Add support for extended grapheme clusters.

[salt-die]

Add support for extended grapheme clusters (egcs).

To facilitate egcs, the Cell type was changed from:

Cell = np.dtype(
    [
        ("char", "U1"),
        ("bold", "?"),
        ("italic", "?"),
        ("underline", "?"),
        ("strikethrough", "?"),
        ("overline", "?"),
        ("reverse", "?"),
        ("fg_color", "u1", (3,)),
        ("bg_color", "u1", (3,)),
    ]
)

to

Cell = np.dtype(
    [
        ("ord", "uint32"),
        ("style", "u1"),
        ("fg_color", "u1", (3,)),
        ("bg_color", "u1", (3,)),
    ]
)

The "style" field is now a bit field of the SGR parameters of a terminal cell. The style int flags are given by:

  class Style(IntFlag):
    DEFAULT = 0
    BOLD = 0b1
    ITALIC = 0b10
    UNDERLINE = 0b100
    STRIKETHROUGH = 0b1000
    OVERLINE = 0b10000
    REVERSE = 0b100000

The "ord" field now stores the unicode codepoint for a cell or, if the cell is an extended grapheme cluster (egc), an index into a pool of egcs (plus EGC_BASE to mark the codepoint as an egc; EGC_BASE is greater than sys.maxunicode).

Additional changes:

• Text gadgets now have a chars property that re-views the "ord" field of their canvas as unicode strings to allow characters to be directly added to their canvas as before, e.g.,

  my_text_gadget.chars[0, 0] = "a"

  which is equivalent to

  my_text_gadget.canvas["ord"][0, 0] = ord("a")

  This method should NOT be used to add egcs to the canvas.
• To add an egc directly to a Text canvas, use the new function text_tools.egc_ord, e.g.,

  my_text_gadget.canvas[0, 0]["ord"] = egc_ord("👩🏽‍🔬")

  However without taking into account the egc column width, this could clip with adjacent cells, so it's recommended to continue to use text_tools.add_text or Text.add_str instead which now account for egcs, e.g.,

  my_text_gadget.add_str("👩🏽‍🔬")

• The inverse of egc_ord is also provided in text_tools, egc_char.
• Added the excellent dependencies uwcwidth and ugrapheme and dropped the custom char_width and str_width functions.