shlesha 0.5.7

High-performance extensible transliteration library with hub-and-spoke architecture
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
use handlebars::Handlebars;
use rustc_hash::FxHashMap;
use serde_json::json;
use std::collections::BTreeSet;
use std::env;
use std::fs;
use std::path::{Path, PathBuf};

#[derive(serde::Deserialize, Debug, Clone)]
struct ScriptMetadata {
    name: String,
    #[allow(dead_code)]
    script_type: String,
    #[allow(dead_code)]
    has_implicit_a: bool,
    aliases: Option<Vec<String>>,
}

#[derive(serde::Deserialize, Debug, Clone)]
struct TokenMappings {
    vowels: Option<FxHashMap<String, TokenMapping>>, // "VowelA" -> ["a", "A"] or "VowelA" -> "a"
    consonants: Option<FxHashMap<String, TokenMapping>>, // "ConsonantK" -> ["k", "K"]
    vowel_signs: Option<FxHashMap<String, TokenMapping>>, // For abugida scripts
    marks: Option<FxHashMap<String, TokenMapping>>,  // "MarkAnusvara" -> ["M", "ṁ"]
    digits: Option<FxHashMap<String, TokenMapping>>, // "Digit0" -> "0"
    special: Option<FxHashMap<String, TokenMapping>>, // "SpecialKs" -> ["kS", "kṣ"]
    extended: Option<FxHashMap<String, TokenMapping>>, // "ExtendedQ" -> "q"
    vedic: Option<FxHashMap<String, TokenMapping>>,  // "MarkUdatta" -> "॑"
}

// Support both single string and array of strings for flexibility
#[derive(serde::Deserialize, serde::Serialize, Debug, Clone)]
#[serde(untagged)]
enum TokenMapping {
    Single(String),        // "a"
    Multiple(Vec<String>), // ["a", "A"]
}

#[derive(serde::Deserialize, Debug, Clone)]
struct CodegenConfig {
    #[allow(dead_code)]
    processor_type: String,
}

impl TokenMapping {
    #[allow(dead_code)]
    fn get_preferred(&self) -> String {
        match self {
            TokenMapping::Single(s) => s.clone(),
            TokenMapping::Multiple(vec) => vec.first().unwrap_or(&"".to_string()).clone(),
        }
    }
}

#[derive(serde::Deserialize, Debug, Clone)]
struct ScriptSchema {
    metadata: ScriptMetadata,
    target: Option<String>, // "alphabet_tokens" or "abugida_tokens" (optional for legacy schemas)
    mappings: TokenMappings,
    #[allow(dead_code)]
    codegen: Option<CodegenConfig>,
}

// Convert TokenMapping mappings to legacy String mappings for compatibility
#[allow(dead_code)]
fn flatten_token_mappings(mappings: &FxHashMap<String, TokenMapping>) -> FxHashMap<String, String> {
    mappings
        .iter()
        .map(|(k, v)| (k.clone(), v.get_preferred()))
        .collect()
}

fn main() {
    println!("cargo:rerun-if-changed=schemas/");
    println!("cargo:rerun-if-changed=templates/");

    if let Err(e) = generate_tokens_from_schemas() {
        println!("cargo:warning=Failed to generate tokens: {e}");
    }

    if let Err(e) = generate_schema_based_converters() {
        println!("cargo:warning=Failed to generate schema-based converters: {e}");
    }
}

/// Collect all unique tokens from schemas and generate tokens.rs
fn generate_tokens_from_schemas() -> Result<(), Box<dyn std::error::Error>> {
    let out_dir = PathBuf::from(env::var("OUT_DIR")?);
    let schemas_dir = Path::new("schemas");

    // Collections for unique tokens
    let mut abugida_vowels = BTreeSet::new();
    let mut abugida_vowel_signs = BTreeSet::new();
    let mut abugida_consonants = BTreeSet::new();
    let mut abugida_marks = BTreeSet::new();
    let mut abugida_special = BTreeSet::new();
    let mut abugida_digits = BTreeSet::new();
    let mut abugida_vedic = BTreeSet::new();

    let mut alphabet_vowels = BTreeSet::new();
    let mut alphabet_consonants = BTreeSet::new();
    let mut alphabet_marks = BTreeSet::new();
    let mut alphabet_special = BTreeSet::new();
    let mut alphabet_digits = BTreeSet::new();
    let mut alphabet_vedic = BTreeSet::new();

