`\filedef/e[3dmenu]{3Dmenu.js}` and then \"`3dmenu`" instead). For creating
your own scripts check out the Acrobat 3D JavaScript
API\fnote{\url{https://wwwimages.adobe.com/content/dam/acom/en/devnet/acrobat/pdfs/AcrobatDC_js_3d_api_reference.pdf}}.
It is possible to do different transformations and achieve animations using
\"time events". The implicit \"context" of 3D scripts can be accessed from
normal JavaScript actions (see~\ref[user:actions-3djs]). Vice versa 3D scripts
may access the global JavaScript environment using `host` object.

More examples of 3D Rich Media, including usage of 3D JavaScript, are available
in the example file `pdfextra-example.tex`. They show how it is possible to
port the examples used by media9.

\sec Actions

Actions are very important aspect of interactivity in the context of PDF. There
are a few very useful types of actions, like \"goto" actions which jump to
other part of document. There are also a few ways how to {\em execute} actions.
The most usual is a clickable area on page, but clickable bookmarks (\"document
outline") also execute actions behind the scenes. \OpTeX/ supports only basic
\"goto" and \"URI" actions using `\ilink` (used for `\ref`, `\cite`, etc.) and
`\ulink` (used for `\url`).

This package offers generalization of this mechanism. The core of it is a way
of specifying an action. This syntax is called <action spec> and is used for
example by `\hlink` command, which can replace both `\ilink` and `\ulink`.
<action spec> is a comma separated list of `<type>:<arguments>`. where <type>
refers to the action type and the syntax of arguments is dependant on <type>.
Leading spaces are ignored, trailing aren't. You probably won't often use the
possiblity of specifying multiple actions, but it is for chaining execution of
several actions.

\^`\pdfaction``[<action spec>]` is available for lower level creation of
different actions, but for clickable areas on page you will use
\^`\hlink``[<action spec>]{<text>}` with a very similiar syntax. Although note,
that `\hlink`'s interface is also not really high level and wrapping it inside
macros like `\ref` or `\url` might be beneficial. <text> will be typeset
directly and the area it occupies will be clickable. Clicking it executes
action defined by <action spec>. Line breaks inside <text> will be possible, in
that case several clickable rectangles will be created, one for each line.
Normally in text you want the areas to be of the same height and depth
(calculated from `\baselineskip`), to achieve sort of a lining, uniform effect.
If you want to define big clickable buttons, you may need to turn off the
lining effect using \^`\nolininglinks`. It respects groups, but a counterpart
(\^`\lininglinks`) is also available.

There are a few predefined action types (<type>): `url`, `extref`, `extpgref`,
`named`, `transition`, `js`, `goto3dview` and `rendition`. They will be explained in a
moment. Any unrecognized <type> is understood as an \"internal link", where
`<type>:<link>` is the destination of the link. Hence it is possible to
use `\hlink` as `\ilink` for example with \OpTeX/'s normal types
of internal links. For example:
\begtt
See section~\hlink[ref:section]{2.2.13} or page~\hlink[pg:5]{5}.
\endtt
`\ulink` may be replaced like this:
\begtt
Visit CTAN's \hlink[url:https://www.ctan.org/]{website}.
\endtt

Before we really get into different types of actions, there is a nicer command
for setting the initial (\"open") action of PDF document, which is executed
when the document is opened. It defaults to opening the page with the zoom
level according to viewing user's preferences. But you can change this using
\^`\openaction``[<action spec>]`:

\begtt
\openaction[pg:2]
\endtt

\secc External references
There are two actions analogous to internal links that can be used to link to
external documents. `[extref:<name>:<named destination>]` can be used to refer
to named locations in a PDF document prepared by `\filedef` with <name>.
`[extpgref:<name>:<page number>]` is similiar but refers to a page number.
Although it would be nice, these actions aren't well supported by all viewers.

Example:

\begtt
\hlink[extref:doc-internet:ref:langphrases]{OpTeX documentation,
        section \"Multilingual phrases and quotation marks".}
\hlink[extpgref:doc-internet:12]{OpTeX documentation, page 12.}
\endtt

(Note that \"`:`" in \"`ref:langphrase`" is not part of the syntactic rule, it
is just the value of <named destination> in this case.)

A little bit of customization is possible, see~\ref[actions-jump].

