1######################## BEGIN LICENSE BLOCK ######################## 2# The Original Code is Mozilla Communicator client code. 3# 4# The Initial Developer of the Original Code is 5# Netscape Communications Corporation. 6# Portions created by the Initial Developer are Copyright (C) 1998 7# the Initial Developer. All Rights Reserved. 8# 9# Contributor(s): 10# Mark Pilgrim - port to Python 11# 12# This library is free software; you can redistribute it and/or 13# modify it under the terms of the GNU Lesser General Public 14# License as published by the Free Software Foundation; either 15# version 2.1 of the License, or (at your option) any later version. 16# 17# This library is distributed in the hope that it will be useful, 18# but WITHOUT ANY WARRANTY; without even the implied warranty of 19# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 20# Lesser General Public License for more details. 21# 22# You should have received a copy of the GNU Lesser General Public 23# License along with this library; if not, write to the Free Software 24# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 25# 02110-1301 USA 26######################### END LICENSE BLOCK ######################### 27 28# Sampling from about 20M text materials include literature and computer technology 29 30# 128 --> 0.79 31# 256 --> 0.92 32# 512 --> 0.986 33# 1024 --> 0.99944 34# 2048 --> 0.99999 35# 36# Idea Distribution Ratio = 0.98653 / (1-0.98653) = 73.24 37# Random Distribution Ration = 512 / (2350-512) = 0.279. 38# 39# Typical Distribution Ratio 40 41EUCKR_TYPICAL_DISTRIBUTION_RATIO = 6.0 42 43EUCKR_TABLE_SIZE = 2352 44 45# Char to FreqOrder table , 46EUCKR_CHAR_TO_FREQ_ORDER = ( 47 13, 130, 120,1396, 481,1719,1720, 328, 609, 212,1721, 707, 400, 299,1722, 87, 481397,1723, 104, 536,1117,1203,1724,1267, 685,1268, 508,1725,1726,1727,1728,1398, 491399,1729,1730,1731, 141, 621, 326,1057, 368,1732, 267, 488, 20,1733,1269,1734, 50 945,1400,1735, 47, 904,1270,1736,1737, 773, 248,1738, 409, 313, 786, 429,1739, 51 116, 987, 813,1401, 683, 75,1204, 145,1740,1741,1742,1743, 16, 847, 667, 622, 52 708,1744,1745,1746, 966, 787, 304, 129,1747, 60, 820, 123, 676,1748,1749,1750, 531751, 617,1752, 626,1753,1754,1755,1756, 653,1757,1758,1759,1760,1761,1762, 856, 54 344,1763,1764,1765,1766, 89, 401, 418, 806, 905, 848,1767,1768,1769, 946,1205, 55 709,1770,1118,1771, 241,1772,1773,1774,1271,1775, 569,1776, 999,1777,1778,1779, 561780, 337, 751,1058, 28, 628, 254,1781, 177, 906, 270, 349, 891,1079,1782, 19, 571783, 379,1784, 315,1785, 629, 754,1402, 559,1786, 636, 203,1206,1787, 710, 567, 581788, 935, 814,1789,1790,1207, 766, 528,1791,1792,1208,1793,1794,1795,1796,1797, 591403,1798,1799, 533,1059,1404,1405,1156,1406, 936, 884,1080,1800, 351,1801,1802, 601803,1804,1805, 