    // Process all YAML schemas
    if schemas_dir.exists() {
        for entry in fs::read_dir(schemas_dir)? {
            let entry = entry?;
            let path = entry.path();

            if path.extension().and_then(|s| s.to_str()) == Some("yaml") {
                let content = fs::read_to_string(&path)?;
                let schema: ScriptSchema = serde_yaml::from_str(&content)
                    .map_err(|e| format!("Failed to parse YAML schema {}: {e}", path.display()))?;

                // Skip debug schemas
                if schema.metadata.name == "abugida_tokens"
                    || schema.metadata.name == "alphabet_tokens"
                {
                    continue;
                }

                // Skip non-token schemas
                let target = match &schema.target {
                    Some(t) => t,
                    None => continue,
                };

                let is_abugida = target == "abugida_tokens";
                let is_alphabet = target == "alphabet_tokens";

                if !is_abugida && !is_alphabet {
                    continue;
                }

                // Collect tokens from each category
                if let Some(vowels) = &schema.mappings.vowels {
                    for token in vowels.keys() {
                        if is_abugida {
                            abugida_vowels.insert(token.clone());
                        } else {
                            alphabet_vowels.insert(token.clone());
                        }
                    }
                }

                if let Some(vowel_signs) = &schema.mappings.vowel_signs {
                    for token in vowel_signs.keys() {
                        if is_abugida {
                            abugida_vowel_signs.insert(token.clone());
                        }
                    }
                }

                if let Some(consonants) = &schema.mappings.consonants {
                    for token in consonants.keys() {
                        if is_abugida {
                            abugida_consonants.insert(token.clone());
                        } else {
                            alphabet_consonants.insert(token.clone());
                        }
                    }
                }

                if let Some(marks) = &schema.mappings.marks {
                    for token in marks.keys() {
                        if is_abugida {
                            abugida_marks.insert(token.clone());
                        } else {
                            alphabet_marks.insert(token.clone());
                        }
                    }
                }

                if let Some(special) = &schema.mappings.special {
                    for token in special.keys() {
                        if is_abugida {
                            abugida_special.insert(token.clone());
                        } else {
                            alphabet_special.insert(token.clone());
                        }
                    }
                }

                if let Some(digits) = &schema.mappings.digits {
                    for token in digits.keys() {
                        if is_abugida {
                            abugida_digits.insert(token.clone());
                        } else {
                            alphabet_digits.insert(token.clone());
                        }
                    }
                }

                if let Some(vedic) = &schema.mappings.vedic {
                    for token in vedic.keys() {
                        if is_abugida {
                            abugida_vedic.insert(token.clone());
                        } else {
                            alphabet_vedic.insert(token.clone());
                        }
                    }
                }
            }
        }
    }

    // Generate tokens.rs using template
    let mut handlebars = Handlebars::new();
    handlebars.set_strict_mode(false);
    handlebars.register_escape_fn(handlebars::no_escape);
    handlebars.register_template_file("tokens", "templates/tokens.hbs")?;

    // Generate vowel to sign mappings
    let mut vowel_to_sign_mappings = Vec::new();
    for vowel in &abugida_vowels {
        if vowel == "VowelA" {
            continue; // No sign for short 'a'
        }
        // Extract the vowel suffix (e.g., "Aa" from "VowelAa")
        if let Some(suffix) = vowel.strip_prefix("Vowel") {
            let sign_name = format!("VowelSign{}", suffix);
            if abugida_vowel_signs.contains(&sign_name) {
                vowel_to_sign_mappings.push(json!({
                    "vowel": vowel,
                    "sign": sign_name,
                }));
            }
        }
    }

    // Generate same sound mappings (tokens with same name exist in both systems)
    let mut same_sound_mappings = Vec::new();
    let mut abugida_to_alphabet_mappings = Vec::new();
    let mut alphabet_to_abugida_mappings = Vec::new();

    // Collect all abugida tokens
    let all_abugida_tokens: Vec<_> = abugida_vowels
        .iter()
        .chain(abugida_consonants.iter())
        .chain(abugida_marks.iter())
        .chain(abugida_special.iter())
        .chain(abugida_vedic.iter())
        .chain(abugida_digits.iter())
        .collect();

    // Collect all alphabet tokens
    let all_alphabet_tokens: Vec<_> = alphabet_vowels
        .iter()
        .chain(alphabet_consonants.iter())
        .chain(alphabet_marks.iter())
        .chain(alphabet_special.iter())
        .chain(alphabet_vedic.iter())
        .chain(alphabet_digits.iter())
        .collect();

    // Generate direct mappings where token names match
    for abugida_token in &all_abugida_tokens {
        if all_alphabet_tokens.iter().any(|a| a == abugida_token) {
            same_sound_mappings.push(json!({
                "abugida": abugida_token,
                "alphabet": abugida_token,
            }));
            abugida_to_alphabet_mappings.push(json!({
                "from": abugida_token,
                "to": abugida_token,
            }));
        }
    }