Because of the poor support you may find luck with the less universal url action with `#fragment`:

\begtt
\hlink[url:http://petr.olsak.net/ftp/olsak/optex/optex-doc.pdf#ref:langphrases]
      {OpTeX documentation, section \"Multilingual phrases and quotation marks".}
\endtt

\secc Named actions
There are four defined in PDF standard, but viewers may support more. All in
examples:

\begtt
\hlink[named:NextPage]{Go to next page,}
\hlink[named:PrevPage]{go to previous page,}
\hlink[named:FirstPage]{go to first page,}
\hlink[named:LastPage]{go to last page.}
\endtt

\secc Transition actions

Transition actions generally make sense only when chained {\em after} jump
actions. The specified transition/animation will occur, before destination is
opened, but it will not override a transition defined for the particular page.
The syntax is `[transition:<transition spec>]`. See~\ref[user:transitions] for
more information about `<transition spec>`.

Example:
\begtt
\hlink[ref:yellow-slide, transition:Box:3:/M /O]
      {Go to yellow slide with long outward Box transition}
\endtt

\secc JavaScript actions

They allow executing pieces of JavaScript code using syntax:
`[js:<name or script>]`. If <name or script> is a validly `\filedef`'d the script from file
<name> will be executed (only embedded files are valid). Otherwise <script> will
be used directly. Apart from reuse in different documents, scripts loaded from
files may be encoded in UTF-16BE and hence support Unicode, inline <script>s
current don't.

Examples:

\begtt
\openaction[js:{%
  app.alert("Javascript alert, open action");
  console.println("printing to console from openaction");
}]

\filedef/e[jstest]{test.js}
\hlink[js:jstest]{JavaScript action from file}
\endtt

(Note how braces were used to guard the comma in JavaScript code from being
interpreted as a action separator.)

In these actions you may want to use your own functions which should be defined
before user has a chance of activating any other JavaScript actions. This is the
purpose of \"document level" JavaScript actions. They function exactly the same
way as normal JavaScript actions, but have names (although meaningless, they
must be unique) and are executed in order of definition. Use
\^`\dljavascript``[<name>]{<name or script>}` for defining these actions:

\begtt \adef![#1]{\url{#1}}
\filedef/e[preamble]{preamble.js}
\dljavascript[preamble]{preamble}

\dljavascript[initialization]{%
  var data = 42;
  function getRandomNumber() {
      return 4; // chosen by fair dice roll, ![https://xkcd.com/221/]
  }
  console.println("initialized with seed " + getRandomNumber());
}
\endtt

The JavaScript API available is not the same as the one in the web browsers. It
is instead specified by Adobe:
\url{https://wwwimages.adobe.com/content/dam/acom/en/devnet/acrobat/pdfs/js_api_reference.pdf}.

\label[user:actions-3djs]
\secc 3D JavaScript actions

There are no \"special" JavaScript actions for dealing with 3D. Normal
JavaScript actions are used. You just have to access the 3D context of the
annotation you want to control. For annotation with <name> the context is made
available in \^`\DDDcontext{<name>}`. Under this context object you find all
global definitions from the respective 3D scripts. For example if a \"`turn`"
function is defined, it is possible to call it like this:

\begtt
\hlink[js:\DDDcontext{part3}.turn();]{Turn by 90 degrees along $x$ axis.}
\endtt

\secc GoTo3Dview actions

GoTo3Dview actions allow changing the view of the 3D scene to one of the
predefined views (those listed by `views` comma separated list of \~`\RM`). The
syntax is `[goto3dview:<name>:<view>]`. <name> is name of the annotation whose
view we want to change and <view> is the intended view. You can refer to last
inserted Rich Media annotation using empty name. For <view> you can either use
\"`(<view name>)`" (e.g. \"`(part)`"), index of the view in the view list (zero
based, e.g.\ \"`0`" for the first view) or one of the special values: \"`/N`"
(next), \"`/P`" (previous), \"`/F`" (first) \"`/L`" (\"last").

