/usr/share/gtk-doc/html/libvips/extending.html

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Extending VIPS: VIPS Reference Manual</title>
<meta name="generator" content="DocBook XSL Stylesheets V1.79.1">
<link rel="home" href="index.html" title="VIPS Reference Manual">
<link rel="up" href="ch01.html" title="VIPS Overview">
<link rel="prev" href="binding.html" title="Writing bindings for libvips">
<link rel="next" href="func-list.html" title="VIPS function list">
<meta name="generator" content="GTK-Doc V1.25 (XML mode)">
<link rel="stylesheet" href="style.css" type="text/css">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
<table class="navigation" id="top" width="100%" summary="Navigation header" cellpadding="2" cellspacing="5"><tr valign="middle">
<td width="100%" align="left" class="shortcuts"></td>
<td><a accesskey="h" href="index.html"><img src="home.png" width="16" height="16" border="0" alt="Home"></a></td>
<td><a accesskey="u" href="ch01.html"><img src="up.png" width="16" height="16" border="0" alt="Up"></a></td>
<td><a accesskey="p" href="binding.html"><img src="left.png" width="16" height="16" border="0" alt="Prev"></a></td>
<td><a accesskey="n" href="func-list.html"><img src="right.png" width="16" height="16" border="0" alt="Next"></a></td>
</tr></table>
<div class="refentry">
<a name="extending"></a><div class="titlepage"></div>
<div class="refnamediv"><table width="100%"><tr>
<td valign="top">
<h2><span class="refentrytitle">Extending VIPS</span></h2>
<p>Extending — How to add operations to VIPS</p>
</td>
<td class="gallery_image" valign="top" align="right"></td>
</tr></table></div>
<div class="refsect3">
<a name="extending-intro"></a><h4>Introduction</h4>
<p>
			This section runs quickly through adding a simple operator to VIPS. 
			For more information, see <a class="link" href="VipsOperation.html" title="VipsOperation"><span class="type">VipsOperation</span></a> and <a class="link" href="VipsRegion.html" title="VipsRegion"><span class="type">VipsRegion</span></a>. A good 
			starting point for a new operation is a similar one in the VIPS library. 
    </p>
<p>
      All VIPS operations are subclasses of <a class="link" href="VipsOperation.html" title="VipsOperation"><span class="type">VipsOperation</span></a>, which in turn
			subclasses <a class="link" href="VipsObject.html" title="VipsObject"><span class="type">VipsObject</span></a> and then <code class="literal">GObject</code>. You add an operation to VIPS
			by defining a new subclass of <a class="link" href="VipsOperation.html" title="VipsOperation"><span class="type">VipsOperation</span></a> and arranging for its 
			<code class="code"><code class="function">class_init()</code></code> to be called, perhaps by calling its <code class="function">get_type()</code>
			function.
    </p>
</div>
<div class="refsect3">
<a name="extending-classstruct"></a><h4>The class and object structures</h4>
<p>
      First you need to define a new
      object struct and a new class struct.

</p>
<pre class="programlisting">
typedef struct _Negative {
  VipsOperation parent_instance;

  VipsImage *in;
  VipsImage *out;

  int image_max;

} Negative;

typedef struct _NegativeClass {
  VipsOperationClass parent_class;

  /* No new class members needed for this op.
   */

} NegativeClass;
</pre>
<p>
    </p>
<p>
      This operation will find the photographic negative of an unsigned
      8-bit image, optionally letting you specify the value which the pixels
      "pivot" about. It doesn't need any class members (ie. values common
      to all operations of this type), so the second struct is empty. See
			the source to <a class="link" href="libvips-arithmetic.html#vips-invert" title="vips_invert ()"><code class="function">vips_invert()</code></a> for a more complete version of this 
			operation that's actually in the library.
    </p>
<p>
      <code class="literal">GObject</code> has a handy macro to write some of the boilerplate for you.