    // Generate reverse mappings
    for alphabet_token in &all_alphabet_tokens {
        if all_abugida_tokens.iter().any(|a| a == alphabet_token) {
            alphabet_to_abugida_mappings.push(json!({
                "from": alphabet_token,
                "to": alphabet_token,
            }));
        }
    }

    // Add vowel sign to vowel mappings
    for sign in &abugida_vowel_signs {
        // Extract the vowel part from VowelSignXxx
        if let Some(vowel_suffix) = sign.strip_prefix("VowelSign") {
            let vowel_name = format!("Vowel{}", vowel_suffix);
            if alphabet_vowels.contains(&vowel_name) {
                abugida_to_alphabet_mappings.push(json!({
                    "from": sign,
                    "to": vowel_name,
                }));
            }
        }
    }

    // Handle special cases where tokens don't exist in one system
    // These could be read from schema files in the future
    let special_mappings = vec![
        // If alphabet doesn't have long e/o, they still map to themselves for preservation
        ("VowelEe", "VowelEe"),
        ("VowelOo", "VowelOo"),
        // Vocalic L often doesn't exist in many scripts
        ("VowelL", "VowelL"),
        ("VowelLl", "VowelLl"),
    ];

    for (abugida, alphabet) in special_mappings {
        if (abugida_vowels.contains(abugida) || abugida_marks.contains(abugida))
            && !alphabet_vowels.contains(alphabet)
            && !alphabet_marks.contains(alphabet)
        {
            // This token exists in abugida but not alphabet - it will be preserved as-is
            abugida_to_alphabet_mappings.push(json!({
                "from": abugida,
                "to": abugida,  // Map to itself for preservation
            }));
        }
    }

    let template_data = json!({
        "abugida_vowels": abugida_vowels.into_iter().collect::<Vec<_>>(),
        "abugida_vowel_signs": abugida_vowel_signs.into_iter().collect::<Vec<_>>(),
        "abugida_consonants": abugida_consonants.into_iter().collect::<Vec<_>>(),
        "abugida_marks": abugida_marks.into_iter().collect::<Vec<_>>(),
        "abugida_special": abugida_special.into_iter().collect::<Vec<_>>(),
        "abugida_vedic": abugida_vedic.into_iter().collect::<Vec<_>>(),
        "abugida_digits": abugida_digits.into_iter().collect::<Vec<_>>(),
        "alphabet_vowels": alphabet_vowels.into_iter().collect::<Vec<_>>(),
        "alphabet_consonants": alphabet_consonants.into_iter().collect::<Vec<_>>(),
        "alphabet_marks": alphabet_marks.into_iter().collect::<Vec<_>>(),
        "alphabet_special": alphabet_special.into_iter().collect::<Vec<_>>(),
        "alphabet_vedic": alphabet_vedic.into_iter().collect::<Vec<_>>(),
        "alphabet_digits": alphabet_digits.into_iter().collect::<Vec<_>>(),
        "vowel_to_sign_mappings": vowel_to_sign_mappings,
        "same_sound_mappings": same_sound_mappings,
        "abugida_to_alphabet_mappings": abugida_to_alphabet_mappings,
        "alphabet_to_abugida_mappings": alphabet_to_abugida_mappings,
    });

    let tokens_code = handlebars.render("tokens", &template_data)?;
    fs::write(out_dir.join("tokens_generated.rs"), tokens_code)?;

    Ok(())
}

fn generate_schema_based_converters() -> Result<(), Box<dyn std::error::Error>> {
    let out_dir = PathBuf::from(env::var("OUT_DIR")?);
    let schemas_dir = Path::new("schemas");

    // Initialize Handlebars template engine - token-based only!
    let mut handlebars = Handlebars::new();
    handlebars.set_strict_mode(false);
    handlebars.register_escape_fn(handlebars::no_escape);
    handlebars.register_template_file(
        "token_based_converter",
        "templates/token_based_converter.hbs",
    )?;
    handlebars.register_template_file("direct_converter", "templates/direct_converter.hbs")?;

    // Register helper functions for templates
    handlebars.register_helper("uppercase", Box::new(uppercase_helper));
    handlebars.register_helper("lowercase", Box::new(lowercase_helper));
    handlebars.register_helper("capitalize", Box::new(capitalize_helper));
    handlebars.register_helper("escape", Box::new(escape_helper));
    handlebars.register_helper("eq", Box::new(eq_helper));

    let mut generated_code = String::new();
    let mut converter_registrations = Vec::new();
    let mut schemas = Vec::new();