801,1806,1807,1808,1119,1809,1157, 714, 474,1407,1810, 298, 899, 61 885,1811,1120, 802,1158,1812, 892,1813,1814,1408, 659,1815,1816,1121,1817,1818, 621819,1820,1821,1822, 319,1823, 594, 545,1824, 815, 937,1209,1825,1826, 573,1409, 631022,1827,1210,1828,1829,1830,1831,1832,1833, 556, 722, 807,1122,1060,1834, 697, 641835, 900, 557, 715,1836,1410, 540,1411, 752,1159, 294, 597,1211, 976, 803, 770, 651412,1837,1838, 39, 794,1413, 358,1839, 371, 925,1840, 453, 661, 788, 531, 723, 66 544,1023,1081, 869, 91,1841, 392, 430, 790, 602,1414, 677,1082, 457,1415,1416, 671842,1843, 475, 327,1024,1417, 795, 121,1844, 733, 403,1418,1845,1846,1847, 300, 68 119, 711,1212, 627,1848,1272, 207,1849,1850, 796,1213, 382,1851, 519,1852,1083, 69 893,1853,1854,1855, 367, 809, 487, 671,1856, 663,1857,1858, 956, 471, 306, 857, 701859,1860,1160,1084,1861,1862,1863,1864,1865,1061,1866,1867,1868,1869,1870,1871, 71 282, 96, 574,1872, 502,1085,1873,1214,1874, 907,1875,1876, 827, 977,1419,1420, 721421, 268,1877,1422,1878,1879,1880, 308,1881, 2, 537,1882,1883,1215,1884,1885, 73 127, 791,1886,1273,1423,1887, 34, 336, 404, 643,1888, 571, 654, 894, 840,1889, 74 0, 886,1274, 122, 575, 260, 908, 938,1890,1275, 410, 316,1891,1892, 100,1893, 751894,1123, 48,1161,1124,1025,1895, 633, 901,1276,1896,1897, 115, 816,1898, 317, 761899, 694,1900, 909, 734,1424, 572, 866,1425, 691, 85, 524,1010, 543, 394, 841, 771901,1902,1903,1026,1904,1905,1906,1907,1908,1909, 30, 451, 651, 988, 310,1910, 781911,1426, 810,1216, 93,1912,1913,1277,1217,1914, 858, 759, 45, 58, 181, 610, 79 269,1915,1916, 131,1062, 551, 443,1000, 821,1427, 957, 895,1086,1917,1918, 375, 801919, 359,1920, 687,1921, 822,1922, 293,1923,1924, 40, 662, 118, 692, 29, 939, 81 887, 640, 482, 174,1925, 69,1162, 728,1428, 910,1926,1278,1218,1279, 386, 870, 82 217, 854,1163, 823,1927,1928,1929,1930, 834,1931, 78,1932, 859,1933,1063,1934, 831935,1936,1937, 438,1164, 208, 595,1938,1939,1940,1941,1219,1125,1942, 280, 888, 841429,1430,1220,1431,1943,1944,1945,1946,1947,1280, 150, 510,1432,1948,1949,1950, 851951,1952,1953,1954,1011,1087,1955,1433,1043,1956, 881,1957, 614, 958,1064,1065, 861221,1958, 638,1001, 860, 967, 896,1434, 989, 492, 553,1281,1165,1959,1282,1002, 871283,1222,1960,1961,1962,1963, 36, 383, 228, 753, 247, 454,1964, 876, 678,1965, 881966,1284, 126, 464, 490, 835, 136, 672, 529, 940,1088,1435, 473,1967,1968, 467, 89 50, 390, 227, 587, 279, 378, 598, 792, 968, 240, 151, 160, 849, 882,1126,1285, 90 639,1044, 133, 140, 288, 360, 811, 563,1027, 561, 142, 523,1969,1970,1971, 7, 91 103, 296, 439, 407, 506, 634, 990,1972,1973,1974,1975, 645,1976,1977,1978,1979, 921980,1981, 236,1982,1436,1983,1984,1089, 192, 828, 618, 518,1166, 333,1127,1985, 93 818,1223,1986,1987,1988,1989,1990,1991,1992,1993, 342,1128,1286, 746, 842,1994, 941995, 