table of contents
Prima::ImageViewer(3) | User Contributed Perl Documentation | Prima::ImageViewer(3) |
NAME¶
Prima::ImageViewer - standard image, icon, and bitmap viewer class.
SYNOPSIS¶
use Prima qw(ImageViewer StdBitmap Application); Prima::ImageViewer-> new( image => Prima::StdBitmap::image(0), zoom => 2.718, ); run Prima;
DESCRIPTION¶
The module contains "Prima::ImageViewer" class, which provides image displaying functionality, including different zoom levels.
"Prima::ImageViewer" is a descendant of "Prima::Widget::ScrollWidget" and inherits its document scrolling behavior and programming interface. See Prima::Widget::ScrollWidget for details.
API¶
Properties¶
- alignment INTEGER
- One of the following "ta::XXX"
constants:
ta::Left ta::Center ta::Right
Selects the horizontal image alignment.
Default value: "ta::Left"
- autoZoom BOOLEAN
- When set, the image is automatically stretched while keeping aspects to the best available fit, given the "zoomPrecision". Scrollbars are turned off if "autoZoom" is set to 1.
- image OBJECT
- Selects the image object to be displayed. OBJECT can be an instance of "Prima::Image", "Prima::Icon", or "Prima::DeviceBitmap" class.
- imageFile FILE
- Set the image FILE to be loaded and displayed. Is rarely used since does not return a loading success flag.
- scaling ist::XX
- Applies scaling when drawing an image.
Default: "ist::Box", default cheap scaling.
Warning: scaling types above the "ist::Box" might be somewhat expensive
- stretch BOOLEAN
- If set, the image is simply stretched over the visual area, without keeping the aspect. Scroll bars, zooming and keyboard navigation become disabled.
- quality BOOLEAN
- A boolean flag, selecting if the palette of
"image" is to be copied into the widget
palette, providing higher visual quality on paletted displays. See also
"palette" in Prima::Widget.
Default value: 1
- valignment INTEGER
- One of the following "ta::XXX"
constants:
ta::Top ta::Middle or ta::Center ta::Bottom
Selects the vertical image alignment.
NB: "ta::Middle" value is not equal to "ta::Center"'s, however the both constants produce equal effect here.
Default value: "ta::Bottom"
- zoom FLOAT
- Selects zoom level for image display. The acceptable value range is
between 0.01 and 100. The zoom value is rounded to the closest value
divisible by 1/"zoomPrecision". For
example, is "zoomPrecision" is 100, the
zoom values will be rounded to the precision of hundredth - to fiftieth
and twentieth fractional values - .02, .04, .05, .06, .08, and 0.1 . When
"zoomPrecision" is 1000, the precision
is one thousandth, and so on.
Default value: 1
- zoomPrecision INTEGER
- Zoom precision of "zoom" property.
Minimal acceptable value is 10, where zoom will be rounded to 0.2, 0.4,
0.5, 0.6, 0.8 and 1.0 .
The reason behind this arithmetics is that when an image of an arbitrary zoom factor is requested to be displayed, the image sometimes must be drawn from a fraction image pixel - for example, 10x zoomed image shifted 3 pixels left, must be displayed so the first image pixel from the left occupies 7 screen pixels, and the next ones - 10 screen pixels. That means, that the correct image display routine must ask the system to draw the image at offset -3 screen pixels, where the first image pixel column would correspond to that offset.