    // Add header with all necessary imports once
    generated_code.push_str(
        r#"
// Auto-generated converters from TOML/YAML schemas
// DO NOT EDIT - Generated by build.rs at compile time

#[allow(unused_imports)]
#[allow(unreachable_patterns)]
#[allow(dead_code)]
#[allow(clippy::new_without_default)]
#[allow(clippy::clone_on_copy)]
#[allow(clippy::match_like_matches_macro)]
#[allow(clippy::duplicated_attributes)]

use once_cell::sync::Lazy;
use crate::modules::hub::HubFormat;
use crate::modules::hub::tokens::{AbugidaToken, AlphabetToken, HubToken, HubTokenSequence};
use aho_corasick::AhoCorasick;

"#,
    );

    // Process YAML schemas
    if schemas_dir.exists() {
        for entry in fs::read_dir(schemas_dir)? {
            let entry = entry?;
            let path = entry.path();

            if path.extension().and_then(|s| s.to_str()) == Some("yaml") {
                println!("cargo:rerun-if-changed={}", path.display());

                let content = fs::read_to_string(&path)?;
                let schema: ScriptSchema = serde_yaml::from_str(&content)
                    .map_err(|e| format!("Failed to parse YAML schema {}: {e}", path.display()))?;

                // Add schema to collection for Hub generation
                schemas.push(schema.clone());

                // Only process token-based schemas
                if let Some(ref target) = schema.target {
                    if target != "alphabet_tokens" && target != "abugida_tokens" {
                        continue; // Skip non-token schemas
                    }
                } else {
                    continue; // Skip schemas without target
                }

                let converter_code =
                    generate_converter_from_schema(&handlebars, &schema).map_err(|e| {
                        format!(
                            "Failed to generate converter for {}: {e}",
                            schema.metadata.name
                        )
                    })?;
                generated_code.push_str(&converter_code);

                // Only register token-based converters!
                if let Some(ref target) = schema.target {
                    if target == "alphabet_tokens" || target == "abugida_tokens" {
                        converter_registrations.push(format!(
                            "{}Converter",
                            capitalize_first(&schema.metadata.name)
                        ));
                    }
                }

                // No more Roman → Devanagari converters - everything goes through tokens!
            }
        }
    }

    // Hub converter is no longer needed - using trait_based_converter instead

    // Generate direct converters for common script pairs to bypass hub overhead
    if let Ok(direct_code) = generate_direct_converters(&handlebars, &schemas) {
        fs::write(out_dir.join("direct_converters_generated.rs"), direct_code)?;
    }

    // Generate token-based converter registry with aliases
    let token_registrations = converter_registrations
        .iter()
        .map(|name| format!("        Box::new({name}::new()),"))
        .collect::<Vec<_>>()
        .join("\n");

    // Generate registration with aliases
    let token_registrations_with_aliases = schemas
        .iter()
        .filter_map(|schema| {
            let converter_name = format!(
                "{}Converter",
                schema
                    .metadata
                    .name
                    .chars()
                    .next()
                    .unwrap()
                    .to_uppercase()
                    .to_string()
                    + &schema.metadata.name[1..]
            );

            if converter_registrations.contains(&converter_name) {
                let aliases = schema
                    .metadata
                    .aliases
                    .as_ref()
                    .map(|aliases| {
                        aliases
                            .iter()
                            .map(|alias| format!("\"{alias}\".to_string()"))
                            .collect::<Vec<_>>()
                            .join(", ")
                    })
                    .unwrap_or_default();

                Some(format!(
                    "        (Box::new({converter_name}::new()), vec![{aliases}]),"
                ))
            } else {
                None
            }
        })
        .collect::<Vec<_>>()
        .join("\n");

    generated_code.push_str(&format!(r#"
/// Register token-based converters
pub fn register_schema_generated_converters(_registry: &mut crate::modules::script_converter::ScriptConverterRegistry) {{
    // Old string-based registration system has been removed
    // Token-based converters are managed separately
}}

/// Get all token-based converters (legacy)
pub fn register_token_converters() -> Vec<Box<dyn crate::modules::script_converter::TokenConverter>> {{
    vec![
{token_registrations}
    ]
}}

/// Get all token-based converters with their aliases
pub fn register_token_converters_with_aliases() -> Vec<(Box<dyn crate::modules::script_converter::TokenConverter>, Vec<String>)> {{
    vec![
{token_registrations_with_aliases}
    ]
}}
"#));

