Pycairo vs. cairocffi vs. Qahirah

Question

I want to create and rasterize vector graphics in Python. I suspect that Pycairo or cairocffi (edit: or Qahirah ) are a great choice. (If not, comments are welcome.)

What are the practical differences between the two?

Specifically, the Pycairo documentation says:

If Pycairo is not what you need, have a look at cairocffi, which is an API compatible package using cffi or Qahirah, which is using ctypes and provides a more "pythonic" API with less focus on matching the cairo C API.

But this raises some questions: In what cases may Pycairo be "not what you need", whereas cairocffi is? In what way are cffi/Qahirah/ctypes better than whatever Pycairo does instead? In what ways is Pycairo not "pythonic"? If cairocffi is better than Pycairo, why is Pycairo more popular, does it have advantages?

Edit: A comma might be missing after "cffi" in the quote above. In that case, it's not about "Pycairo vs. cairocffi", but about "Pycairo vs. cairocffi vs. Qahirah".

Answer 1

Mimicking the C API closely means that functions take simple arguments. As an example, the move_to function/method looks like this:

ctx.move_to(x: float, y: float)→ None

If you want to use (x,y) points or vectors in your program, then at some point you might get annoyed that you need to write, say, move_to(P[0], P[1]) instead of just move_to(P) .

In Qahirah you write this:

p = Vector(x, y)
ctx.move_to(p)

or perhaps

ctx.move_to(Vector(x, y))

or even

ctx.move_to((x, y))

You can think of the C API as a base level on top it is possible to build more convenient API.

As always, it is a trade-off. Using a direct style C API might be faster (but not much).

More information can be found in the README for Qahirah:

https://github.com/ldo/qahirah 

Then again - if you find a Cairo tutorial using C, it might be easier to use the direct style C API.

Try them both and see what you like.