Compress bitmaps by a factor of two

Compresses bm_bitmap() objects by a factor of two by re-mapping to a “block elements” scheme. For pixmap objects like bm_pixmap() we simply shrink the pixmap by a factor of two using bm_distort().

Usage

bm_compress(x, direction = "vertical", ...)

# S3 method for class 'bm_bitmap'
bm_compress(x, direction = "vertical", ...)

# S3 method for class 'bm_pixmap'
bm_compress(x, direction = "vertical", ..., filter = "Point")

# S3 method for class '`magick-image`'
bm_compress(x, direction = "vertical", ..., filter = "Point")

# S3 method for class 'nativeRaster'
bm_compress(x, direction = "vertical", ..., filter = "Point")

# S3 method for class 'raster'
bm_compress(x, direction = "vertical", ..., filter = "Point")

# S3 method for class 'bm_list'
bm_compress(x, ...)

Arguments

x

Either a bm_bitmap(), bm_font(), bm_list(), "magick-image", "nativeRaster", bm_pixmap(), or "raster" object.

direction

Either "vertical" or "v", "horizontal" or "h", OR "both" or "b".

...

Additional arguments to be passed to or from methods.

filter

Passed to magick::image_resize(). Use magick::filter_types() for list of supported filters. The default "Point" filter will maintain your sprite's color palette. NULL will give you the magick's default filter which may work better if you are not trying to maintain a sprite color palette.

Value

Depending on x either a bm_bitmap(), bm_font(), bm_list(), magick-image, "nativeRaster", bm_pixmap(), or raster object.

Details

Depending on direction we shrink the bitmaps height and/or width by a factor of two and re-encode pairs/quartets of pixels to a “block elements” scheme. If necessary we pad the right/bottom of the bitmap(s) by a pixel. For each pair/quartet we determine the most-common non-zero element and map them to a length twenty set of integers representing the “block elements” scheme. For integers greater than zero we map it to higher twenty character sets i.e. 1's get mapped to 0:19, 2's get mapped to 20:39, 3's get mapped to 40:59, etc. Using the default px_unicode will give you the exact matching “Block Elements” glyphs while px_ascii gives the closest ASCII approximation. Hence print.bm_bitmap() should produce reasonable results for compressed bitmaps if either of them are used as the px argument.

See also

See https://en.wikipedia.org/wiki/Block_Elements for more info on the Unicode Block Elements block.

Examples

font_file <- system.file("fonts/spleen/spleen-8x16.hex.gz", package = "bittermelon")
font <- read_hex(font_file)
r <- font[[str2ucp("R")]]
print(r)
#> ░░░░░░░░
#> ░░░░░░░░
#> ██████░░
#> ██░░░██░
#> ██░░░██░
#> ██░░░██░
#> ██████░░
#> ██░░░██░
#> ██░░░██░
#> ██░░░██░
#> ██░░░██░
#> ██░░░██░
#> ░░░░░░░░
#> ░░░░░░░░
#> ░░░░░░░░
#> ░░░░░░░░
print(bm_compress(r, "vertical"))
#>         
#> ██▀▀▀█▄ 
#> ██   ██ 
#> ██▀▀▀█▄ 
#> ██   ██ 
#> ██   ██ 
#>         
#>         
print(bm_compress(r, "horizontal"))
#>     
#>     
#> ███ 
#> █ ▐▌
#> █ ▐▌
#> █ ▐▌
#> ███ 
#> █ ▐▌
#> █ ▐▌
#> █ ▐▌
#> █ ▐▌
#> █ ▐▌
#>     
#>     
#>     
#>     
print(bm_compress(r, "both"))
#>     
#> █▀▜▖
#> █ ▐▌
#> █▀▜▖
#> █ ▐▌
#> █ ▐▌
#>     
#>     

img <- png::readPNG(system.file("img", "Rlogo.png", package="png"))
logo <- as_bm_pixmap(img)
if (cli::is_utf8_output() && 
    cli::num_ansi_colors() > 256L &&
    requireNamespace("magick", quietly = TRUE)) {
  logo_c <- bm_compress(pm, "both", filter = NULL)
  print(logo_c, compress = "v")
}