    // Generate script type helper functions
    let mut brahmic_scripts = Vec::new();
    let mut roman_scripts = Vec::new();

    for schema in &schemas {
        match schema.metadata.script_type.as_str() {
            "brahmic" => {
                brahmic_scripts.push(format!("\"{}\"", schema.metadata.name.to_lowercase()));
                if let Some(aliases) = &schema.metadata.aliases {
                    for alias in aliases {
                        brahmic_scripts.push(format!("\"{}\"", alias.to_lowercase()));
                    }
                }
            }
            "roman" => {
                roman_scripts.push(format!("\"{}\"", schema.metadata.name.to_lowercase()));
                if let Some(aliases) = &schema.metadata.aliases {
                    for alias in aliases {
                        roman_scripts.push(format!("\"{}\"", alias.to_lowercase()));
                    }
                }
            }
            _ => {}
        }
    }

    let script_helpers = format!(
        r#"
/// Check if a script is an Indic/Brahmic script
pub fn is_indic_script(script: &str) -> bool {{
    matches!(
        script.to_lowercase().as_str(),
        {}
    )
}}

/// Check if a script is a Roman script
pub fn is_roman_script(script: &str) -> bool {{
    matches!(
        script.to_lowercase().as_str(),
        {}
    )
}}
"#,
        brahmic_scripts.join("\n            | "),
        roman_scripts.join("\n            | ")
    );

    generated_code.push_str(&script_helpers);

    // Write generated code
    fs::write(out_dir.join("schema_generated.rs"), generated_code)?;
    Ok(())
}

fn generate_converter_from_schema(
    handlebars: &Handlebars,
    schema: &ScriptSchema,
) -> Result<String, Box<dyn std::error::Error>> {
    // Token-based only! All converters must use tokens
    generate_token_based_converter(handlebars, schema)
}

fn capitalize_first(s: &str) -> String {
    // Convert kebab-case and snake_case to PascalCase
    s.split(&['-', '_'][..])
        .map(|part| {
            let mut chars = part.chars();
            match chars.next() {
                None => String::new(),
                Some(first) => first.to_uppercase().collect::<String>() + chars.as_str(),
            }
        })
        .collect()
}

fn escape_string(s: &str) -> String {
    s.replace('\\', "\\\\").replace('"', "\\\"")
}

// Template-based converter generators

#[allow(dead_code)]
fn generate_roman_converter_with_template(
    handlebars: &Handlebars,
    struct_name: &str,
    script_name: &str,
    mappings: &FxHashMap<String, String>,
    metadata: &ScriptMetadata,
    canonical_forms: &Option<FxHashMap<String, String>>,
    use_aho_corasick: bool,
) -> Result<String, Box<dyn std::error::Error>> {
    // Sort by length (longest first) for proper matching
    let mut sorted_mappings: Vec<_> = mappings.iter().collect();
    sorted_mappings.sort_by_key(|b| std::cmp::Reverse(b.0.len()));

    // Prepare ALL reverse mappings for template (not just multi-character ones)
    // Note: reverse mapping means ISO → source_script, so ISO should be the key
    let mut reverse_mappings: Vec<(&str, &str)> = sorted_mappings
        .iter()
        .map(|(from, to)| (to.as_str(), from.as_str())) // (ISO, source_script)
        .collect();

    // Sort by length (longest first) for proper matching priority
    reverse_mappings.sort_by_key(|b| std::cmp::Reverse(b.0.len()));

    // Create reverse mappings with preference for canonical forms
    let mut reverse_priority_mappings: FxHashMap<&str, &str> = FxHashMap::default();

    // If canonical forms are specified in schema, use them; otherwise use first occurrence
    if let Some(canonical) = canonical_forms {
        // Pre-build sets for faster lookup (avoids repeated iteration)
        let iso_values: std::collections::HashSet<&str> =
            mappings.values().map(|s| s.as_str()).collect();
        let source_keys: std::collections::HashSet<&str> =
            mappings.keys().map(|s| s.as_str()).collect();

        // First, add all mappings
        for (iso_char, source_char) in reverse_mappings {
            reverse_priority_mappings.insert(iso_char, source_char);
        }

        // Then override with canonical forms where specified
        for (iso_char, canonical_source) in canonical {
            // Quick verification using pre-built sets
            if iso_values.contains(iso_char.as_str())
                && source_keys.contains(canonical_source.as_str())
            {
                reverse_priority_mappings.insert(iso_char, canonical_source);
            }
        }
    } else {
        // No canonical forms specified, use first occurrence (original behavior)
        for (iso_char, source_char) in reverse_mappings {
            reverse_priority_mappings
                .entry(iso_char)
                .or_insert(source_char);
        }
    }