\begtt
% let's define an annotation with a few views
\RM[part.prc][
  activation=auto,
  views={front, left, top, isometric},
]{{\White\vrule width\hsize height\vsize}}

% try the different methods of reffering to views
\hlink[goto3dview::/N]{Next view},
\hlink[goto3dview::(left)]{left view} and
\hlink[goto3dview:part.prc:3]{third view}.
\endtt

\secc Rendition actions

Rendition actions can be used to control playback of \"Rendition annotations".
They use the syntax `[rendition:<name>:<operation>]`, where <name> refers to
the name of the rendition to control and <operation> is one of `play`, `stop`,
`pause` or `resume`. As a convenience, you can refer to last inserted rendition
using empty name. If a file has been \~`\render`ed more than once with the same
name, the action will influence the first instance.

Beware that currently these actions do not work in Evince and Okular (but do in
Acrobat and Foxit).

Examples:

\begtt
% rendition with name=video, that we want to control
\render[video]{\inspic{example-image.pdf}}

% we want the rendition action to have yellow border and red content
\def\_renditionborder{1 1 0}
\let\_renditionlinkcolor\Red

To start playing the video, click \hlink[rendition::play]{\"Play"}.
After that you can \hlink[rendition:video]{pause}.
\endtt

\label[user:transitions]
\sec Transitions and other page attributes

In PDF there are a few settings that can be set as {\em page attributes}. This
means that they apply only to said page. For setting these page attributes,
there are two options:

\begitems
* value for \"current page" (or rather the page where the command appears),
* default value used if \"current page" value is not set.
\enditems

While this package contains mechanism to handle all page attributes, not that
many are useful for end user. The interesting ones remaining are:

\begitems
* Transitions and page durations. When page has the transition attribute any
jump to this page will display the requested animation (customized by the
corresponding parameters). Page duration is the time before PDF viewer will auto
advance to the next page. \^`\transition``[<transition spec>]` sets transition
for the \"current page", \^`\transitions` sets the default.
`<transition spec>` has three parts:\nl\indent
`<animation type>:<duration>:<raw PDF attributes>`\nl
\noindent <animation type> is one of `Split`, `Blinds`, `Box`, `Wipe`,
`Dissolve`, `Glitter`, `Fly`, `Push`, `Cover`, `Uncover` and `Fade`, or the
special value `R` which essentially means no animation and instantaneous
transition (regardless of the set duration of transition). <duration> is the
duration of transition in seconds (integer or decimal number).
<raw PDF attributes> may be used to customize
the animations (for example `/M` can set the direction of motion of `Split`,
`Box` and `Fly`, e.g. `/M /I` for inward motion). For the raw PDF attributes
refer to the standard
itself\fnote{\url{https://wwwimages2.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf\#G11.2295795}}.
`:<raw PDF attributes>` or even `:<duration>:<raw PDF attributes>` may be omitted.
Default values specified by PDF standard will be used in that case (1 second
duration and default values for all attributes). Page durations can be set
either using \^`\defaultpageduration``[<duration>]` or
\^`\pageduration``[<duration>]`, where duration is in seconds (default is no auto
advancement, i.e. $\infty$).
Examples:

\begtt
% unless stated otherwise all pages will have Wipe animation
% with 1 second duration
\transitions[Wipe:1]

% but this page has a 1 second Fade animation
\transition[Fade]

\pg+ % (if using \slides)

% and this page has 3 second Split animation
% with vertical direction and inward motion
\transition[Split:3:/Dm /V /M /I]
\endtt

Note that transitions are only displayed when in {\em full-screen mode}. You
can use \~`\fullscreen` to have the document automatically open in full-screen
mode.

* Additional actions. It is possible to define actions which will respond to
page events: page open (`/O`) and page close (`/C`). They can be set using
\^`\defaultpageactions``[<additional actions spec>]` (the default for all pages) and
\^`\pageactions``[<additional actions spec>]` (current page override).
<additional actions spec> consists of braced pairs of \"event" (O or C in this
case) and <action spec>. Example:

\begtt
\pageactions[
  {O} {js:{app.alert("Page open, random = " + getRandomNumber());}}
  {C} {js:{app.alert("Page close!");}}
]
\endtt

\enditems

\sec Attachments

Every file embedded into a PDF file may optionally be presented in the user
interface as an embedded file. This allows readers of the document to save or
open the file.

PDF allows two ways of presenting attachments -- using annotations (the
attachments is represented by a rectangular icon on a page) or global array of
attachments (attachments are visible in PDF viewers toolbar). The first method
is intended more for document reviewers than for primary insertion. Hence this
package supports only the second type.

