| 1 | package i18n |
| 2 | |
| 3 | import "strings" |
| 4 | |
| 5 | // Transliterate writes English phonetically in Sumero-Akkadian |
| 6 | // cuneiform (Unicode U+12000). Digits and punctuation stay, so a |
| 7 | // sentence keeps its shape. Printf verbs (%s, %d, %q, %%) stay |
| 8 | // literal, so a dialog can still substitute a name. The spelling |
| 9 | // is syllabic: consonants without a following vowel take an |
| 10 | // inherent a, o is written u, and there is no separate sign for |
| 11 | // f, v or o. igh and eigh are the same diphthong as ai. |
| 12 | func Transliterate(s string) string { |
| 13 | var b strings.Builder |
| 14 | word := strings.Builder{} |
| 15 | flush := func() { |
| 16 | if word.Len() == 0 { |
| 17 | return |
| 18 | } |
| 19 | b.WriteString(encodeWord(word.String())) |
| 20 | word.Reset() |
| 21 | } |
| 22 | rs := []rune(s) |
| 23 | for i := 0; i < len(rs); i++ { |
| 24 | if n := printfVerb(rs[i:]); n > 0 { |
| 25 | flush() |
| 26 | b.WriteString(string(rs[i : i+n])) |
| 27 | i += n - 1 |
| 28 | continue |
| 29 | } |
| 30 | r := rs[i] |
| 31 | if isLetter(r) || r == '\'' { |
| 32 | word.WriteRune(r) |
| 33 | continue |
| 34 | } |
| 35 | flush() |
| 36 | b.WriteRune(r) |
| 37 | } |
| 38 | flush() |
| 39 | return b.String() |
| 40 | } |
| 41 | |
| 42 | // printfVerb is the length of a Go fmt verb at the start of rs, |
| 43 | // including %%. Zero means rs does not begin with a verb. |
| 44 | func printfVerb(rs []rune) int { |
| 45 | if len(rs) < 2 || rs[0] != '%' { |
| 46 | return 0 |
| 47 | } |
| 48 | if rs[1] == '%' { |
| 49 | return 2 |
| 50 | } |
| 51 | i := 1 |
| 52 | for i < len(rs) && strings.ContainsRune("#0-+ ", rs[i]) { |
| 53 | i++ |
| 54 | } |
| 55 | if i < len(rs) && rs[i] == '[' { |
| 56 | j := i + 1 |
| 57 | for j < len(rs) && rs[j] >= '0' && rs[j] <= '9' { |
| 58 | j++ |
| 59 | } |
| 60 | if j > i+1 && j < len(rs) && rs[j] == ']' { |
| 61 | i = j + 1 |
| 62 | } |
| 63 | } |
| 64 | for i < len(rs) && rs[i] >= '0' && rs[i] <= '9' { |
| 65 | i++ |
| 66 | } |
| 67 | if i < len(rs) && rs[i] == '.' { |
| 68 | k := i + 1 |
| 69 | for k < len(rs) && rs[k] >= '0' && rs[k] <= '9' { |
| 70 | k++ |
| 71 | } |
| 72 | if k > i+1 { |
| 73 | i = k |
| 74 | } |
| 75 | } |
| 76 | if i < len(rs) && isPrintfVerb(rs[i]) { |
| 77 | return i + 1 |
| 78 | } |
| 79 | return 0 |
| 80 | } |
| 81 | |
| 82 | func isPrintfVerb(r rune) bool { |
| 83 | switch r { |
| 84 | case 'b', 'c', 'd', 'e', 'E', 'f', 'g', 'G', 'o', 'p', 'q', 's', 't', 'T', 'U', 'v', 'w', 'x', 'X': |
| 85 | return true |
| 86 | default: |
| 87 | return false |
| 88 | } |
| 89 | } |
| 90 | |
| 91 | func isLetter(r rune) bool { |
| 92 | return (r >= 'a' && r <= 'z') || (r >= 'A' && r <= 'Z') |
| 93 | } |
| 94 | |
| 95 | func encodeWord(raw string) string { |
| 96 | ph := phonemes(raw) |
| 97 | if len(ph) == 0 { |
| 98 | return "" |
| 99 | } |
| 100 | var out []rune |
| 101 | i := 0 |
| 102 | for i < len(ph) { |
| 103 | if i+2 < len(ph) && isC(ph[i]) && isV(ph[i+1]) && isC(ph[i+2]) && (i+3 == len(ph) || isC(ph[i+3])) { |
