Incorporate manual modifications of font data to gen_font.py

[obrusvit]

If we run core/tools/codegen/gen_font.py script, it creates a diff in generated font definitions. This is because of manual changes done to individual glyps for better visual appearance on certain models. The changes are documented by commentary but IMHO they should be defined in the script so that manual changes to generated code are not necessary.

Changes are done to

• Caesar UI:
  • font_unifont_regular_16
  • font_unifont_bold_16 - no idea how these changes were produced?
  • font_pixeloperatormono_regular_8
  • font_pixeloperator_regular_8
  • font_pixeloperator_bold_8
• Delizia UI:
  • font_ttsatoshi_demibold_18.c and font_ttsatoshi_demibold_21.c - alternative numeral glyphs used
  • font_ttsatoshi_demibold_42.c - uppercase fonts completely removed to save space - they are not used anywhere at the moment

[obrusvit]

Note on font_ttsatoshi_demibold_18/21.c. Numerals (0 - 9 mapped to glyph indices 1298-1307 as of #3954 . Done manually by modifying gen_font.py as follows. A systematic solution would be better.

def _load_char(self, c: str) -> None:
        glyph_map = {
            '0': 1298,
            '1': 1299,
            '2': 1300,
            '3': 1301,
            '4': 1302,
            '5': 1303,
            '6': 1304,
            '7': 1305,
            '8': 1306,
            '9': 1307,
        }

        if c in glyph_map:
            self.face.load_glyph(glyph_map[c], freetype.FT_LOAD_RENDER | freetype.FT_LOAD_TARGET_NORMAL)
        else:
            self.face.load_char(c, freetype.FT_LOAD_RENDER | freetype.FT_LOAD_TARGET_NORMAL)  # type: ignore