GD.pm - Interface to Gd Graphics Library
use GD;
# create a new image
$im = new GD::Image(100,100);
# allocate some colors
$white = $im->colorAllocate(255,255,255);
$black = $im->colorAllocate(0,0,0);
$red = $im->colorAllocate(255,0,0);
$blue = $im->colorAllocate(0,0,255);
# make the background transparent and interlaced
$im->transparent($white);
$im->interlaced('true');
# Put a black frame around the picture
$im->rectangle(0,0,99,99,$black);
# Draw a blue oval
$im->arc(50,50,95,75,0,360,$blue);
# And fill it with red
$im->fill(50,50,$red);
# make sure we are writing to a binary stream
binmode STDOUT;
# Convert the image to PNG and print it on standard output
print $im->png;
GD.pm is a Perl interface to Thomas Boutell's gd graphics library (version 2.01 or higher; see below). GD allows you to create color drawings using a large number of graphics primitives, and emit the drawings as PNG files.
GD defines the following four classes:
GD::Image
GD::Font
GD::Polygon
GD::Simple
A Simple Example:
#!/usr/local/bin/perl
use GD;
# create a new image
$im = new GD::Image(100,100);
# allocate some colors
$white = $im->colorAllocate(255,255,255);
$black = $im->colorAllocate(0,0,0);
$red = $im->colorAllocate(255,0,0);
$blue = $im->colorAllocate(0,0,255);
# make the background transparent and interlaced
$im->transparent($white);
$im->interlaced('true');
# Put a black frame around the picture
$im->rectangle(0,0,99,99,$black);
# Draw a blue oval
$im->arc(50,50,95,75,0,360,$blue);
# And fill it with red
$im->fill(50,50,$red);
# make sure we are writing to a binary stream
binmode STDOUT;
# Convert the image to PNG and print it on standard output
print $im->png;
Notes:
binmode() on the file you are writing
the image to.
The following class methods allow you to create new GD::Image objects.
The new() method is the main constructor for the GD::Image class. Called with two integer arguments, it creates a new blank image of the specified width and height. For example:
$myImage = new GD::Image(100,100) || die;
This will create an image that is 100 x 100 pixels wide. If you don't specify the dimensions, a default of 64 x 64 will be chosen.
The optional third argument, $truecolor, tells new() to create a truecolor GD::Image object. Truecolor images have 24 bits of color data (eight bits each in the red, green and blue channels respectively), allowing for precise photograph-quality color usage. If not specified, the image will use an 8-bit palette for compatibility with older versions of libgd.
Alternatively, you may create a GD::Image object based on an existing image by providing an open filehandle, a filename, or the image data itself. The image formats automatically recognized and accepted are: PNG, JPEG, XPM and GD2. Other formats, including WBMP, and GD version 1, cannot be recognized automatically at this time.
If something goes wrong (e.g. insufficient memory), this call will return undef.
For backwards compatibility with scripts previous versions of GD, new images created from scratch (width, height) are palette based by default. To change this default to create true color images use:
GD::Image->trueColor(1);
somewhere before creating new images. To switch back to palette based by default, use:
GD::Image->trueColor(0);
The newFromPng() method will create an image from a PNG file read in
through the provided filehandle or file path. The filehandle must
previously have been opened on a valid PNG file or pipe. If
successful, this call will return an initialized image which you can
then manipulate as you please. If it fails, which usually happens if
the thing at the other end of the filehandle is not a valid PNG file,
the call returns undef. Notice that the call doesn't automatically
close the filehandle for you. But it does call binmode(FILEHANDLE)
for you, on platforms where this matters.
You may use any of the following as the argument:
1) a simple filehandle, such as STDIN 2) a filehandle glob, such as *PNG 3) a reference to a glob, such as \*PNG 4) an IO::Handle object 5) the pathname of a file
In the latter case, newFromPng() will attempt to open the file for you and read the PNG information from it.
Example1:
open (PNG,"barnswallow.png") || die;
$myImage = newFromPng GD::Image(\*PNG) || die;
close PNG;
Example2:
$myImage = newFromPng GD::Image('barnswallow.png');
To get information about the size and color usage of the information, you can call the image query methods described below. Images created by reading PNG images will be truecolor if the image file itself is truecolor. To force the image to be palette-based, pass a value of 0 in the optional $truecolor argument.
The newFromPngData() method will create a new GD::Image initialized
with the PNG format data contained in $data.
