Support text transforms

[kkaefer]

We should allow text transforms (uppercase/lowercase/small caps?) when creating the vector tiles.

[nickidlugash]

This would great, especially while only one font stack is supported.

[nickidlugash]

This could be useful for the Outdoors style. Since it is unlikely that multiple font stacks will be supported in time for WWDC, this would allow for more differentiation between similar sized/colored labels.

Tracking for Outdoors: https://github.com/mapbox/gl-ported-maps/issues/2

[mikemorris]

Multiple fontstacks are now supported.

[nickidlugash]

Is this something that could be possible to implement soon?

[kkaefer]

Hrm, text transformation is another can of worms, e.g. in German, you translate the lowercase "ß" to "SS". For small caps, we need a font that supports small caps and replace the code points with different glyphs (i.e. no 1:1 mapping between unicode points and glyphs).

The Unicode consortium helpfully has a list of case conversions: http://www.unicode.org/charts/case/

[yhahn]

@kkaefer can we start with something dumb, ie. the what js is doing when you

> ('Hello World').toUpperCase();
'HELLO WORLD'
> ('Hello World').toLowerCase();
'hello world'

? Will it be weird/hard to achieve parity to this on native?

[kkaefer]

ICU's C++ libraries have code for handling that. We're currently using boost-unicode on systems that don't have unicode support in stdlib. I know @springmeyer wants to switch to https://github.com/dropbox/miniutf, but this library doesn't seem to have support for uppercasing/lowercasing/titlecasing. We could contribute the code though. Using ICU is not something I'd like to do, since it is quite a large dependency.

[yhahn]

word

[springmeyer]

yes, unicode regex and upper/lower/title casing is about all Mapnik really needs/uses ICU for.

[kkaefer]

It looks like miniutf has lowercase support but no titlecase or uppercase support.

[mikemorris]

With http://www.unicode.org/charts/case/ it doesn't look like it would be too hard to add uppercase/titlecase support to miniutf, either in a fork or PR.

[mikemorris]

miniutf implements http://www.unicode.org/Public/6.3.0/ucd/UnicodeData.txt directly at https://raw.githubusercontent.com/dropbox/miniutf/master/data-6.3.0/UnicodeData.txt, which does already contain support for titlecase and uppercase mapping, we would just need to add a method to use that data. We may want to add http://www.unicode.org/Public/6.3.0/ucd/SpecialCasing.txt too if necessary.

0061;LATIN SMALL LETTER A;Ll;0;L;;;;;N;;;0041;;0041

[kkaefer]

@mikemorris Yeah, the data is in the text file, but it's currently not in the cpp files that get compiled. The preprocess script does that, but it's currently commented out: https://github.com/dropbox/miniutf/blob/73b871ff1e6f5e716f9da90f64bd9c19fea18dcc/preprocess.py#L340

Also see dropbox/miniutf#2