stet-pdf-reader 0.8.1

// stet-pdf-reader
// Copyright (c) 2026 Scott Bowman
// SPDX-License-Identifier: Apache-2.0 OR MIT

//! Hardcoded GID-to-Unicode mapping tables for common non-embedded fonts.
//!
//! When a CIDFontType2 font with Identity-H encoding is not embedded in the PDF,
//! the CID values in text strings are glyph indices (GIDs) from the original font.
//! A substitute font has different glyph ordering, so CID-as-GID produces garbled
//! text. These tables map known fonts' GIDs to Unicode code points, enabling
//! correct rendering with any substitute font.
//!
//! Tables derived from Mozilla PDF.js (Apache-2.0 licensed):
//! - `getGlyphMapForStandardFonts()` — base table for common TrueType fonts
//! - `getSupplementalGlyphMapForCalibri()` — Calibri-specific overrides

use std::collections::HashMap;

/// Build a GID-to-Unicode map for a known font family.
///
/// Returns `Some(map)` if the font family has a known mapping, `None` otherwise.
/// The font family name should be lowercase, stripped of subset prefix and style
/// suffix (e.g. "calibri" not "ABCDEF+Calibri,Bold").
pub fn get_gid_to_unicode_map(font_family: &str) -> Option<HashMap<u16, u32>> {
    match font_family {
        "calibri" => {
            let mut map = standard_glyph_map();
            // Calibri overrides: these GIDs differ from the generic table
            for (gid, unicode) in CALIBRI_SUPPLEMENT {
                map.insert(*gid, *unicode);
            }
            Some(map)
        }
        // These fonts share the standard TrueType glyph ordering
        "tahoma" | "verdana" | "arial" | "trebuchet ms" | "trebuchetms" | "centurygothic"
        | "century gothic" | "comicsansms" | "comic sans ms" | "consolas" | "georgia"
        | "impact" | "lucidaconsole" | "lucida console" | "palatinolinotype"
        | "palatino linotype" | "segoeui" | "segoe ui" | "timesnewroman" | "times new roman"
        | "couriernew" | "courier new" | "cambria" | "candara" | "constantia" | "corbel" => {
            Some(standard_glyph_map())
        }
        // ArialBlack uses the standard map plus overrides for Polish chars
        "arialblack" | "arial black" => {
            let mut map = standard_glyph_map();
            for (gid, unicode) in ARIAL_BLACK_SUPPLEMENT {
                map.insert(*gid, *unicode);
            }
            Some(map)
        }
        // Fallback: for any non-embedded CIDFontType2 + Identity-H font,
        // the standard TrueType glyph ordering is the best guess. Without
        // any mapping the text is guaranteed garbled; the standard map is
        // correct for the vast majority of Western TrueType fonts.
        _ => Some(standard_glyph_map()),
    }
}

/// Base GID-to-Unicode table for standard TrueType fonts.
///
/// Covers basic Latin, extended Latin, Cyrillic, Hebrew, Arabic diacriticals,
/// and assorted symbols. 496 entries.
fn standard_glyph_map() -> HashMap<u16, u32> {
    let mut m = HashMap::with_capacity(STANDARD_GLYPHS.len());
    for (gid, unicode) in STANDARD_GLYPHS {
        m.insert(*gid, *unicode);
    }
    m
}

