simoncozens / tensorfont Goto Github PK

Please update the support for numpy version of the package. When I call the function as_matrix() for the glyph the numpy library broke the system for deprecated function.

Numpy no longer uses np.float, but float built-in

There are a couple of doc strings that I think are not as clear as they could be.

For example, in the Font properties dealing with ascenders & descenders, the docs say things like "The font’s ascender height, in font units and pixels."

From a non–font-engineer's perspective, I worry that could be ambiguous as to whether it's referring to "baseline to the top of d/h/l/b", "x-height to the top of d/h/l/b", hheaAscender, typoAscender, or winAscent. So it'd be better to be explicit.

If I understand the intent of the Font class, its properties are meant to capture "inherent" geometric/optical characteristics.

Assuming that's the case, I'd think there are a few additional ones to consider — such as stroke & aperture metrics. Certainly the width of vertical strokes. I would actually think that the width of n would be more immediately useful than m, but I may be misinterpreting the goal.

Essentially, what the set of valuable properties is could be an endless debate, but for classifying typefaces I've been trying to stick to the TypeNetwork parametric-axes model, and I think that works pretty well. So would it make sense to use that as a guide?

It'd essentially be adding horizontal & vertical stem measurements for capitals and lowercase, plus counter widths for capitals & lowercase. That gives you context-free way to measure weight and width (well, in Latin and sister scripts, anyway).