</p>
<pre class="programlisting">
G_DEFINE_TYPE( Negative, negative, VIPS_TYPE_OPERATION );
</pre>
<p>

      <code class="function">G_DEFINE_TYPE()</code> defines a function called <code class="function">negative_get_type()</code>,
      which registers this new class and returns its <code class="literal">GType</code> (a
      pointer-sized integer). <code class="function">negative_get_type()</code> in turn needs two
      functions, <code class="function">negative_init()</code>, to initialise a new instance, and
      <code class="function">negative_class_init()</code>, to initialise a new class.
    </p>
</div>
<div class="refsect3">
<a name="extending-init"></a><h4>Class and object initialisation</h4>
<p>
      <code class="function">negative_init()</code> is very simple, it just sets the default value for
      our optional parameter.

</p>
<pre class="programlisting">
static void
negative_init( Negative *negative )
{
  negative-&gt;image_max = 255;
}
</pre>
<p>
    </p>
<p>
      <code class="function">negative_class_init()</code> is more complicated: it has to set various
      fields in various superclasses and define the operation's parameters. 

</p>
<pre class="programlisting">
static void
negative_class_init( NegativeClass *class )
{
  GObjectClass *gobject_class = G_OBJECT_CLASS( class );
  VipsObjectClass *object_class = VIPS_OBJECT_CLASS( class );

  gobject_class-&gt;set_property = vips_object_set_property;
  gobject_class-&gt;get_property = vips_object_get_property;

  object_class-&gt;nickname = "negative";
  object_class-&gt;description = "photographic negative";
  object_class-&gt;build = negative_build;

  VIPS_ARG_IMAGE( class, "in", 1, 
    "Input", 
    "Input image",
    VIPS_ARGUMENT_REQUIRED_INPUT,
    G_STRUCT_OFFSET( Negative, in ) );

  VIPS_ARG_IMAGE( class, "out", 2, 
    "Output", 
    "Output image",
    VIPS_ARGUMENT_REQUIRED_OUTPUT, 
    G_STRUCT_OFFSET( Negative, out ) );

  VIPS_ARG_INT( class, "image_max", 4, 
    "Image maximum", 
    "Maximum value in image: pivot about this",
    VIPS_ARGUMENT_OPTIONAL_INPUT,
    G_STRUCT_OFFSET( Negative, image_max ),
    0, 255, 255 );
}
</pre>
<p>
    </p>
<p>
      In <code class="literal">GObject</code>, it needs to set the getters and setters for this class. vips
      has a generic get/set system, so any subclass of <a class="link" href="VipsObject.html" title="VipsObject"><span class="type">VipsObject</span></a> needs to
      use the vips ones.
    </p>
<p>
      In <a class="link" href="VipsObject.html" title="VipsObject"><span class="type">VipsObject</span></a>, it needs to set the operation <em class="parameter"><code>nickname</code></em> and <em class="parameter"><code>description</code></em>,
      and set a build function (see below). <em class="parameter"><code>nickname</code></em> is used to refer to
      this operation in the API, <em class="parameter"><code>description</code></em> is used to explain this
      operation to users and will be translated into their language.
    </p>
<p>
      Finally, it needs to define the arguments the constructor for this class
			takes. There are a set of handy macros for doing this, see 
			<a class="link" href="VipsObject.html#VIPS-ARG-INT:CAPS" title="VIPS_ARG_INT()"><code class="function">VIPS_ARG_INT()</code></a> and friends.
    </p>
<p>
			The first few
      parameters are always the same and mean: class pointer for argument,
      argument name, argument priority (bindings expect required arguments in
      order of priority), long argument name (this one is internationalised
      and displayed to users), description (again, users can see this),
      some flags describing the argument, and finally the position of the
      member in the struct.
    </p>
<p>
      Integer arguments take three more values: the minimum, maximum and
      default value for the argument.
    </p>
</div>
<div class="refsect3">
<a name="extending-build"></a><h4>The <code class="function">build()</code> function</h4>
<p>
			The build function is the thing <a class="link" href="VipsObject.html" title="VipsObject"><span class="type">VipsObject</span></a> calls during object 
			construction, after all arguments have been supplied and before the 
			object is used. It has two roles: to verify that arguments are correct, 
			and then to construct the object.  After <code class="function">build()</code>, the object is expected
      to be ready for use.