/// Standard fonts GID→Unicode table (from PDF.js getGlyphMapForStandardFonts).
static STANDARD_GLYPHS: &[(u16, u32)] = &[
    (2, 10),
    (3, 32),
    (4, 33),
    (5, 34),
    (6, 35),
    (7, 36),
    (8, 37),
    (9, 38),
    (10, 39),
    (11, 40),
    (12, 41),
    (13, 42),
    (14, 43),
    (15, 44),
    (16, 45),
    (17, 46),
    (18, 47),
    (19, 48),
    (20, 49),
    (21, 50),
    (22, 51),
    (23, 52),
    (24, 53),
    (25, 54),
    (26, 55),
    (27, 56),
    (28, 57),
    (29, 58),
    (30, 894),
    (31, 60),
    (32, 61),
    (33, 62),
    (34, 63),
    (35, 64),
    (36, 65),
    (37, 66),
    (38, 67),
    (39, 68),
    (40, 69),
    (41, 70),
    (42, 71),
    (43, 72),
    (44, 73),
    (45, 74),
    (46, 75),
    (47, 76),
    (48, 77),
    (49, 78),
    (50, 79),
    (51, 80),
    (52, 81),
    (53, 82),
    (54, 83),
    (55, 84),
    (56, 85),
    (57, 86),
    (58, 87),
    (59, 88),
    (60, 89),
    (61, 90),
    (62, 91),
    (63, 92),
    (64, 93),
    (65, 94),
    (66, 95),
    (67, 96),
    (68, 97),
    (69, 98),
    (70, 99),
    (71, 100),
    (72, 101),
    (73, 102),
    (74, 103),
    (75, 104),
    (76, 105),
    (77, 106),
    (78, 107),
    (79, 108),
    (80, 109),
    (81, 110),
    (82, 111),
    (83, 112),
    (84, 113),
    (85, 114),
    (86, 115),
    (87, 116),
    (88, 117),
    (89, 118),
    (90, 119),
    (91, 120),
    (92, 121),
    (93, 122),
    (94, 123),
    (95, 124),
    (96, 125),
    (97, 126),
    (98, 196),
    (99, 197),
    (100, 199),
    (101, 201),
    (102, 209),
    (103, 214),
    (104, 220),
    (105, 225),
    (106, 224),
    (107, 226),
    (108, 228),
    (109, 227),
    (110, 229),
    (111, 231),
    (112, 233),
    (113, 232),
    (114, 234),
    (115, 235),
    (116, 237),
    (117, 236),
    (118, 238),
    (119, 239),
    (120, 241),
    (121, 243),
    (122, 242),
    (123, 244),
    (124, 246),
    (125, 245),
    (126, 250),
    (127, 249),
    (128, 251),
    (129, 252),
    (130, 8224),
    (131, 176),
    (132, 162),
    (133, 163),
    (134, 167),
    (135, 8226),
    (136, 182),
    (137, 223),
    (138, 174),
    (139, 169),
    (140, 8482),
    (141, 180),
    (142, 168),
    (143, 8800),
    (144, 198),
    (145, 216),
    (146, 8734),
    (147, 177),
    (148, 8804),
    (149, 8805),
    (150, 165),
    (151, 181),
    (152, 8706),
    (153, 8721),
    (154, 8719),
    (156, 8747),
    (157, 170),
    (158, 186),
    (159, 8486),
    (160, 230),
    (161, 248),
    (162, 191),
    (163, 161),
    (164, 172),
    (165, 8730),
    (166, 402),
    (167, 8776),
    (168, 8710),
    (169, 171),
    (170, 187),
    (171, 8230),
    (179, 8220),
    (180, 8221),
    (181, 8216),
    (182, 8217),
    (200, 193),
    (203, 205),
    (207, 211),
    (210, 218),
    (223, 711),
    (224, 321),
    (225, 322),
    (226, 352),
    (227, 353),
    (228, 381),
    (229, 382),
    (233, 221),
    (234, 253),
    (252, 263),
    (253, 268),
    (254, 269),
    (258, 258),
    (260, 260),
    (261, 261),
    (265, 280),
    (266, 281),
    (267, 282),
    (268, 283),
    (269, 313),
    (275, 323),
    (276, 324),
    (277, 327),
    (278, 328),
    (279, 336),
    (280, 337),
    (281, 338),
    (282, 339),
    (283, 344),
    (284, 345),
    (285, 346),
    (286, 347),
    (287, 350),
    (288, 351),
    (289, 354),
    (290, 355),
    (291, 366),
    (292, 367),
    (293, 368),
    (294, 369),
    (295, 377),
    (296, 378),
    (297, 379),
    (298, 380),
    (305, 963),
    (306, 964),
    (307, 966),
    (308, 8215),
    (309, 8252),
    (310, 8319),
    (311, 8359),
    (312, 8592),
    (313, 8593),
    (337, 9552),
    (493, 1039),
    (494, 1040),
    (570, 1040),
    (571, 1041),
    (572, 1042),
    (573, 1043),
    (574, 1044),
    (575, 1045),
    (576, 1046),
    (577, 1047),
    (578, 1048),
    (579, 1049),
    (580, 1050),
    (581, 1051),
    (582, 1052),
    (583, 1053),
    (584, 1054),
    (585, 1055),
    (586, 1056),
    (587, 1057),
    (588, 1058),
    (589, 1059),
    (590, 1060),
    (591, 1061),
    (592, 1062),
    (593, 1063),
    (594, 1064),
    (595, 1065),
    (596, 1066),
    (597, 1067),
    (598, 1068),
    (599, 1069),
    (600, 1070),
    (601, 1071),
    (602, 1072),