These methods will create an image from a JPEG file. They work just like newFromPng() and newFromPngData(), and will accept the same filehandle and pathname arguments.
Images created by reading JPEG images will always be truecolor. To force the image to be palette-based, pass a value of 0 in the optional $truecolor argument.
These methods will create an image from a GIF file. They work just like newFromPng() and newFromPngData(), and will accept the same filehandle and pathname arguments.
Images created from GIFs are always 8-bit palette images. To convert to truecolor, you must create a truecolor image and then perform a copy.
This works in exactly the same way as newFromPng, but reads the
contents of an X Bitmap (black & white) file:
open (XBM,"coredump.xbm") || die; $myImage = newFromXbm GD::Image(\*XBM) || die; close XBM;
There is no newFromXbmData() function, because there is no corresponding function in the gd library.
These methods initialize a GD::Image from a Gd file, filehandle, or data. Gd is Tom Boutell's disk-based storage format, intended for the rare case when you need to read and write the image to disk quickly. It's not intended for regular use, because, unlike PNG or JPEG, no image compression is performed and these files can become BIG.
$myImage = newFromGd GD::Image("godzilla.gd") || die;
close GDF;
newFromGd() and
newFromGdData, but use the new compressed GD2 image format.
This class method allows you to read in just a portion of a GD2 image file. In addition to a filehandle, it accepts the top-left corner and dimensions (width,height) of the region of the image to read. For example:
open (GDF,"godzilla.gd2") || die; $myImage = GD::Image->newFromGd2Part(\*GDF,10,20,100,100) || die; close GDF;
This reads a 100x100 square portion of the image starting from position (10,20).
This creates a new GD::Image object starting from a filename. This is unlike the other newFrom() functions because it does not take a filehandle. This difference comes from an inconsistency in the underlying gd library.
$myImage = newFromXpm GD::Image('earth.xpm') || die;
This function is only available if libgd was compiled with XPM support.
NOTE: The libgd library is unable to read certain XPM files, returning an all-black image instead.
Once a GD::Image object is created, you can draw with it, copy it, and merge two images. When you are finished manipulating the object, you can convert it into a standard image file format to output or save to a file.
The following methods convert the internal drawing format into standard output file formats.
This returns the image data in PNG format. You can then print it, pipe it to a display program, or write it to a file. Example:
$png_data = $myImage->png; open (DISPLAY,"| display -") || die; binmode DISPLAY; print DISPLAY $png_data; close DISPLAY;
Note the use of binmode(). This is crucial for portability to
DOSish platforms.
The optional $compression_level argument controls the amount of compression to apply to the output PNG image. Values range from 0-9, where 0 means no compression (largest files, highest quality) and 9 means maximum compression (smallest files, worst quality). A compression level of -1 uses the default compression level selected when zlib was compiled on your system, and is the same as calling png() with no argument. Be careful not to confuse this argument with the jpeg() quality argument, which ranges from 0-100 and has the opposite meaning from compression (higher numbers give higher quality).
For libgd version 2.0.33 and higher, this call begins an animated GIF by returning the data that comprises animated gif image file header. After you call this method, call gifanimadd() one or more times to add the frames of the image. Then call gifanimend(). Each frame must be the same width and height.
A typical sequence will look like this:
my $gifdata = $image->gifanimbegin;
$gifdata .= $image->gifanimadd; # first frame
for (1..100) {
# make a frame of right size
my $frame = GD::Image->new($image->getBounds);
add_frame_data($frame); # add the data for this frame
$gifdata .= $frame->gifanimadd; # add frame
}
$gifdata .= $image->gifanimend; # finish the animated GIF
print $gifdata; # write animated gif to STDOUT
If you do not wish to store the data in memory, you can print it to stdout or a file.
The image that you call gifanimbegin on is used to set the image size, color resolution and color map. If argument $GlobalCM is 1, the image color map becomes the GIF89a global color map. If $Loops is given and >= 0, the NETSCAPE2.0 application extension is created, with looping count. Looping count 0 means forever.
This returns the image data in GD format. You can then print it, pipe it to a display program, or write it to a file. Example:
binmode MYOUTFILE; print MYOUTFILE $myImage->gd;
These methods allow you to control and manipulate the GD::Image color table.