560, 223,1287, 98, 8, 189, 650, 978,1288,1996,1437,1997, 17, 345, 250, 95 423, 277, 234, 512, 226, 97, 289, 42, 167,1998, 201,1999,2000, 843, 836, 824, 96 532, 338, 783,1090, 182, 576, 436,1438,1439, 527, 500,2001, 947, 889,2002,2003, 972004,2005, 262, 600, 314, 447,2006, 547,2007, 693, 738,1129,2008, 71,1440, 745, 98 619, 688,2009, 829,2010,2011, 147,2012, 33, 948,2013,2014, 74, 224,2015, 61, 99 191, 918, 399, 637,2016,1028,1130, 257, 902,2017,2018,2019,2020,2021,2022,2023, 1002024,2025,2026, 837,2027,2028,2029,2030, 179, 874, 591, 52, 724, 246,2031,2032, 1012033,2034,1167, 969,2035,1289, 630, 605, 911,1091,1168,2036,2037,2038,1441, 912, 1022039, 623,2040,2041, 253,1169,1290,2042,1442, 146, 620, 611, 577, 433,2043,1224, 103 719,1170, 959, 440, 437, 534, 84, 388, 480,1131, 159, 220, 198, 679,2044,1012, 104 819,1066,1443, 113,1225, 194, 318,1003,1029,2045,2046,2047,2048,1067,2049,2050, 1052051,2052,2053, 59, 913, 112,2054, 632,2055, 455, 144, 739,1291,2056, 273, 681, 106 499,2057, 448,2058,2059, 760,2060,2061, 970, 384, 169, 245,1132,2062,2063, 414, 1071444,2064,2065, 41, 235,2066, 157, 252, 877, 568, 919, 789, 580,2067, 725,2068, 1082069,1292,2070,2071,1445,2072,1446,2073,2074, 55, 588, 66,1447, 271,1092,2075, 1091226,2076, 960,1013, 372,2077,2078,2079,2080,2081,1293,2082,2083,2084,2085, 850, 1102086,2087,2088,2089,2090, 186,2091,1068, 180,2092,2093,2094, 109,1227, 522, 606, 1112095, 867,1448,1093, 991,1171, 926, 353,1133,2096, 581,2097,2098,2099,1294,1449, 1121450,2100, 596,1172,1014,1228,2101,1451,1295,1173,1229,2102,2103,1296,1134,1452, 113 949,1135,2104,2105,1094,1453,1454,1455,2106,1095,2107,2108,2109,2110,2111,2112, 1142113,2114,2115,2116,2117, 804,2118,2119,1230,1231, 805,1456, 405,1136,2120,2121, 1152122,2123,2124, 720, 701,1297, 992,1457, 927,1004,2125,2126,2127,2128,2129,2130, 116 22, 417,2131, 303,2132, 385,2133, 971, 520, 513,2134,1174, 73,1096, 231, 274, 117 962,1458, 673,2135,1459,2136, 152,1137,2137,2138,2139,2140,1005,1138,1460,1139, 1182141,2142,2143,2144, 11, 374, 844,2145, 154,1232, 46,1461,2146, 838, 830, 721, 1191233, 106,2147, 90, 428, 462, 578, 566,1175, 352,2148,2149, 538,1234, 124,1298, 1202150,1462, 761, 565,2151, 686,2152, 649,2153, 72, 173,2154, 460, 415,2155,1463, 1212156,1235, 305,2157,2158,2159,2160,2161,2162, 579,2163,2164,2165,2166,2167, 747, 1222168,2169,2170,2171,1464, 669,2172,2173,2174,2175,2176,1465,2177, 23, 530, 285, 1232178, 335, 729,2179, 397,2180,2181,2182,1030,2183,2184, 698,2185,2186, 325,2187, 1242188, 369,2189, 799,1097,1015, 348,2190,1069, 680,2191, 851,1466,2192,2193, 10, 1252194, 613, 424,2195, 979, 108, 449, 589, 27, 172, 81,1031, 80, 774, 281, 350, 1261032, 525, 301, 582,1176,2196, 674,1045,2197,2198,1467, 730, 762,2199,2200,2201, 1272202,1468,2203, 993,2204,2205, 266,1070, 