    (603, 1073),
    (604, 1074),
    (605, 1075),
    (606, 1076),
    (607, 1077),
    (608, 1078),
    (609, 1079),
    (610, 1080),
    (611, 1081),
    (612, 1082),
    (613, 1083),
    (614, 1084),
    (615, 1085),
    (616, 1086),
    (617, 1087),
    (618, 1088),
    (619, 1089),
    (620, 1090),
    (621, 1091),
    (622, 1092),
    (623, 1093),
    (624, 1094),
    (625, 1095),
    (626, 1096),
    (627, 1097),
    (628, 1098),
    (629, 1099),
    (630, 1100),
    (631, 1101),
    (632, 1102),
    (633, 1103),
    (672, 1488),
    (673, 1489),
    (674, 1490),
    (675, 1491),
    (676, 1492),
    (677, 1493),
    (678, 1494),
    (679, 1495),
    (680, 1496),
    (681, 1497),
    (682, 1498),
    (683, 1499),
    (684, 1500),
    (685, 1501),
    (686, 1502),
    (687, 1503),
    (688, 1504),
    (689, 1505),
    (690, 1506),
    (691, 1507),
    (692, 1508),
    (693, 1509),
    (694, 1510),
    (695, 1511),
    (696, 1512),
    (697, 1513),
    (698, 1514),
    (705, 1524),
    (706, 8362),
    (710, 64288),
    (711, 64298),
    (759, 1617),
    (761, 1776),
    (763, 1778),
    (775, 1652),
    (777, 1764),
    (778, 1780),
    (779, 1781),
    (780, 1782),
    (782, 771),
    (783, 64726),
    (786, 8363),
    (788, 8532),
    (790, 768),
    (791, 769),
    (792, 768),
    (795, 803),
    (797, 64336),
    (798, 64337),
    (799, 64342),
    (800, 64343),
    (801, 64344),
    (802, 64345),
    (803, 64362),
    (804, 64363),
    (805, 64364),
    (2424, 7821),
    (2425, 7822),
    (2426, 7823),
    (2427, 7824),
    (2428, 7825),
    (2429, 7826),
    (2430, 7827),
    (2433, 7682),
    (2678, 8045),
    (2679, 8046),
    (2830, 1552),
    (2838, 686),
    (2840, 751),
    (2842, 753),
    (2843, 754),
    (2844, 755),
    (2846, 757),
    (2856, 767),
    (2857, 848),
    (2858, 849),
    (2862, 853),
    (2863, 854),
    (2864, 855),
    (2865, 861),
    (2866, 862),
    (2906, 7460),
    (2908, 7462),
    (2909, 7463),
    (2910, 7464),
    (2912, 7466),
    (2913, 7467),
    (2914, 7468),
    (2916, 7470),
    (2917, 7471),
    (2918, 7472),
    (2920, 7474),
    (2921, 7475),
    (2922, 7476),
    (2924, 7478),
    (2925, 7479),
    (2926, 7480),
    (2928, 7482),
    (2929, 7483),
    (2930, 7484),
    (2932, 7486),
    (2933, 7487),
    (2934, 7488),
    (2936, 7490),
    (2937, 7491),
    (2938, 7492),
    (2940, 7494),
    (2941, 7495),
    (2942, 7496),
    (2944, 7498),
    (2946, 7500),
    (2948, 7502),
    (2950, 7504),
    (2951, 7505),
    (2952, 7506),
    (2954, 7508),
    (2955, 7509),
    (2956, 7510),
    (2958, 7512),
    (2959, 7513),
    (2960, 7514),
    (2962, 7516),
    (2963, 7517),
    (2964, 7518),
    (2966, 7520),
    (2967, 7521),
    (2968, 7522),
    (2970, 7524),
    (2971, 7525),
    (2972, 7526),
    (2974, 7528),
    (2975, 7529),
    (2976, 7530),
    (2978, 1537),
    (2979, 1538),
    (2980, 1539),
    (2982, 1549),
    (2983, 1551),
    (2984, 1552),
    (2986, 1554),
    (2987, 1555),
    (2988, 1556),
    (2990, 1623),
    (2991, 1624),
    (2995, 1775),
    (2999, 1791),
    (3002, 64290),
    (3003, 64291),
    (3004, 64292),
    (3006, 64294),
    (3007, 64295),
    (3008, 64296),
    (3011, 1900),
    (3014, 8223),
    (3015, 8244),
    (3017, 7532),
    (3018, 7533),
    (3019, 7534),
    (3075, 7590),
    (3076, 7591),
    (3079, 7594),
    (3080, 7595),
    (3083, 7598),
    (3084, 7599),
    (3087, 7602),
    (3088, 7603),
    (3091, 7606),
    (3092, 7607),
    (3095, 7610),
    (3096, 7611),
    (3099, 7614),
    (3100, 7615),
    (3103, 7618),
    (3104, 7619),
    (3107, 8337),
    (3108, 8338),
    (3116, 1884),
    (3119, 1885),
    (3120, 1885),
    (3123, 1886),
    (3124, 1886),
    (3127, 1887),
    (3128, 1887),
    (3131, 1888),
    (3132, 1888),
    (3135, 1889),
    (3136, 1889),
    (3139, 1890),
    (3140, 1890),
    (3143, 1891),
    (3144, 1891),
    (3147, 1892),
    (3148, 1892),
    (3153, 580),
    (3154, 581),
    (3157, 584),
    (3158, 585),
    (3161, 588),
    (3162, 589),
    (3165, 891),
    (3166, 892),
    (3169, 1274),
    (3170, 1275),
    (3173, 1278),
    (3174, 1279),
    (3181, 7622),
    (3182, 7623),
    (3282, 11799),
    (3316, 578),
    (3379, 42785),
    (3393, 1159),
    (3416, 8377),
];