This allocates a color with the specified red, green and blue components and returns its index in the color table, if specified. The first color allocated in this way becomes the image's background color. (255,255,255) is white (all pixels on). (0,0,0) is black (all pixels off). (255,0,0) is fully saturated red. (127,127,127) is 50% gray. You can find plenty of examples in /usr/X11/lib/X11/rgb.txt.
If no colors are allocated, then this function returns -1.
Example:
$white = $myImage->colorAllocate(0,0,0); #background color $black = $myImage->colorAllocate(255,255,255); $peachpuff = $myImage->colorAllocate(255,218,185);
alphaBlending function changes the way this
alpha channel affects the resulting image.
This marks the color at the specified index as being ripe for reallocation. The next time colorAllocate is used, this entry will be replaced. You can call this method several times to deallocate multiple colors. There's no function result from this call.
Example:
$myImage->colorDeallocate($peachpuff); $peachy = $myImage->colorAllocate(255,210,185);
This returns the index of the color closest in the color table to the red green and blue components specified. If no colors have yet been allocated, then this call returns -1.
Example:
$apricot = $myImage->colorClosest(255,200,180);
This also attempts to return the color closest in the color table to the red green and blue components specified. It uses a Hue/White/Black color representation to make the selected color more likely to match human perceptions of similar colors.
If no colors have yet been allocated, then this call returns -1.
Example:
$mostred = $myImage->colorClosestHWB(255,0,0);
This returns the index of a color that exactly matches the specified red green and blue components. If such a color is not in the color table, this call returns -1.
$rosey = $myImage->colorExact(255,100,80); warn "Everything's coming up roses.\n" if $rosey >= 0;
This returns the index of a color that exactly matches the specified red green and blue components. If such a color is not in the color table and there is room, then this method allocates the color in the color table and returns its index.
$rosey = $myImage->colorResolve(255,100,80); warn "Everything's coming up roses.\n" if $rosey >= 0;
This returns the total number of colors allocated in the object.
$maxColors = $myImage->colorsTotal;
In the case of a TrueColor image, this call will return undef.
This returns the color table index underneath the specified point. It can be combined with rgb() to obtain the rgb color underneath the pixel.
Example:
$index = $myImage->getPixel(20,100);
($r,$g,$b) = $myImage->rgb($index);
This returns a list containing the red, green and blue components of the specified color index.
Example:
@RGB = $myImage->rgb($peachy);
This marks the color at the specified index as being transparent. Portions of the image drawn in this color will be invisible. This is useful for creating paintbrushes of odd shapes, as well as for making PNG backgrounds transparent for displaying on the Web. Only one color can be transparent at any time. To disable transparency, specify -1 for the index.
If you call this method without any parameters, it will return the current index of the transparent color, or -1 if none.
Example:
open(PNG,"test.png"); $im = newFromPng GD::Image(PNG); $white = $im->colorClosest(255,255,255); # find white $im->transparent($white); binmode STDOUT; print $im->png;
GD implements a number of special colors that can be used to achieve special effects. They are constants defined in the GD:: namespace, but automatically exported into your namespace when the GD module is loaded.
You can draw lines and shapes using a brush pattern. Brushes are just images that you can create and manipulate in the usual way. When you draw with them, their contents are used for the color and shape of the lines.
To make a brushed line, you must create or load the brush first, then assign it to the image using setBrush(). You can then draw in that with that brush using the gdBrushed special color. It's often useful to set the background of the brush to transparent so that the non-colored parts don't overwrite other parts of your image.
Example:
# Create a brush at an angle $diagonal_brush = new GD::Image(5,5); $white = $diagonal_brush->colorAllocate(255,255,255); $black = $diagonal_brush->colorAllocate(0,0,0); $diagonal_brush->transparent($white); $diagonal_brush->line(0,4,4,0,$black); # NE diagonal # Set the brush $myImage->setBrush($diagonal_brush); # Draw a circle using the brush $myImage->arc(50,50,25,25,0,360,gdBrushed);
Styled lines consist of an arbitrary series of repeated colors and are useful for generating dotted and dashed lines. To create a styled line, use setStyle() to specify a repeating series of colors. It accepts an array consisting of one or more color indexes. Then draw using the gdStyled special color. Another special color, gdTransparent can be used to introduce holes in the line, as the example shows.