963,1140,2206,2207,2208, 664,1098, 972, 1282209,2210,2211,1177,1469,1470, 871,2212,2213,2214,2215,2216,1471,2217,2218,2219, 1292220,2221,2222,2223,2224,2225,2226,2227,1472,1236,2228,2229,2230,2231,2232,2233, 1302234,2235,1299,2236,2237, 200,2238, 477, 373,2239,2240, 731, 825, 777,2241,2242, 1312243, 521, 486, 548,2244,2245,2246,1473,1300, 53, 549, 137, 875, 76, 158,2247, 1321301,1474, 469, 396,1016, 278, 712,2248, 321, 442, 503, 767, 744, 941,1237,1178, 1331475,2249, 82, 178,1141,1179, 973,2250,1302,2251, 297,2252,2253, 570,2254,2255, 1342256, 18, 450, 206,2257, 290, 292,1142,2258, 511, 162, 99, 346, 164, 735,2259, 1351476,1477, 4, 554, 343, 798,1099,2260,1100,2261, 43, 171,1303, 139, 215,2262, 1362263, 717, 775,2264,1033, 322, 216,2265, 831,2266, 149,2267,1304,2268,2269, 702, 1371238, 135, 845, 347, 309,2270, 484,2271, 878, 655, 238,1006,1478,2272, 67,2273, 138 295,2274,2275, 461,2276, 478, 942, 412,2277,1034,2278,2279,2280, 265,2281, 541, 1392282,2283,2284,2285,2286, 70, 852,1071,2287,2288,2289,2290, 21, 56, 509, 117, 140 432,2291,2292, 331, 980, 552,1101, 148, 284, 105, 393,1180,1239, 755,2293, 187, 1412294,1046,1479,2295, 340,2296, 63,1047, 230,2297,2298,1305, 763,1306, 101, 800, 142 808, 494,2299,2300,2301, 903,2302, 37,1072, 14, 5,2303, 79, 675,2304, 312, 1432305,2306,2307,2308,2309,1480, 6,1307,2310,2311,2312, 1, 470, 35, 24, 229, 1442313, 695, 210, 86, 778, 15, 784, 592, 779, 32, 77, 855, 964,2314, 259,2315, 145 501, 380,2316,2317, 83, 981, 153, 689,1308,1481,1482,1483,2318,2319, 716,1484, 1462320,2321,2322,2323,2324,2325,1485,2326,2327, 128, 57, 68, 261,1048, 211, 170, 1471240, 31,2328, 51, 435, 742,2329,2330,2331, 635,2332, 264, 456,2333,2334,2335, 148 425,2336,1486, 143, 507, 263, 943,2337, 363, 920,1487, 256,1488,1102, 243, 601, 1491489,2338,2339,2340,2341,2342,2343,2344, 861,2345,2346,2347,2348,2349,2350, 395, 1502351,1490,1491, 62, 535, 166, 225,2352,2353, 668, 419,1241, 138, 604, 928,2354, 1511181,2355,1492,1493,2356,2357,2358,1143,2359, 696,2360, 387, 307,1309, 682, 476, 1522361,2362, 332, 12, 222, 156,2363, 232,2364, 641, 276, 656, 517,1494,1495,1035, 153 416, 736,1496,2365,1017, 586,2366,2367,2368,1497,2369, 242,2370,2371,2372,1498, 1542373, 965, 713,2374,2375,2376,2377, 740, 982,1499, 944,1500,1007,2378,2379,1310, 1551501,2380,2381,2382, 785, 329,2383,2384,1502,2385,2386,2387, 932,2388,1503,2389, 1562390,2391,2392,1242,2393,2394,2395,2396,2397, 994, 950,2398,2399,2400,2401,1504, 1571311,2402,2403,2404,2405,1049, 749,2406,2407, 853, 718,1144,1312,2408,1182,1505, 1582409,2410, 255, 516, 479, 564, 550, 214,1506,1507,1313, 413, 239, 444, 339,1145, 1591036,1508,1509,1314,1037,1510,1315,2411,1511,2412,2413,2414, 176, 703, 497, 624, 160 593, 921, 302,2415, 341, 165,1103,1512,2416,1513,2417,2418,2419, 376,2420, 700, 