Sizing: fit large images, but do not zoom small ones

[candidtim]

Hi! Is it possible to configure the widget such that a large image is downscaled to fit the available space, but a smaller one does not zoom in?

When setting width and/or height to auto, larger images are scaled to fit the available space, but so are the small ones - they are effectively zoomed in. When keeping width and height at none, the images keep their original size - meaning that larger images do not fit and overflow.

Is it possible to have an image container that displays any image in full and keeps the image original size, unless its too big in which case it is downscaled to fit the available space?

Thank you!

[lnqs]

Hey! I'm afraid this is not possible with the CSS rules. I'm also not sure how to model that on the API in a reasonable way if we wanted to add it.
For now, I guess calculating the desired size in your code and setting width/height to cell values is the best way to go.

[candidtim]

OK, thanks! So, I ended up with something along the lines of this:

image: PIL.Image = ...
widget: textual_image.widget import Image = ...

cell_size = textual_image._terminal.get_cell_size()

width = math.floor(image.width / cell_size.width)
height = math.floor(image.height / cell_size.height)

if width > content_size.width:
    widget.styles.width = content_size.width
    widget.styles.height = "auto"
elif height > content_size.height:
    widget.styles.width = "auto"
    widget.styles.height = content_size.height
else:
    widget.styles.width = width
    widget.styles.height = height

The only part I don't like about it is using the textual_image._terminal private API. Is there a more robust way to access this function?

I understand your point about CSS, I agree that it would be too limited to express "fit" or "shrink". Maybe an alternative is to add a method in the Image class, or in a separate module to not pollute Image class with it? Or do you prefer keeping it outside of the textual-image?

[lnqs]

The only part I don't like about it is using the textual_image._terminal private API. Is there a more robust way to access this function?

Fair point. You're right, calling private APIs is not exactly recommended generally.

I understand your point about CSS, I agree that it would be too limited to express "fit" or "shrink". Maybe an alternative is to add a method in the Image class, or in a separate module to not pollute Image class with it? Or do you prefer keeping it outside of the textual-image?

I'd be happy to add it, given we have a reasonable API for it. Could you elaborate a bit more how you image it to be used?