You may add an embedded file to the global attachments array using
\^`\attach``[<name>]`. As usual, name is either a <name> defined with
`\filedef` or alternatively a path to file which will be embedded (and
`\filedef`d) automatically. It is not possible to `\attach` files referenced by
path or URL.

If you want to automatically display toolbar with embedded files, consider
using `\showattached` (see section~\ref[user:document-view]).

\label[user:document-view]
\sec Document view

You can choose what is shown when document is opened with commands:

\begitems
* \^`\fullscreen` (the document is opened in full-screen mode),
* \^`\showoutlines` (show bookmarks/outlines toolbar)
* \^`\showattached` (show attachments in toolbar)
\enditems

The commands are mutually exclusive and only the first appearing one will be respected.

You can request two page view (odd pages on the right) using
\^`\duplexdisplay`. It is useful for more natural display of double sided
documents. Because it may not be desirable to automatically apply this, it is
independent of `\margins`.

\label[user:plain+latex]
\sec Usage in plain \LuaTeX/ or LuaLa\TeX/

You can use this package also from plain LuaTeX by adding to your document:

\begtt
\input pdfextra
\endtt

\noindent or for LuaLa\TeX/:

\begtt
\usepackage{pdfextra}
\endtt

See the file `pdfextra-example-latex.tex` for the adaptation of the \OpTeX/
examples to \LaTeX/.

The usage of the macros described in this document is the same, but there are
limitations:

\begitems
* {\em Color}. Where this package expects \"\OpTeX/ color" key value argument
(e.g. `\Blue`), you have to use an RGB triplet instead (e.g. `0.0 0.0 0.0` or
the shorter `0 0 0`).

But for text color setting, you can get away with wrapping commands from
La\TeX/'s `color` package, e.g. to customize link border/color:

  \begtt
  \sdef{_renditionlinkcolor}{\color[red]}
  \endtt

* {\em Initialization}. In \OpTeX/ you normally have to initialize hyperlinks
with the command\nl
`\hyperlinks<color for internal links><color for external links>`.\nl
This is not required by this package. You can instead set the color by setting
`\_linkcolor` (fallback for all link types), `\_ilinkcolor` / `\_elinkcolor`
(internal / external links). E.g. if you have the \LaTeX/ `color` package
loaded, you can get blue links like this:

  \begtt
  \sdef{_linkcolor}{\color[blue]}
  \endtt

In addition to the above this means all links will be blue except \"rendition"
links.

* {\em Openaction}. If the package `hyperref` is used, then `\openaction` will
not work.

* {\em Labels / hyperlink destinations}. \OpTeX/ uses very simple and consistent
scheme for labels / hyperlink destinations:

  \begitems
  * `ref:<label>`     -- result of `\label[<label>]`
  * `toc:<tocrefnum>` -- result of `\chap`/`\sec`/`\secc` titles
  * `pg:<gpageno>`    -- created on each page with global numbering from 1
  * `cite:<bibpart>/<bibnum>` -- bibliography references,
  * `fnt:<gfnotenum>` -- link form text to footnote
  * `fnf:<gfnotenum>` -- link from footnote to text
  * `url:<url>`       -- used by `\url` or `\ulink`,
  \enditems

  Hence these labels / destinations can be used with `\hlink`, e.g. to make text
  `page 5` a link to page 5, one can use:

  \begtt
  \hlink[pg:5]{page 5}
  \endtt

  This is not possible in plain \LuaTeX/ (no destinations are created) or
  LuaLa\TeX/ (different and incompatible destination names). You would have to
  create your own destinations adhering to the naming convention
  `<type>:<arguments>` to be able to use `\hlink` as intended for some links.

\enditems

\chap Technical documentation

This is the technical documentation. It is intended for those who want to know
how this package works internally. Casual users shouldn't need to read this. But
if you would like to customize anything or perhaps just use some part of this
package, feel free to copy paste and use anything you want in \OpTeX/'s spirit.

This documentation is interleaved within the source itself, both are contained
in a single file, `pdfextra.opm` (according to \OpTeX/ conventions). The user
documentation is instead contained in `pdfextra-doc.tex`, which itself
`\input`'s the documented source file `pdfextra.opm` so that the user and
technical documentation is available in a single PDF file, `pdfextra-doc.pdf`.

\printdoc     pdfextra.opm

\bye