Example:
# Set a style consisting of 4 pixels of yellow, # 4 pixels of blue, and a 2 pixel gap $myImage->setStyle($yellow,$yellow,$yellow,$yellow, $blue,$blue,$blue,$blue, gdTransparent,gdTransparent); $myImage->arc(50,50,25,25,0,360,gdStyled);
To combine the gdStyled and gdBrushed behaviors, you can specify
gdStyledBrushed. In this case, a pixel from the current brush
pattern is rendered wherever the color specified in setStyle() is
neither gdTransparent nor 0.
setTile in order to define the particular tile pattern you'll use
for drawing when you specify the gdTiled color.
details.
gdAntiAliased color is used for drawing lines with antialiasing
turned on. Antialiasing will blend the jagged edges of lines with the
background, creating a smoother look. The actual color drawn is set
with setAntiAliased().
"Antialiasing" is a process by which jagged edges associated with line drawing can be reduced by blending the foreground color with an appropriate percentage of the background, depending on how much of the pixel in question is actually within the boundaries of the line being drawn. All line-drawing methods, such as line() and polygon, will draw antialiased lines if the special "color" gdAntiAliased is used when calling them.
setAntiAliased() is used to specify the actual foreground color to be used when drawing antialiased lines. You may set any color to be the foreground, however as of libgd version 2.0.12 an alpha channel component is not supported.
Antialiased lines can be drawn on both truecolor and palette-based images. However, attempts to draw antialiased lines on highly complex palette-based backgrounds may not give satisfactory results, due to the limited number of colors available in the palette. Antialiased line-drawing on simple backgrounds should work well with palette-based images; otherwise create or fetch a truecolor image instead. When using palette-based images, be sure to allocate a broad spectrum of colors in order to have sufficient colors for the antialiasing to use.
Normally, when drawing lines with the special gdAntiAliased "color," blending with the background to reduce jagged edges is the desired behavior. However, when it is desired that lines not be blended with one particular color when it is encountered in the background, the setAntiAliasedDontBlend() method can be used to indicate the special color that the foreground should stand out more clearly against.
Once turned on, you can turn this feature off by calling setAntiAliasedDontBlend() with a second argument of 0:
$image->setAntiAliasedDontBlend($color,0);
These methods allow you to draw lines, rectangles, and ellipses, as well as to perform various special operations like flood-fill.
This sets the pixel at (x,y) to the specified color index. No value is returned from this method. The coordinate system starts at the upper left at (0,0) and gets larger as you go down and to the right. You can use a real color, or one of the special colors gdBrushed, gdStyled and gdStyledBrushed can be specified.
Example:
# This assumes $peach already allocated $myImage->setPixel(50,50,$peach);
This draws a line from (x1,y1) to (x2,y2) of the specified color. You can use a real color, or one of the special colors gdBrushed, gdStyled and gdStyledBrushed.
Example:
# Draw a diagonal line using the currently defined # paintbrush pattern. $myImage->line(0,0,150,150,gdBrushed);
DEPRECATED: The libgd library provides this method solely for backward compatibility with libgd version 1.0, and there have been reports that it no longer works as expected. Please use the setStyle() and gdStyled methods as described below.
This draws a dashed line from (x1,y1) to (x2,y2) in the specified color. A more powerful way to generate arbitrary dashed and dotted lines is to use the setStyle() method described below and to draw with the special color gdStyled.
Example:
$myImage->dashedLine(0,0,150,150,$blue);
This draws a rectangle with the specified color. (x1,y1) and (x2,y2) are the upper left and lower right corners respectively. Both real color indexes and the special colors gdBrushed, gdStyled and gdStyledBrushed are accepted.
Example:
$myImage->rectangle(10,10,100,100,$rose);
This draws a rectangle filed with the specified color. You can use a real color, or the special fill color gdTiled to fill the polygon with a pattern.
Example:
# read in a fill pattern and set it
$tile = newFromPng GD::Image('happyface.png');
$myImage->setTile($tile);
# draw the rectangle, filling it with the pattern
$myImage->filledRectangle(10,10,150,200,gdTiled);
This draws a polygon with the specified color. The polygon must be created first (see below). The polygon must have at least three vertices. If the last vertex doesn't close the polygon, the method will close it for you. Both real color indexes and the special colors gdBrushed, gdStyled and gdStyledBrushed can be specified.
Example:
$poly = new GD::Polygon; $poly->addPt(50,0); $poly->addPt(99,99); $poly->addPt(0,99); $myImage->openPolygon($poly,$blue);
This draws a sequence of connected lines with the specified color, without connecting the first and last point to a closed polygon. The polygon must be created first (see below). The polygon must have at least three vertices. Both real color indexes and the special colors gdBrushed, gdStyled and gdStyledBrushed can be specified.