When the zoom factor is fractional, the picture is getting more complex. For example, with zoom factor 12.345, and zero screen offset, the first image pixel begins at the 12th screen pixel, the next one - at the 25th ( because of the roundoff ), then the 37th etc etc. If the image is 2000x2000 pixels wide, and is asked to be drawn so that it appears shifted 499 screen image pixels left, it needs to be drawn from the 499/12.345=40.42122th image pixel. Is might seem that indeed it would be enough to ask the system to begin drawing from image pixel 40, and offset int(0.42122*12.345)=5 screen pixels to the left, however, that procedure will not account for the correct fixed point roundoff that accumulates as system scales the image. For zoom factor 12.345 this roundoff sequence is, as we seen before, (12,25,37,49,62,74,86,99,111,123) for the first 10 pixels displayed, that occupy (12,13,12,12,13,12,12,13,12,12) screen pixels correspondingly. For the pixels starting at 499, the sequence is (506,519,531,543,556,568,580,593,605,617) offsets or (13,12,12,13,13,12,12,13,12,12) widths -- note the two subsequent 13s there. This sequence begins to repeat itself after 200 iterations (12.345*200=2469.000), which means that in order to achieve correct display results, the image must be asked to be displayed from as far as image pixel 0 if image's first pixel on the screen is between 0 and 199 ( or for screen pixels 0-2468), then from image pixel 200 for offsets 200-399, ( screen pixels 2469-4937), and so on.
Since the system internally allocates memory for image scaling, that means that up to 2*200*min(window_width,image_width)*bytes_per_pixel unneccessary bytes will be allocated for each image drawing call (2 because the calculations are valid for both the vertical and horizontal strips), and this can lead to slowdown or even request failure when image or window dimensions are large. The proposed solution is to roundoff accepted zoom factors, so these offsets are kept small - for example, N.25 zoom factors require only max 1/.25=4 extra pixels. When "zoomPrecision" value is 100, zoom factors are rounded to 0.X2, 0.X4, 0.X5, 0.X6, 0.X8, 0.X0, thus requiring max 50 extra pixels.
NB. If, despite the efforts, the property gets in the way, increase it to 1000 or even 10000, but note that this may lead to problems.
Default value: 100
Methods¶
- on_paint SELF, CANVAS
- The "Paint" notification handler is
mentioned here for the specific case of its return value, that is the
return value of internal "put_image"
call. For those who might be interested in
"put_image" failures, that mostly occur
when trying to draw an image that is too big, the following code might be
useful:
sub on_paint { my ( $self, $canvas) = @_; warn "put_image() error:$@" unless $self-> SUPER::on_paint($canvas); }
- screen2point X, Y, [ X, Y, ... ]
- Performs translation of integer pairs integers as (X,Y)-points from widget
coordinates to pixel offset in image coordinates. Takes in account zoom
level, image alignments, and offsets. Returns array of same length as the
input.
Useful for determining correspondence, for example, of a mouse event to a image point.
The reverse function is "point2screen".
- point2screen X, Y, [ X, Y, ... ]
- Performs translation of integer pairs as (X,Y)-points from image pixel
offset to widget image coordinates. Takes in account zoom level, image
alignments, and offsets. Returns array of same length as the input.
Useful for determining a screen location of an image point.
The reverse function is "screen2point".
- watch_load_progress IMAGE
- When called, image viewer watches as IMAGE is being loaded ( see
"load" in Prima::Image ) and displays the progress. As soon as
IMAGE begins to load, it replaces the existing
"image" property. Example:
$i = Prima::Image-> new; $viewer-> watch_load_progress( $i); $i-> load('huge.jpg'); $viewer-> unwatch_load_progress;
Similar functionality is present in Prima::Dialog::ImageDialog.
- unwatch_load_progress CLEAR_IMAGE=1
- Stops monitoring of image loading progress. If CLEAR_IMAGE is 0, the leftovers of the incremental loading stay intact in "image" propery. Otherwise, "image" is set to "undef".
- zoom_round ZOOM
- Rounds the zoom factor to "zoomPrecision" precision, returns the rounded zoom value. The algorithm is the same as used internally in "zoom" property.
AUTHOR¶
Dmitry Karasik, <dmitry@karasik.eu.org>.
SEE ALSO¶
Prima, Prima::Image, Prima::Widget::ScrollWidget, Prima::Dialog::ImageDialog, examples/iv.pl.
2023-09-05 | perl v5.26.1 |