/// Calibri-specific GID→Unicode overrides (from PDF.js getSupplementalGlyphMapForCalibri).
///
/// ArialBlack overrides for Polish/Central European characters whose GIDs
/// differ from the standard TrueType table (from PDF.js
/// getSupplementalGlyphMapForArialBlack).
static ARIAL_BLACK_SUPPLEMENT: &[(u16, u32)] = &[
    (227, 322), // ł
    (264, 261), // ą
    (291, 346), // Ś
];

/// These override the standard table entries for GIDs that differ in Calibri's
/// glyph ordering from the generic TrueType convention.
static CALIBRI_SUPPLEMENT: &[(u16, u32)] = &[
    (1, 32),
    (4, 65),
    (5, 192),
    (6, 193),
    (9, 196),
    (17, 66),
    (18, 67),
    (21, 268),
    (24, 68),
    (28, 69),
    (29, 200),
    (30, 201),
    (32, 282),
    (38, 70),
    (39, 71),
    (44, 72),
    (47, 73),
    (48, 204),
    (49, 205),
    (58, 74),
    (60, 75),
    (62, 76),
    (68, 77),
    (69, 78),
    (75, 79),
    (76, 210),
    (80, 214),
    (87, 80),
    (89, 81),
    (90, 82),
    (92, 344),
    (94, 83),
    (97, 352),
    (100, 84),
    (104, 85),
    (109, 220),
    (115, 86),
    (116, 87),
    (121, 88),
    (122, 89),
    (124, 221),
    (127, 90),
    (129, 381),
    (258, 97),
    (259, 224),
    (260, 225),
    (263, 228),
    (268, 261),
    (271, 98),
    (272, 99),
    (273, 263),
    (275, 269),
    (282, 100),
    (286, 101),
    (287, 232),
    (288, 233),
    (290, 283),
    (295, 281),
    (296, 102),
    (336, 103),
    (346, 104),
    (349, 105),
    (350, 236),
    (351, 237),
    (361, 106),
    (364, 107),
    (367, 108),
    (371, 322),
    (373, 109),
    (374, 110),
    (381, 111),
    (382, 242),
    (383, 243),
    (386, 246),
    (393, 112),
    (395, 113),
    (396, 114),
    (398, 345),
    (400, 115),
    (401, 347),
    (403, 353),
    (410, 116),
    (437, 117),
    (442, 252),
    (448, 118),
    (449, 119),
    (454, 120),
    (455, 121),
    (457, 253),
    (460, 122),
    (462, 382),
    (463, 380),
    (853, 44),
    (855, 58),
    (856, 46),
    (876, 47),
    (878, 45),
    (882, 45),
    (894, 40),
    (895, 41),
    (896, 91),
    (897, 93),
    (923, 64),
    (940, 163),
    (1004, 48),
    (1005, 49),
    (1006, 50),
    (1007, 51),
    (1008, 52),
    (1009, 53),
    (1010, 54),
    (1011, 55),
    (1012, 56),
    (1013, 57),
    (1081, 37),
    (1085, 43),
    (1086, 45),
];