Is there a way to use skia.Surface to export high quality images? about skia-python

[HinTak]

I think you generally get an image which is the same pixel size as the declared dimension of the surface. So one answer to your question is probably just drawing on a larger surface. One other possibility is the SamplingOptions class, which is a translation of the upstream SkSamplingOptions class. This is used as an argument to many image-related methods. In skia-python we only provides the default at the moment, but it won't surprise me that upstream has a resolution-doubling etc option. If you look at the upstream documentation and see if you can find what you need there, we can put it in the next (122) release. A 3rd possibility is to output as one of the vector formats (svg, ps, pdf etc) and rasterize externally with mupdf/batik/ghostscript. For that, you need to worry about pen width, etc.

[HinTak]

The SamplingOptions class seems to have a number of enums, kLinear and kNearest for interpolation. I don't know which might be relevant to you (or none), but if you are trying to draw a bitmap onto a surface that is not exactly the same pixel dimension as the bitmap, you might find changing from kLinear to KNearest or vice versa more suited to your needs.

[adelhpour]

Thanks for your quick reply.
What I am actually looking for is that how I can use the following methods mentioned in the documentation:

image = surface.makeImageSnapshot()
image.save('output.png', skia.kPNG)

and get a figure saved with a higher quality than the one I currently get? Is there a way like increasing the dpi of the created image there?

[HinTak]

Well, I wrote everything in my initial reply already: easiest would be starting from a larger surface. 2nd way is to use image.resize(), which takes a SamplingOption argument for interpolation. If the current python interface to the upstream SkSamplingOption class is insufficient and you see things upstream we could add, we'll add it for next release. Upstream docs at api.skia.org or something similar.

[HinTak]

https://api.skia.org/structSkSamplingOptions.html