You need libgd 2.0.33 or higher to use this feature.
Example:
$poly = new GD::Polygon; $poly->addPt(50,0); $poly->addPt(99,99); $poly->addPt(0,99); $myImage->unclosedPolygon($poly,$blue);
This draws a polygon filled with the specified color. You can use a real color, or the special fill color gdTiled to fill the polygon with a pattern.
Example:
# make a polygon $poly = new GD::Polygon; $poly->addPt(50,0); $poly->addPt(99,99); $poly->addPt(0,99); # draw the polygon, filling it with a color $myImage->filledPolygon($poly,$peachpuff);
This draws arcs and ellipses. (cx,cy) are the center of the arc, and (width,height) specify the width and height, respectively. The portion of the ellipse covered by the arc are controlled by start and end, both of which are given in degrees from 0 to 360. Zero is at the top of the ellipse, and angles increase clockwise. To specify a complete ellipse, use 0 and 360 as the starting and ending angles. To draw a circle, use the same value for width and height.
You can specify a normal color or one of the special colors gdBrushed, gdStyled, or gdStyledBrushed.
Example:
# draw a semicircle centered at 100,100 $myImage->arc(100,100,50,50,0,180,$blue);
This method is like arc() except that it colors in the pie wedge with the selected color. $arc_style is optional. If present it is a bitwise OR of the following constants:
gdArc connect start & end points of arc with a rounded edge gdChord connect start & end points of arc with a straight line gdPie synonym for gdChord gdNoFill outline the arc or chord gdEdged connect beginning and ending of the arc to the center
gdArc and gdChord are mutually exclusive. gdChord just connects the starting and ending angles with a straight line, while gdArc produces a rounded edge. gdPie is a synonym for gdArc. gdNoFill indicates that the arc or chord should be outlined, not filled. gdEdged, used together with gdNoFill, indicates that the beginning and ending angles should be connected to the center; this is a good way to outline (rather than fill) a "pie slice."
Example:
$image->filledArc(100,100,50,50,0,90,$blue,gdEdged|gdNoFill);
This method flood-fills regions with the specified color. The color will spread through the image, starting at point (x,y), until it is stopped by a pixel of a different color from the starting pixel (this is similar to the "paintbucket" in many popular drawing toys). You can specify a normal color, or the special color gdTiled, to flood-fill with patterns.
Example:
# Draw a rectangle, and then make its interior blue $myImage->rectangle(10,10,100,100,$black); $myImage->fill(50,50,$blue);
Like fill, this method flood-fills regions with the specified
color, starting at position (x,y). However, instead of stopping when
it hits a pixel of a different color than the starting pixel, flooding
will only stop when it hits the color specified by bordercolor. You
must specify a normal indexed color for the bordercolor. However, you
are free to use the gdTiled color for the fill.
Example:
# This has the same effect as the previous example $myImage->rectangle(10,10,100,100,$black); $myImage->fillToBorder(50,50,$black,$blue);
Two methods are provided for copying a rectangular region from one image to another. One method copies a region without resizing it. The other allows you to stretch the region during the copy operation.
With either of these methods it is important to know that the routines will attempt to flesh out the destination image's color table to match the colors that are being copied from the source. If the destination's color table is already full, then the routines will attempt to find the best match, with varying results.
$srcX,$srcY,$width,$height)
This is the simplest of the several copy operations, copying the specified region from the source image to the destination image (the one performing the method call). (srcX,srcY) specify the upper left corner of a rectangle in the source image, and (width,height) give the width and height of the region to copy. (dstX,dstY) control where in the destination image to stamp the copy. You can use the same image for both the source and the destination, but the source and destination regions must not overlap or strange things will happen.
Example:
$myImage = new GD::Image(100,100); ... various drawing stuff ... $srcImage = new GD::Image(50,50); ... more drawing stuff ... # copy a 25x25 pixel region from $srcImage to # the rectangle starting at (10,10) in $myImage $myImage->copy($srcImage,10,10,0,0,25,25);
Make a copy of the image and return it as a new object. The new image will look identical. However, it may differ in the size of the color palette and other nonessential details.
Example:
$myImage = new GD::Image(100,100);
... various drawing stuff ...