| 104 | if sign, ok := signs[ph[i]+ph[i+1]+ph[i+2]]; ok { |
| 105 | out = append(out, sign) |
| 106 | i += 3 |
| 107 | continue |
| 108 | } |
| 109 | } |
| 110 | if i+1 < len(ph) && isC(ph[i]) && isV(ph[i+1]) { |
| 111 | if sign, ok := cv(ph[i], ph[i+1]); ok { |
| 112 | out = append(out, sign) |
| 113 | i += 2 |
| 114 | continue |
| 115 | } |
| 116 | } |
| 117 | if i+1 < len(ph) && isV(ph[i]) && isC(ph[i+1]) && (i+2 == len(ph) || isC(ph[i+2])) { |
| 118 | if sign, ok := signs[ph[i]+ph[i+1]]; ok { |
| 119 | out = append(out, sign) |
| 120 | i += 2 |
| 121 | continue |
| 122 | } |
| 123 | } |
| 124 | if isV(ph[i]) { |
| 125 | if sign, ok := signs[ph[i]]; ok { |
| 126 | out = append(out, sign) |
| 127 | } |
| 128 | i++ |
| 129 | continue |
| 130 | } |
| 131 | if sign, ok := cv(ph[i], "a"); ok { |
| 132 | out = append(out, sign) |
| 133 | } |
| 134 | i++ |
| 135 | } |
| 136 | return string(out) |
| 137 | } |
| 138 | |
| 139 | func cv(c, v string) (rune, bool) { |
| 140 | if v == "e" { |
| 141 | if sign, ok := signs[c+"e"]; ok { |
| 142 | return sign, true |
| 143 | } |
| 144 | if sign, ok := signs[c+"i"]; ok { |
| 145 | return sign, true |
| 146 | } |
| 147 | } |
| 148 | sign, ok := signs[c+v] |
| 149 | return sign, ok |
| 150 | } |
| 151 | |
| 152 | func isC(s string) bool { return s != "a" && s != "e" && s != "i" && s != "u" } |
| 153 | func isV(s string) bool { return !isC(s) } |
| 154 | |
| 155 | func phonemes(raw string) []string { |
| 156 | s := strings.ToLower(raw) |
| 157 | var out []string |
| 158 | for i := 0; i < len(s); i++ { |
| 159 | if i+3 < len(s) && s[i:i+4] == "eigh" { |
| 160 | out = append(out, "e") |
| 161 | i += 3 |
| 162 | continue |
| 163 | } |
| 164 | if i+2 < len(s) && s[i:i+3] == "igh" { |
| 165 | out = append(out, "e") |
| 166 | i += 2 |
| 167 | continue |
| 168 | } |
| 169 | if i+1 < len(s) { |
| 170 | switch s[i : i+2] { |
| 171 | case "qu": |
| 172 | out = append(out, "k", "w") |
| 173 | i++ |
| 174 | continue |
| 175 | case "ck": |
| 176 | out = append(out, "k") |
| 177 | i++ |
| 178 | continue |
| 179 | case "ph": |
| 180 | out = append(out, "p") |
| 181 | i++ |
| 182 | continue |
| 183 | case "th": |
| 184 | out = append(out, "t") |
| 185 | i++ |
| 186 | continue |
| 187 | case "sh": |
| 188 | out = append(out, "š") |
| 189 | i++ |
| 190 | continue |
| 191 | case "ch": |
| 192 | out = append(out, "š") |
| 193 | i++ |
| 194 | continue |
| 195 | case "wh": |
| 196 | out = append(out, "w") |
| 197 | i++ |
| 198 | continue |
| 199 | case "ng": |
| 200 | out = append(out, "n") |
| 201 | i++ |
| 202 | continue |
| 203 | case "ee", "ea": |
| 204 | out = append(out, "i") |
| 205 | i++ |
| 206 | continue |
| 207 | case "oo", "ou": |
| 208 | out = append(out, "u") |
| 209 | i++ |
| 210 | continue |
| 211 | case "ai", "ay": |
| 212 | out = append(out, "e") |
| 213 | i++ |
| 214 | continue |
| 215 | } |
| 216 | } |
| 217 | c := s[i] |
| 218 | switch c { |
| 219 | case 'a', 'e', 'i', 'u': |
| 220 | out = append(out, string(c)) |
| 221 | case 'o': |
| 222 | out = append(out, "u") |
| 223 | case 'y': |