</p>
<pre class="programlisting">
static int
negative_build( VipsObject *object )
{
  VipsObjectClass *class = VIPS_OBJECT_GET_CLASS( object );
  Negative *negative = (Negative *) object;

  if( VIPS_OBJECT_CLASS( negative_parent_class )-&gt;build( object ) )
    return( -1 );

  if( vips_check_uncoded( class-&gt;nickname, negative-&gt;in ) ||
    vips_check_format( class-&gt;nickname, negative-&gt;in, VIPS_FORMAT_UCHAR ) )
    return( -1 );

  g_object_set( object, "out", vips_image_new(), NULL ); 

  if( vips_image_pipelinev( negative-&gt;out, 
    VIPS_DEMAND_STYLE_THINSTRIP, negative-&gt;in, NULL ) )
    return( -1 );

  if( vips_image_generate( negative-&gt;out, 
    vips_start_one, 
    negative_generate, 
    vips_stop_one, 
    negative-&gt;in, negative ) )
    return( -1 );

  return( 0 );
}
</pre>
<p>
    </p>
<p>
      <code class="function">negative_build()</code> first chains up to the superclass: this will check
      that all input arguments have been supplied and are sane.
    </p>
<p>
      Next, it adds its own checks. This is a demo operation, so we just
			work for uncoded, unsigned 8-bit images. There are a lot of 
			convenience functions like <a class="link" href="libvips-error.html#vips-check-format" title="vips_check_format ()"><code class="function">vips_check_format()</code></a>, see the docs. 
    </p>
<p>
      Next, it creates the output image. This needs to be set with
      <code class="function">g_object_set()</code> so that vips can see that it has been assigned. vips
      will also handle the reference counting for you.
    </p>
<p>
      <a class="link" href="libvips-generate.html#vips-image-pipelinev" title="vips_image_pipelinev ()"><code class="function">vips_image_pipelinev()</code></a> links our new image onto the input image and
      notes that this operation prefers to work in lines. You can request 
      other input geometries, see <a class="link" href="VipsImage.html#VipsDemandStyle" title="enum VipsDemandStyle"><span class="type">VipsDemandStyle</span></a>. 
    </p>
<p>
			The geometry hint is just a hint, an operation needs to be able to 
			supply any size 
      <a class="link" href="VipsRegion.html" title="VipsRegion"><span class="type">VipsRegion</span></a> on request. If you must have a certain size request, you can 
      put a cache in the pipeline after your operation, see <a class="link" href="libvips-conversion.html#vips-linecache" title="vips_linecache ()"><code class="function">vips_linecache()</code></a> 
			and <a class="link" href="libvips-conversion.html#vips-tilecache" title="vips_tilecache ()"><code class="function">vips_tilecache()</code></a>. You can also make requests to your operation 
			ordered, see <a class="link" href="libvips-conversion.html#vips-sequential" title="vips_sequential ()"><code class="function">vips_sequential()</code></a>. 
    </p>
<p>
      Finally, <a class="link" href="libvips-generate.html#vips-image-generate" title="vips_image_generate ()"><code class="function">vips_image_generate()</code></a> attaches a set of callbacks to the
      output image to generate chunks of it on request. <a class="link" href="libvips-generate.html#vips-start-one" title="vips_start_one ()"><code class="function">vips_start_one()</code></a>
      and <a class="link" href="libvips-generate.html#vips-stop-one" title="vips_stop_one ()"><code class="function">vips_stop_one()</code></a> are convenience functions that make the input
      region for you, see below. 
    </p>
</div>
<div class="refsect3">
<a name="extending-gen"></a><h4>The <code class="function">generate()</code> function</h4>
<p>
			The <code class="function">generate()</code> function does the actual image processing. 
			<code class="function">negative_generate()</code> (of type <a class="link" href="VipsImage.html#VipsGenerateFn" title="VipsGenerateFn ()"><span class="type">VipsGenerateFn</span></a>, supplied to 
			<a class="link" href="libvips-generate.html#vips-image-generate" title="vips_image_generate ()"><code class="function">vips_image_generate()</code></a> above) is 
			called whenever some pixels of our output image are required. 