$copy = $myImage->clone;
$srcX,$srcY,$width,$height,$percent)
This copies the indicated rectangle from the source image to the destination image, merging the colors to the extent specified by percent (an integer between 0 and 100). Specifying 100% has the same effect as copy() -- replacing the destination pixels with the source image. This is most useful for highlighting an area by merging in a solid rectangle.
Example:
$myImage = new GD::Image(100,100); ... various drawing stuff ... $redImage = new GD::Image(50,50); ... more drawing stuff ... # copy a 25x25 pixel region from $srcImage to # the rectangle starting at (10,10) in $myImage, merging 50% $myImage->copyMerge($srcImage,10,10,0,0,25,25,50);
$srcX,$srcY,$width,$height,$percent)
This is identical to copyMerge() except that it preserves the hue of the source by converting all the pixels of the destination rectangle to grayscale before merging.
$srcX,$srcY,$destW,$destH,$srcW,$srcH)
This method is similar to copy() but allows you to choose different sizes for the source and destination rectangles. The source and destination rectangle's are specified independently by (srcW,srcH) and (destW,destH) respectively. copyResized() will stretch or shrink the image to accommodate the size requirements.
Example:
$myImage = new GD::Image(100,100); ... various drawing stuff ... $srcImage = new GD::Image(50,50); ... more drawing stuff ... # copy a 25x25 pixel region from $srcImage to # a larger rectangle starting at (10,10) in $myImage $myImage->copyResized($srcImage,10,10,0,0,50,50,25,25);
$srcX,$srcY,$destW,$destH,$srcW,$srcH)
This method is similar to copyResized() but provides "smooth" copying from a large image to a smaller one, using a weighted average of the pixels of the source area rather than selecting one representative pixel. This method is identical to copyResized() when the destination image is a palette image.
$srcX,$srcY,$width,$height,$angle)
Like copyResized() but the $angle argument specifies an arbitrary amount to rotate the image clockwise (in degrees). In addition, $dstX and $dstY species the center of the destination image, and not the top left corner.
Gd also provides some common image transformations:
GD allows you to draw characters and strings, either in normal horizontal orientation or rotated 90 degrees. These routines use a GD::Font object, described in more detail below. There are four built-in monospaced fonts, available in the global variables gdGiantFont, gdLargeFont, gdMediumBoldFont, gdSmallFont and gdTinyFont.
In addition, you can use the load() method to load GD-formatted bitmap font files at runtime. You can create these bitmap files from X11 BDF-format files using the bdf2gd.pl script, which should have been installed with GD (see the bdf_scripts directory if it wasn't). The format happens to be identical to the old-style MSDOS bitmap ".fnt" files, so you can use one of those directly if you happen to have one.
For writing proportional scaleable fonts, GD offers the stringFT() method, which allows you to load and render any TrueType font on your system.
This method draws a string starting at position (x,y) in the specified font and color. Your choices of fonts are gdSmallFont, gdMediumBoldFont, gdTinyFont, gdLargeFont and gdGiantFont.
Example:
$myImage->string(gdSmallFont,2,10,"Peachy Keen",$peach);
This method dynamically loads a font file, returning a font that you can use in subsequent calls to drawing methods. For example:
my $courier = GD::Font->load('./courierR12.fnt') or die "Can't load font";
$image->string($courier,2,10,"Peachy Keen",$peach);
Font files must be in GD binary format, as described above.
This method uses TrueType to draw a scaled, antialiased string using the TrueType vector font of your choice. It requires that libgd to have been compiled with TrueType support, and for the appropriate TrueType font to be installed on your system.
The arguments are as follows:
fgcolor Color index to draw the string in fontname A path to the TrueType (.ttf) font file or a font pattern. ptsize The desired point size (may be fractional) angle The rotation angle, in radians (positive values rotate counter clockwise) x,y X and Y coordinates to start drawing the string string The string itself
If successful, the method returns an eight-element list giving the boundaries of the rendered string:
@bounds[0,1] Lower left corner (x,y) @bounds[2,3] Lower right corner (x,y) @bounds[4,5] Upper right corner (x,y) @bounds[6,7] Upper left corner (x,y)
In case of an error (such as the font not being available, or FT support not being available), the method returns an empty list and sets $@ to the error message.
The string may contain UTF-8 sequences like: "À"
You may also call this method from the GD::Image class name, in which case it doesn't do any actual drawing, but returns the bounding box using an inexpensive operation. You can use this to perform layout operations prior to drawing.