    // Convert to JSON for template
    let mappings_json: FxHashMap<&str, &str> = mappings
        .iter()
        .map(|(k, v)| (k.as_str(), v.as_str()))
        .collect();

    // Add script-specific aliases
    let aliases = match script_name {
        "harvard_kyoto" => vec![
            ("hk_to_iso", "harvard_kyoto_to_iso"),
            ("iso_to_hk", "iso_to_harvard_kyoto"),
        ],
        _ => vec![],
    };

    let template_data = json!({
        "struct_name": struct_name,
        "script_name": script_name,
        "has_implicit_a": metadata.has_implicit_a,
        "mappings": mappings_json,
        "reverse_priority_mappings": reverse_priority_mappings,
        "aliases": aliases
    });

    let template_name = if use_aho_corasick {
        "roman_converter_aho_corasick"
    } else {
        "roman_converter"
    };

    let code = handlebars.render(template_name, &template_data)?;
    Ok(code)
}

// Handlebars helper functions
fn uppercase_helper(
    h: &handlebars::Helper,
    _: &handlebars::Handlebars,
    _: &handlebars::Context,
    _rc: &mut handlebars::RenderContext,
    out: &mut dyn handlebars::Output,
) -> handlebars::HelperResult {
    let param = h.param(0).unwrap().value().as_str().unwrap();
    out.write(&param.to_uppercase())?;
    Ok(())
}

fn lowercase_helper(
    h: &handlebars::Helper,
    _: &handlebars::Handlebars,
    _: &handlebars::Context,
    _rc: &mut handlebars::RenderContext,
    out: &mut dyn handlebars::Output,
) -> handlebars::HelperResult {
    let param = h.param(0).unwrap().value().as_str().unwrap();
    out.write(&param.to_lowercase())?;
    Ok(())
}

fn capitalize_helper(
    h: &handlebars::Helper,
    _: &handlebars::Handlebars,
    _: &handlebars::Context,
    _rc: &mut handlebars::RenderContext,
    out: &mut dyn handlebars::Output,
) -> handlebars::HelperResult {
    let param = h.param(0).unwrap().value().as_str().unwrap();
    out.write(&capitalize_first(param))?;
    Ok(())
}

fn escape_helper(
    h: &handlebars::Helper,
    _: &handlebars::Handlebars,
    _: &handlebars::Context,
    _rc: &mut handlebars::RenderContext,
    out: &mut dyn handlebars::Output,
) -> handlebars::HelperResult {
    let param = h.param(0).unwrap().value().as_str().unwrap();
    out.write(&escape_string(param))?;
    Ok(())
}

fn eq_helper(
    h: &handlebars::Helper,
    _: &handlebars::Handlebars,
    _: &handlebars::Context,
    _rc: &mut handlebars::RenderContext,
    out: &mut dyn handlebars::Output,
) -> handlebars::HelperResult {
    let param1 = h.param(0).unwrap().value().as_str().unwrap();
    let param2 = h.param(1).unwrap().value().as_str().unwrap();
    if param1 == param2 {
        out.write("true")?;
    } else {
        out.write("false")?;
    }
    Ok(())
}

#[allow(dead_code)]
fn generate_indic_converter_with_template(
    handlebars: &Handlebars,
    struct_name: &str,
    script_name: &str,
    mappings: &FxHashMap<String, String>,
    metadata: &ScriptMetadata,
) -> Result<String, Box<dyn std::error::Error>> {
    // Filter to only single-character mappings for optimized HashMap<char, char>
    let char_mappings: FxHashMap<&str, &str> = mappings
        .iter()
        .filter(|(from, to)| from.chars().count() == 1 && to.chars().count() == 1)
        .map(|(k, v)| (k.as_str(), v.as_str()))
        .collect();

    let template_data = json!({
        "struct_name": struct_name,
        "script_name": script_name,
        "has_implicit_a": metadata.has_implicit_a,
        "char_mappings": char_mappings
    });

    let code = handlebars.render("indic_standard_converter", &template_data)?;
    Ok(code)
}

#[allow(dead_code)]
fn generate_extended_indic_converter_with_template(
    handlebars: &Handlebars,
    struct_name: &str,
    script_name: &str,
    mappings: &FxHashMap<String, String>,
    metadata: &ScriptMetadata,
) -> Result<String, Box<dyn std::error::Error>> {
    // Use all mappings as strings for complex script support
    let string_mappings: FxHashMap<&str, &str> = mappings
        .iter()
        .map(|(k, v)| (k.as_str(), v.as_str()))
        .collect();

    let template_data = json!({
        "struct_name": struct_name,
        "script_name": script_name,
        "has_implicit_a": metadata.has_implicit_a,
        "string_mappings": string_mappings
    });

    let code = handlebars.render("indic_extended_converter", &template_data)?;
    Ok(code)
}