</p>
<pre class="programlisting">
static int
negative_generate( VipsRegion *or, 
  void *vseq, void *a, void *b, gboolean *stop )
{
  /* The area of the output region we have been asked to make.
   */
  VipsRect *r = &amp;or-&gt;valid;

  /* The sequence value ... the thing returned by vips_start_one().
   */
  VipsRegion *ir = (VipsRegion *) vseq;

  VipsImage *in = (VipsImage *) a;
  Negative *negative = (Negative *) b;
  int line_size = r-&gt;width * negative-&gt;in-&gt;Bands; 

  int x, y;

  /* Request matching part of input region.
   */
  if( vips_region_prepare( ir, r ) )
    return( -1 );

  for( y = 0; y &lt; r-&gt;height; y++ ) {
    unsigned char *p = (unsigned char *)
      VIPS_REGION_ADDR( ir, r-&gt;left, r-&gt;top + y ); 
    unsigned char *q = (unsigned char *)
      VIPS_REGION_ADDR( or, r-&gt;left, r-&gt;top + y ); 

    for( x = 0; x &lt; line_size; x++ ) 
      q[x] = negative-&gt;image_max - p[x];
  }

  return( 0 );
}
</pre>
<p>
    </p>
<p>
      This has to calculate a section of the output image. The output
      <a class="link" href="VipsRegion.html" title="VipsRegion"><span class="type">VipsRegion</span></a>, <em class="parameter"><code>or</code></em>, contains a <a class="link" href="libvips-rect.html#VipsRect" title="VipsRect"><span class="type">VipsRect</span></a> called <em class="parameter"><code>valid</code></em> which is the
			area needing calculation. This call to <code class="function">negative_generate()</code> must 
			somehow make this part of <em class="parameter"><code>or</code></em> contain pixel data. 
    </p>
<p>
			<em class="parameter"><code>vseq</code></em> is the sequence value. This is the 
			per-thread state for this generate, created (in this example) by 
			<a class="link" href="libvips-generate.html#vips-start-one" title="vips_start_one ()"><code class="function">vips_start_one()</code></a>. In this simple case it's just a <a class="link" href="VipsRegion.html" title="VipsRegion"><span class="type">VipsRegion</span></a> defined on 
			the input image. If you need more per-thread state you can write your 
			own start and stop functions and have a struct you create and pass as a 
			sequence value. There are plenty of examples in the VIPS source code, 
			see <a class="link" href="libvips-morphology.html#vips-rank" title="vips_rank ()"><code class="function">vips_rank()</code></a>. 
    </p>
<p>
			<em class="parameter"><code>a</code></em> and <em class="parameter"><code>b</code></em> are the last two arguments to <a class="link" href="libvips-generate.html#vips-image-generate" title="vips_image_generate ()"><code class="function">vips_image_generate()</code></a> above. 
			<em class="parameter"><code>stop</code></em> is a bool pointer you can set to stop computation early. <a class="link" href="libvips-arithmetic.html#vips-min" title="vips_min ()"><code class="function">vips_min()</code></a>
			on an unsigned int image, for example, will set <em class="parameter"><code>stop</code></em> as soon as it sees 
			a zero, and will not scan the entire image.
    </p>
<p>
			The first thing <code class="function">negative_generate()</code> does is 
			use <a class="link" href="VipsRegion.html#vips-region-prepare" title="vips_region_prepare ()"><code class="function">vips_region_prepare()</code></a> to 
			ask for the corresponding pixels from the input image. Operations which 
			do coordinate transforms or which need an area of input for each output 
			point will need to calculate a new rect before calling 
			<a class="link" href="VipsRegion.html#vips-region-prepare" title="vips_region_prepare ()"><code class="function">vips_region_prepare()</code></a>. 
    </p>
<p>
			Finally, it can calculate some pixels. <code class="function">negative_generate()</code> loops
			over the valid area of the output and calls <a class="link" href="VipsRegion.html#VIPS-REGION-ADDR:CAPS" title="VIPS_REGION_ADDR()"><code class="function">VIPS_REGION_ADDR()</code></a> for each 
			line. This macro is reasonaly quick, but it's best not to call it for 
			each pixel. Once per line is fine though. 
    </p>
</div>
<div class="refsect3">
<a name="extending-add"></a><h4>Adding to VIPS</h4>
<p>
			To add the operation to vips, just call <code class="function">negative_get_type()</code>. You can 
			include the source in your program, or use <code class="literal">GModule</code> to make a binary 
			plugin that will be loaded by libvips at startup. There are some <a class="ulink" href="https://github.com/jcupitt/vips-gmic" target="_top">example 
			plugins available</a>.
    </p>
<p>
			You
      can then use <em class="parameter"><code>negative</code></em> from any of the vips interfaces. For example,
      in Python you'd use it like this:

</p>
<pre class="programlisting">
out = in.negative(image_max = 128)
</pre>
<p>
    </p>
<p>
      From the command-line it'd look like this:

</p>
<pre class="programlisting">
$ vips negative in.png out.tif --image-max 128
</pre>
<p>
    </p>
<p>
      And from C like this:

</p>
<pre class="programlisting">
VipsImage *in;
VipsImage *out;
if( vips_call( "negative", in, &amp;out, "image_max", 128, NULL ) )
  ... error
</pre>
<p>
    </p>
<p>
      Unfortunately that will do almost no compile-time type checking,
      so all vips operations have a tiny extra wrapper to add a bit of
      safety. For example:

</p>
<pre class="programlisting">
static int 
negative( VipsImage *in, VipsImage **out, ... )
{
  va_list ap;
  int result;

  va_start( ap, out );
  result = vips_call_split( "negative", ap, in, out );
  va_end( ap );

  return( result );
}
</pre>
<p>
    </p>
<p>
      And now you can write:

</p>
<pre class="programlisting">
if( negative( in, &amp;out, "image_max", 128, NULL ) )
  ... error
</pre>
<p>

      and it's at least a bit safer.
    </p>
</div>
<div class="refsect3">
<a name="extending-othertypes"></a><h4>Other types of operation</h4>
<p>
      Change the <code class="function">_build()</code> function to make other types of operation. 
    </p>
<p>
      Use <a class="link" href="libvips-generate.html#vips-image-generate" title="vips_image_generate ()"><code class="function">vips_image_generate()</code></a> with <a class="link" href="libvips-generate.html#vips-start-many" title="vips_start_many ()"><code class="function">vips_start_many()</code></a> to make operations 
      which demand pixels from more than one image at once, such as image 
      plus image. 
    </p>
<p>
      Use <a class="link" href="libvips-generate.html#vips-sink" title="vips_sink ()"><code class="function">vips_sink()</code></a> instead of <a class="link" href="libvips-generate.html#vips-image-generate" title="vips_image_generate ()"><code class="function">vips_image_generate()</code></a> to loop over an image 
      and calculate a value. vips uses this for the statistics operations, 
      like <a class="link" href="libvips-arithmetic.html#vips-avg" title="vips_avg ()"><code class="function">vips_avg()</code></a>.
    </p>
<p>
      Use <a class="link" href="VipsImage.html#vips-image-wio-input" title="vips_image_wio_input ()"><code class="function">vips_image_wio_input()</code></a> to get an entire image into memory so you
      can read it with a pointer. This will obviously not scale well to
      very large images, but some operations, like FFTs or flood-fill, need 
      the whole image to be available at once.
    </p>
<p>
      Make area operations, like filters, by enlarging the <a class="link" href="libvips-rect.html#VipsRect" title="VipsRect"><span class="type">VipsRect</span></a> that
      <code class="function">_generate()</code> is given before calling <a class="link" href="VipsRegion.html#vips-region-prepare" title="vips_region_prepare ()"><code class="function">vips_region_prepare()</code></a>. You can
      enlarge the input image, so that the output image is the same size as
      the original input, by using <a class="link" href="libvips-conversion.html#vips-embed" title="vips_embed ()"><code class="function">vips_embed()</code></a> within the <code class="function">_build()</code> function.
    </p>
<p>
      Make things like flips and rotates by making larger changes to the
      <a class="link" href="libvips-rect.html#VipsRect" title="VipsRect"><span class="type">VipsRect</span></a> in <code class="function">_generate()</code>.
    </p>
<p>
      Make zero-copy operations, like <a class="link" href="libvips-conversion.html#vips-insert" title="vips_insert ()"><code class="function">vips_insert()</code></a>, with <a class="link" href="VipsRegion.html#vips-region-region" title="vips_region_region ()"><code class="function">vips_region_region()</code></a>.
    </p>
</div>
</div>
<div class="footer">
<hr>Generated by GTK-Doc V1.25</div>
</body>
</html>
libvips-doc 8.4.5-1build1 / usr / share / gtk-doc / html / libvips / extending.html
