text_width property of FontMetric doesn't seem accurate

[ProfessorSalty]

I'm trying to add text to an image that I'm making from a resized SVG. I've got a simple algorithm to figure out the largest font size that will fit in the final image, but I seem to be misunderstanding the properties in FontMetric. Here's my minimum working example:

#!/usr/bin/env python3
from wand.color import Color
from wand.image import Image
from wand.drawing import Drawing


def generate_message(message: str):
    with Image(filename='Opensource.svg', resolution=4096, background=Color('transparent')) as image:
        image.resize(2400, 1697 * 2)
        image.extent(width=2400, height=1697 * 2,gravity='south')
        with Drawing() as draw:
            draw.gravity = 'north'
            draw.font_size = 10
            while True:
                metrics = draw.get_font_metrics(image, message)
                #if metrics.text_width >= image.width: #this also doesn't work
                if metrics.text_width >= 2048: #checking the width of the text here against the desired width
                    draw.font_size -= 0.1
                else:
                    break
            draw.text(0, 0, message)
            draw(image)
        image.save(filename='test.png')


if __name__ == '__main__':
    generate_message('hey yous guys')

Here's the SVG I'm using

Does this make sense or am I thinking of this wrong? I expect the text to be as large as it can be without going past the left or right edges. Is there a better way to do this?

[emcconville]

Is there a better way to do this?

Have you tried just generating the text as a new image, and composing over the source image?

from wand.image import Image

kFontPath = '/usr/share/fonts/google-noto-vf/NotoSans-VF.ttf'


def faster_generate_message(msg):
    with Image(filename='wizard:') as img:
        dim = img.size
        with Image() as txt:
            # Set typeface properties before reading label:
            txt.background_color = 'transparent'
            txt.font_color = 'magenta'
            txt.font_path = kFontPath
            # Picking either the min / max will generate a real large image.
            txt.font_size = min(*dim)
            txt.read(filename='label:' + msg)
            # If we want message to hit the border of the image.
            txt.trim(fuzz=0.1*txt.quantum_range)
            # Resize respecting the largest side.
            txt.transform(resize='{0}x{1}>'.format(*dim))
            # Compose over the original image.
            img.composite(txt, gravity='center')
        # Just to show the image when displayed online.
        img.border('cyan', 1, 1)
        img.save(filename='output.png')


if __name__ == '__main__':
    faster_generate_message('hey yous guys\nI\'ve got multiple lines!')

[ProfessorSalty]

That is incredible! Thank you!