#[allow(dead_code)]
fn generate_roman_to_devanagari_converter(
    handlebars: &Handlebars,
    schema: &ScriptSchema,
) -> Result<String, Box<dyn std::error::Error>> {
    let script_name = &schema.metadata.name;
    let struct_name = format!("{}DevanagariConverter", capitalize_first(script_name));

    // Get ISO-15919 to Devanagari mappings from hub (at compile time)
    let iso_to_deva_mappings = get_iso_to_devanagari_mappings();

    // Compose Roman → ISO-15919 → Devanagari mappings at COMPILE TIME
    let mut roman_to_deva_mappings = FxHashMap::default();

    // Process all mapping categories from the schema
    let all_roman_mappings = [
        &schema.mappings.vowels,
        &schema.mappings.consonants,
        &schema.mappings.marks,
        &schema.mappings.digits,
        &schema.mappings.special,
    ];

    for mappings in all_roman_mappings.iter().copied().flatten() {
        for (roman_key, iso_token_mapping) in mappings.iter() {
            let iso_value = iso_token_mapping.get_preferred(); // Get the preferred ISO mapping

            // Direct mapping for all keys
            if let Some(deva_value) = iso_to_deva_mappings.get(&iso_value) {
                roman_to_deva_mappings.insert(roman_key.clone(), deva_value.clone());
            }

            // For consonants, also try with 'a' suffix (for inherent vowel)
            let iso_with_a = format!("{iso_value}a");
            if let Some(deva_value) = iso_to_deva_mappings.get(&iso_with_a) {
                let roman_with_a = format!("{roman_key}a");
                roman_to_deva_mappings.insert(roman_with_a, deva_value.clone());
            }
        }
    }

    // Add vowel sign mappings by deriving them from hub data
    let vowels_as_strings = schema.mappings.vowels.as_ref().map(flatten_token_mappings);
    add_vowel_sign_mappings(
        &mut roman_to_deva_mappings,
        &iso_to_deva_mappings,
        &vowels_as_strings,
    );

    // Sort by length (longest first) for proper matching
    let mut sorted_mappings: Vec<_> = roman_to_deva_mappings.iter().collect();
    sorted_mappings.sort_by_key(|b| std::cmp::Reverse(b.0.len()));

    // Convert to template format - use the original String keys
    let mappings_for_template = &roman_to_deva_mappings;

    let template_data = json!({
        "struct_name": struct_name,
        "script_name": format!("{}_devanagari", script_name),
        "original_script": script_name,
        "has_implicit_a": schema.metadata.has_implicit_a,
        "mappings": mappings_for_template,
    });

    let code = handlebars.render("roman_to_devanagari_direct", &template_data)?;
    Ok(code)
}

#[allow(dead_code)]
fn get_iso_to_devanagari_mappings() -> FxHashMap<String, String> {
    // Hardcode the hub mappings since we can't access the crate from build.rs
    // This is still better than before since we're deriving vowel signs dynamically
    let mut iso_to_deva_mappings = FxHashMap::default();

    // Core vowels
    iso_to_deva_mappings.insert("a".to_string(), "अ".to_string());
    iso_to_deva_mappings.insert("ā".to_string(), "आ".to_string());
    iso_to_deva_mappings.insert("i".to_string(), "इ".to_string());
    iso_to_deva_mappings.insert("ī".to_string(), "ई".to_string());
    iso_to_deva_mappings.insert("u".to_string(), "उ".to_string());
    iso_to_deva_mappings.insert("ū".to_string(), "ऊ".to_string());
    iso_to_deva_mappings.insert("r̥".to_string(), "ऋ".to_string());
    iso_to_deva_mappings.insert("r̥̄".to_string(), "ॠ".to_string());
    iso_to_deva_mappings.insert("l̥".to_string(), "ऌ".to_string());
    iso_to_deva_mappings.insert("l̥̄".to_string(), "ॡ".to_string());
    iso_to_deva_mappings.insert("e".to_string(), "ए".to_string());
    iso_to_deva_mappings.insert("ai".to_string(), "ऐ".to_string());
    iso_to_deva_mappings.insert("o".to_string(), "ओ".to_string());
    iso_to_deva_mappings.insert("au".to_string(), "औ".to_string());