| 224 | if i == 0 || isVowelByte(s[i-1]) { |
| 225 | out = append(out, "y") |
| 226 | } else { |
| 227 | out = append(out, "i") |
| 228 | } |
| 229 | case 'c': |
| 230 | if i+1 < len(s) && (s[i+1] == 'e' || s[i+1] == 'i' || s[i+1] == 'y') { |
| 231 | out = append(out, "s") |
| 232 | } else { |
| 233 | out = append(out, "k") |
| 234 | } |
| 235 | case 'f': |
| 236 | out = append(out, "p") |
| 237 | case 'v': |
| 238 | out = append(out, "b") |
| 239 | case 'j': |
| 240 | out = append(out, "y") |
| 241 | case 'q': |
| 242 | out = append(out, "k") |
| 243 | case 'x': |
| 244 | out = append(out, "k", "s") |
| 245 | case 'b', 'd', 'g', 'h', 'k', 'l', 'm', 'n', 'p', 'r', 's', 't', 'w', 'z': |
| 246 | out = append(out, string(c)) |
| 247 | } |
| 248 | } |
| 249 | compact := out[:0] |
| 250 | for _, p := range out { |
| 251 | if len(compact) > 0 && compact[len(compact)-1] == p && isC(p) { |
| 252 | continue |
| 253 | } |
| 254 | compact = append(compact, p) |
| 255 | } |
| 256 | return compact |
| 257 | } |
| 258 | |
| 259 | func isVowelByte(c byte) bool { |
| 260 | return c == 'a' || c == 'e' || c == 'i' || c == 'o' || c == 'u' || c == 'y' |
| 261 | } |
| 262 | |
| 263 | // signs maps a conventional Akkadian reading to one sign. |
| 264 | // pu is written with BU, and w+vowel with PI, which are the usual values. |
| 265 | var signs = map[string]rune{ |
| 266 | "a": '𒀀', "e": '𒂊', "i": '𒄿', "u": '𒌋', |
| 267 | "ba": '𒁀', "bi": '𒁉', "bu": '𒁍', "pu": '𒁍', |
| 268 | "da": '𒁕', "di": '𒁲', "du": '𒁺', |
| 269 | "ga": '𒂵', "gi": '𒄀', "gu": '𒄖', |
| 270 | "ha": '𒄩', "hi": '𒄭', "hu": '𒄷', |
| 271 | "ka": '𒅗', "ki": '𒆠', "ku": '𒆪', |
| 272 | "la": '𒆷', "li": '𒇷', "lu": '𒇻', |
| 273 | "ma": '𒈠', "me": '𒈨', "mi": '𒈪', "mu": '𒈬', |
| 274 | "na": '𒈾', "ne": '𒉈', "ni": '𒉌', "nu": '𒉡', |
| 275 | "pa": '𒉺', "pi": '𒉿', |
| 276 | "ra": '𒊏', "ri": '𒊑', "ru": '𒊒', |
| 277 | "sa": '𒊓', "si": '𒋛', "su": '𒋢', |
| 278 | "ša": '𒊭', "še": '𒊺', "ši": '𒅆', "šu": '𒋗', |
| 279 | "ta": '𒋫', "te": '𒋼', "ti": '𒋾', "tu": '𒌅', |
| 280 | "wa": '𒉿', "we": '𒉿', "wi": '𒉿', "wu": '𒉿', |
| 281 | "ya": '𒅀', "za": '𒍝', "zi": '𒍣', "zu": '𒍪', |
| 282 | "ab": '𒀊', "ad": '𒀜', "ak": '𒀝', "al": '𒀠', "an": '𒀭', "aš": '𒀸', |
| 283 | "el": '𒂖', "en": '𒂗', "eš": '𒂠', |
| 284 | "ib": '𒅁', "ig": '𒅅', "il": '𒅋', "im": '𒅎', "in": '𒅔', "ir": '𒅕', "iš": '𒅖', |
| 285 | "ub": '𒌒', "ud": '𒌓', "um": '𒌝', "un": '𒌦', "ur": '𒌨', "uš": '𒍑', |
| 286 | "bad": '𒁁', "bal": '𒁄', "bar": '𒁇', |
| 287 | "dag": '𒁖', "dam": '𒁮', "dar": '𒁯', |
| 288 | "dib": '𒁳', "dim": '𒁴', "din": '𒁷', |
| 289 | "dub": '𒁾', "dug": '𒂁', "dun": '𒂄', |
| 290 | "gal": '𒃲', "gam": '𒃵', "gan": '𒃶', "gar": '𒃻', |
| 291 | "gul": '𒄢', "gur": '𒄥', |
| 292 | "hal": '𒄬', |
| 293 | "kal": '𒆗', "kin": '𒆥', "kul": '𒆰', "kur": '𒆳', |
| 294 | "lal": '𒇲', "lam": '𒇴', "lum": '𒈝', |
| 295 | "mar": '𒈥', |
| 296 | "nam": '𒉆', "nun": '𒉣', |
| 297 | "pad": '𒉻', "pap": '𒉽', |
| 298 | "rab": '𒊐', |
| 299 | "sag": '𒊕', "sar": '𒊬', |
| 300 | "tab": '𒋰', "tag": '𒋳', "tar": '𒋻', |
| 301 | "tum": '𒌈', "tur": '𒌉', |
| 302 | "zag": '𒍠', |
| 303 | } |