How to translate text into English image?

[mrcego]

According commit #d92d7f6, we should translate English text into images in docs. How to do that? Any guide about this? I'm interested.

[tony19]

I'm wondering myself. 🙂 I'll have to update the steps with instructions once I figure it out.

[tony19]

It looks like Evan created the PNGs a while back ({bundler,esm,graph}.png). I think we should convert them to SVGs to enable translations.

[mrcego]

What about this kind of images? I was testing some tools to parse SVG path from images and standards isn't the same at all. Trying to edit SVG Path file later was ponderous and need time. IMHO, there are two scenarios here:

1. Convince Evan to create shareable SVG files with a tool of his preference and test several tools to edit that files, or
2. Create a tool satisfying both issues (shareable and editable SVG files with global standard).

While we are discussing this, I will continue to search and test the correct tools to achieve this with old images.

[tony19]

I think even those images (which are just graphs with text) can be recreated in SVG. Once we convert those PNGs to SVG, we could easily maintain the translations in the SVGs going forward. We don't need to create a tool to auto convert the PNGs (that would be overkill IMO).

[mrcego]

Yeah, I suggested second option just in case with.

I have tested Adobe Express online tool and convert images with an OK fidelity quality. Now, I'm searching a graphic editor tool for generated SVG.

[tony19]

I re-drew the PNGs into SVGs with Figma. Translations can be applied to the SVGs themselves by editing them directly. If you need to expand boxes for your translation to fit the text, you could also use Figma or any other SVG editor to adjust the diagram.

related: vitejs/vite#9208

[mrcego]

I was playing a little, trying to understand how to vectorize PNG's. Some advices to get SVG files with at least a similar quality from source?

[tony19]

I'm not aware of a tool that creates perfect SVGs from PNGs. There's a Figma plugin (Vectorized) that traces the outlines of an image, but the output would have to be colored in.