1612421,2422,2423, 258, 768,1316,2424,1183,2425, 995, 608,2426,2427,2428,2429, 221, 1622430,2431,2432,2433,2434,2435,2436,2437, 195, 323, 726, 188, 897, 983,1317, 377, 163 644,1050, 879,2438, 452,2439,2440,2441,2442,2443,2444, 914,2445,2446,2447,2448, 164 915, 489,2449,1514,1184,2450,2451, 515, 64, 427, 495,2452, 583,2453, 483, 485, 1651038, 562, 213,1515, 748, 666,2454,2455,2456,2457, 334,2458, 780, 996,1008, 705, 1661243,2459,2460,2461,2462,2463, 114,2464, 493,1146, 366, 163,1516, 961,1104,2465, 167 291,2466,1318,1105,2467,1517, 365,2468, 355, 951,1244,2469,1319,2470, 631,2471, 1682472, 218,1320, 364, 320, 756,1518,1519,1321,1520,1322,2473,2474,2475,2476, 997, 1692477,2478,2479,2480, 665,1185,2481, 916,1521,2482,2483,2484, 584, 684,2485,2486, 170 797,2487,1051,1186,2488,2489,2490,1522,2491,2492, 370,2493,1039,1187, 65,2494, 171 434, 205, 463,1188,2495, 125, 812, 391, 402, 826, 699, 286, 398, 155, 781, 771, 172 585,2496, 590, 505,1073,2497, 599, 244, 219, 917,1018, 952, 646,1523,2498,1323, 1732499,2500, 49, 984, 354, 741,2501, 625,2502,1324,2503,1019, 190, 357, 757, 491, 174 95, 782, 868,2504,2505,2506,2507,2508,2509, 134,1524,1074, 422,1525, 898,2510, 175 161,2511,2512,2513,2514, 769,2515,1526,2516,2517, 411,1325,2518, 472,1527,2519, 1762520,2521,2522,2523,2524, 985,2525,2526,2527,2528,2529,2530, 764,2531,1245,2532, 1772533, 25, 204, 311,2534, 496,2535,1052,2536,2537,2538,2539,2540,2541,2542, 199, 178 704, 504, 468, 758, 657,1528, 196, 44, 839,1246, 272, 750,2543, 765, 862,2544, 1792545,1326,2546, 132, 615, 933,2547, 732,2548,2549,2550,1189,1529,2551, 283,1247, 1801053, 607, 929,2552,2553,2554, 930, 183, 872, 616,1040,1147,2555,1148,1020, 441, 181 249,1075,2556,2557,2558, 466, 743,2559,2560,2561, 92, 514, 426, 420, 526,2562, 1822563,2564,2565,2566,2567,2568, 185,2569,2570,2571,2572, 776,1530, 658,2573, 362, 1832574, 361, 922,1076, 793,2575,2576,2577,2578,2579,2580,1531, 251,2581,2582,2583, 1842584,1532, 54, 612, 237,1327,2585,2586, 275, 408, 647, 111,2587,1533,1106, 465, 185 3, 458, 9, 38,2588, 107, 110, 890, 209, 26, 737, 498,2589,1534,2590, 431, 186 202, 88,1535, 356, 287,1107, 660,1149,2591, 381,1536, 986,1150, 445,1248,1151, 187 974,2592,2593, 846,2594, 446, 953, 184,1249,1250, 727,2595, 923, 193, 883,2596, 1882597,2598, 102, 324, 539, 817,2599, 421,1041,2600, 832,2601, 94, 175, 197, 406, 1892602, 459,2603,2604,2605,2606,2607, 330, 555,2608,2609,2610, 706,1108, 389,2611, 1902612,2613,2614, 233,2615, 833, 558, 931, 954,1251,2616,2617,1537, 546,2618,2619, 1911009,2620,2621,2622,1538, 690,1328,2623, 955,2624,1539,2625,2626, 772,2627,2628, 1922629,2630,2631, 924, 648, 863, 603,2632,2633, 934,1540, 864, 865,2634, 642,1042, 193 670,1190,2635,2636,2637,2638, 168,2639, 652, 873, 542,1054,1541,2640,2641,2642, # 512, 256 194) 195 196