Generating text over transparent background

It looks like one of the most basic tasks and yet it seems quite hard to do at least without reading through all of the documentation and looking at only the related methods and classes.

def output(filename, text, color):
    image, feedback = pyvips.Image.text(
        text, font='Liberation-Sans-Narrow', width=32, height=32,
        autofit_dpi=True)
    image = image.colourspace('srgb')
    image = image.addalpha()
    image = image.ifthenelse([0, 0, 0, 255], [255, 255, 255, 0], blend=True)
    image.write_to_file(f'{DIR}/{filename}.png')

I guess I’m not working correctly with ifthenelse. But what are the right values to make the image transparent? Where is the alpha channel supposed to be?

Also, the resulting image is not 32x32, it may be good to add hints on how to make it the required size somewhere around here.

Additional context #203 https://github.com/libvips/libvips/issues/1208

[jcupitt]

a.embed(x, y, width, height) will position a at (x, y) within an image of size (width, height).

[jcupitt]

Hello again, you could make this a little simpler: use the text for the alpha and have a solid colour in RGB.

For example:

alpha, feedback = pyvips.Image.text('Hello World!', 
        font='Liberation-Sans-Narrow', width=32, height=32,
        autofit_dpi=True)
alpha = alpha.gravity('centre', 32, 32)
image = alpha.new_from_image([0, 0, 255]).bandjoin(alpha)

The gravity will expand the alpha up to 32x32 if it’s smaller.

text uses 0 - 255 for black - white. If you want to make the text semi-transparent, you’ll need to set a smaller range, perhaps 0 - 64? You could write:

image = alpha.new_from_image([0, 0, 255]).bandjoin(alpha * 0.25)