Using a negative color index will disable antialiasing, as described in the libgd manual page at http://www.boutell.com/gd/manual2.0.9.html#gdImageStringFT.
An optional 8th argument allows you to pass a hashref of options to stringFT(). Several hashkeys are recognized: linespacing, charmap, resolution, and kerning.
The value of linespacing is supposed to be a multiple of the character height, so setting linespacing to 2.0 will result in double-spaced lines of text. However the current version of libgd (2.0.12) does not do this. Instead the linespacing seems to be double what is provided in this argument. So use a spacing of 0.5 to get separation of exactly one line of text. In practice, a spacing of 0.6 seems to give nice results. Another thing to watch out for is that successive lines of text should be separated by the "\r\n" characters, not just "\n".
The value of charmap is one of "Unicode", "Shift_JIS" and "Big5". The interaction between Perl, Unicode and libgd is not clear to me, and you should experiment a bit if you want to use this feature.
The value of resolution is the vertical and horizontal resolution, in DPI, in the format "hdpi,vdpi". If present, the resolution will be passed to the Freetype rendering engine as a hint to improve the appearance of the rendered font.
The value of kerning is a flag. Set it to false to turn off the default kerning of text.
Example:
$gd->stringFT($black,'/dosc/windows/Fonts/pala.ttf',40,0,20,90,
"hi there\r\nbye now",
{linespacing=>0.6,
charmap => 'Unicode',
});
If GD was compiled with fontconfig support, and the fontconfig library is available on your system, then you can use a font name pattern instead of a path. Patterns are described in fontconfig and will look something like this "Times:italic". For backward compatibility, this feature is disabled by default. You must enable it by calling useFontConfig(1) prior to the stringFT() call.
$image->useFontConfig(1);
For backward compatibility with older versions of the FreeType library, the alias stringTTF() is also recognized.
This draws text in a circle. Currently (libgd 2.0.33) this function does not work for me, but the interface is provided for completeness. The call signature is somewhat complex. Here is an excerpt from the libgd manual page:
Draws the text strings specified by top and bottom on the image, curved along the edge of a circle of radius radius, with its center at cx and cy. top is written clockwise along the top; bottom is written counterclockwise along the bottom. textRadius determines the "height" of each character; if textRadius is 1/2 of radius, characters extend halfway from the edge to the center. fillPortion varies from 0 to 1.0, with useful values from about 0.4 to 0.9, and determines how much of the 180 degrees of arc assigned to each section of text is actually occupied by text; 0.9 looks better than 1.0 which is rather crowded. font is a freetype font; see gdImageStringFT. points is passed to the freetype engine and has an effect on hinting; although the size of the text is determined by radius, textRadius, and fillPortion, you should pass a point size that "hints" appropriately -- if you know the text will be large, pass a large point size such as 24.0 to get the best results. fgcolor can be any color, and may have an alpha component, do blending, etc.
Returns a true value on success.
The alpha channel methods allow you to control the way drawings are processed according to the alpha channel. When true color is turned on, colors are encoded as four bytes, in which the last three bytes are the RGB color values, and the first byte is the alpha channel. Therefore the hexadecimal representation of a non transparent RGB color will be: C=0x00(rr)(bb)(bb)
When alpha blending is turned on, you can use the first byte of the color to control the transparency, meaning that a rectangle painted with color 0x00(rr)(bb)(bb) will be opaque, and another one painted with 0x7f(rr)(gg)(bb) will be transparent. The Alpha value must be >= 0 and <= 0x7f.
The alphaBlending() method allows for two different modes of drawing
on truecolor images. In blending mode, which is on by default (libgd
2.0.2 and above), the alpha channel component of the color supplied to
all drawing functions, such as setPixel, determines how much of the
underlying color should be allowed to shine through. As a result, GD
automatically blends the existing color at that point with the drawing
color, and stores the result in the image. The resulting pixel is
opaque. In non-blending mode, the drawing color is copied literally
with its alpha channel information, replacing the destination
pixel. Blending mode is not available when drawing on palette images.
Pass a value of 1 for blending mode, and 0 for non-blending mode.
saveAlpha(1) to turn on saving of such information, and call
alphaBlending(0) to turn off alpha blending within the library so
that alpha channel information is actually stored in the image rather
than being composited immediately at the time that drawing functions
are invoked.
These are various utility methods that are useful in some circumstances.
Compare two images and return a bitmap describing the differences found, if any. The return value must be logically ANDed with one or more constants in order to determine the differences. The following constants are available:
GD_CMP_IMAGE The two images look different GD_CMP_NUM_COLORS The two images have different numbers of colors GD_CMP_COLOR The two images' palettes differ GD_CMP_SIZE_X The two images differ in the horizontal dimension GD_CMP_SIZE_Y The two images differ in the vertical dimension GD_CMP_TRANSPARENT The two images have different transparency GD_CMP_BACKGROUND The two images have different background colors GD_CMP_INTERLACE The two images differ in their interlace GD_CMP_TRUECOLOR The two images are not both true color
The most important of these is GD_CMP_IMAGE, which will tell you whether the two images will look different, ignoring differences in the order of colors in the color palette and other invisible changes. The constants are not imported by default, but must be imported individually or by importing the :cmp tag. Example:
use GD qw(:DEFAULT :cmp);
# get $image1 from somewhere
# get $image2 from somewhere
if ($image1->compare($image2) & GD_CMP_IMAGE) {
warn "images differ!";
}
A few primitive polygon creation and manipulation methods are provided. They aren't part of the Gd library, but I thought they might be handy to have around (they're borrowed from my qd.pl Quickdraw library). Also see GD::Polyline.
Create an empty polygon with no vertices.
$poly = new GD::Polygon;
Add point (x,y) to the polygon.
$poly->addPt(0,0); $poly->addPt(0,50); $poly->addPt(25,25); $myImage->fillPoly($poly,$blue);
Retrieve the point at the specified vertex.
($x,$y) = $poly->getPt(2);
Change the value of an already existing vertex. It is an error to set a vertex that isn't already defined.
$poly->setPt(2,100,100);
Delete the specified vertex, returning its value.
($x,$y) = $poly->deletePt(1);
Draw from current vertex to a new vertex, using relative (dx,dy) coordinates. If this is the first point, act like addPt().
$poly->addPt(0,0); $poly->toPt(0,50); $poly->toPt(25,-25); $myImage->fillPoly($poly,$blue);
Return the number of vertices in the polygon.
$points = $poly->length;
Return a list of all the vertices in the polygon object. Each member of the list is a reference to an (x,y) array.
@vertices = $poly->vertices;
foreach $v (@vertices)
print join(",",@$v),"\n";
}
Return the smallest rectangle that completely encloses the polygon. The return value is an array containing the (left,top,right,bottom) of the rectangle.
($left,$top,$right,$bottom) = $poly->bounds;
Offset all the vertices of the polygon by the specified horizontal (dh) and vertical (dy) amounts. Positive numbers move the polygon down and to the right.
$poly->offset(10,30);
Map the polygon from a source rectangle to an equivalent position in a destination rectangle, moving it and resizing it as necessary. See polys.pl for an example of how this works. Both the source and destination rectangles are given in (left,top,right,bottom) coordinates. For convenience, you can use the polygon's own bounding box as the source rectangle.
# Make the polygon really tall $poly->map($poly->bounds,0,0,50,200);
Please see GD::Polyline for information on creating open polygons and splines.
The libgd library (used by the Perl GD library) has built-in support for about half a dozen fonts, which were converted from public-domain X Windows fonts. For more fonts, compile libgd with TrueType support and use the stringFT() call.
If you wish to add more built-in fonts, the directory bdf_scripts contains two contributed utilities that may help you convert X-Windows BDF-format fonts into the format that libgd uses internally. However these scripts were written for earlier versions of GD which included its own mini-gd library. These scripts will have to be adapted for use with libgd, and the libgd library itself will have to be recompiled and linked! Please do not contact me for help with these scripts: they are unsupported.
Each of these fonts is available both as an imported global (e.g. gdSmallFont) and as a package method (e.g. GD::Font->Small).
This returns the number of characters in the font.
print "The large font contains ",gdLargeFont->nchars," characters\n";
height
These return the width and height of the font.
($w,$h) = (gdLargeFont->width,gdLargeFont->height);
libgd, the C-language version of gd, can be obtained at URL http://www.boutell.com/gd/. Directions for installing and using it can be found at that site. Please do not contact me for help with libgd.
The GD.pm interface is copyright 1995-2007, Lincoln D. Stein. It is distributed under GPL and the Artistic License 2.0.
The latest versions of GD.pm are available at
http://stein.cshl.org/WWW/software/GD