    // Consonants with inherent 'a' vowel
    iso_to_deva_mappings.insert("ka".to_string(), "क".to_string());
    iso_to_deva_mappings.insert("kha".to_string(), "ख".to_string());
    iso_to_deva_mappings.insert("ga".to_string(), "ग".to_string());
    iso_to_deva_mappings.insert("gha".to_string(), "घ".to_string());
    iso_to_deva_mappings.insert("ṅa".to_string(), "ङ".to_string());
    iso_to_deva_mappings.insert("ca".to_string(), "च".to_string());
    iso_to_deva_mappings.insert("cha".to_string(), "छ".to_string());
    iso_to_deva_mappings.insert("ja".to_string(), "ज".to_string());
    iso_to_deva_mappings.insert("jha".to_string(), "झ".to_string());
    iso_to_deva_mappings.insert("ña".to_string(), "ञ".to_string());
    iso_to_deva_mappings.insert("ṭa".to_string(), "ट".to_string());
    iso_to_deva_mappings.insert("ṭha".to_string(), "ठ".to_string());
    iso_to_deva_mappings.insert("ḍa".to_string(), "ड".to_string());
    iso_to_deva_mappings.insert("ḍha".to_string(), "ढ".to_string());
    iso_to_deva_mappings.insert("ṇa".to_string(), "ण".to_string());
    iso_to_deva_mappings.insert("ta".to_string(), "त".to_string());
    iso_to_deva_mappings.insert("tha".to_string(), "थ".to_string());
    iso_to_deva_mappings.insert("da".to_string(), "द".to_string());
    iso_to_deva_mappings.insert("dha".to_string(), "ध".to_string());
    iso_to_deva_mappings.insert("na".to_string(), "न".to_string());
    iso_to_deva_mappings.insert("pa".to_string(), "प".to_string());
    iso_to_deva_mappings.insert("pha".to_string(), "फ".to_string());
    iso_to_deva_mappings.insert("ba".to_string(), "ब".to_string());
    iso_to_deva_mappings.insert("bha".to_string(), "भ".to_string());
    iso_to_deva_mappings.insert("ma".to_string(), "म".to_string());
    iso_to_deva_mappings.insert("ya".to_string(), "य".to_string());
    iso_to_deva_mappings.insert("ra".to_string(), "र".to_string());
    iso_to_deva_mappings.insert("la".to_string(), "ल".to_string());
    iso_to_deva_mappings.insert("va".to_string(), "व".to_string());
    iso_to_deva_mappings.insert("śa".to_string(), "श".to_string());
    iso_to_deva_mappings.insert("ṣa".to_string(), "ष".to_string());
    iso_to_deva_mappings.insert("sa".to_string(), "स".to_string());
    iso_to_deva_mappings.insert("ha".to_string(), "ह".to_string());

    // Consonants with other vowels for vowel sign extraction
    iso_to_deva_mappings.insert("kā".to_string(), "का".to_string());
    iso_to_deva_mappings.insert("ki".to_string(), "कि".to_string());
    iso_to_deva_mappings.insert("kī".to_string(), "की".to_string());
    iso_to_deva_mappings.insert("ku".to_string(), "कु".to_string());
    iso_to_deva_mappings.insert("kū".to_string(), "कू".to_string());
    iso_to_deva_mappings.insert("kr̥".to_string(), "कृ".to_string());
    iso_to_deva_mappings.insert("kr̥̄".to_string(), "कॄ".to_string());
    iso_to_deva_mappings.insert("kl̥".to_string(), "कॢ".to_string());
    iso_to_deva_mappings.insert("kl̥̄".to_string(), "कॣ".to_string());
    iso_to_deva_mappings.insert("ke".to_string(), "के".to_string());
    iso_to_deva_mappings.insert("kai".to_string(), "कै".to_string());
    iso_to_deva_mappings.insert("ko".to_string(), "को".to_string());
    iso_to_deva_mappings.insert("kau".to_string(), "कौ".to_string());

    // Marks
    iso_to_deva_mappings.insert("ṁ".to_string(), "ं".to_string());
    iso_to_deva_mappings.insert("ḥ".to_string(), "ः".to_string());
    iso_to_deva_mappings.insert("m̐".to_string(), "ँ".to_string());
    iso_to_deva_mappings.insert("'".to_string(), "ऽ".to_string());

    // Digits
    iso_to_deva_mappings.insert("0".to_string(), "०".to_string());
    iso_to_deva_mappings.insert("1".to_string(), "१".to_string());
    iso_to_deva_mappings.insert("2".to_string(), "२".to_string());
    iso_to_deva_mappings.insert("3".to_string(), "३".to_string());
    iso_to_deva_mappings.insert("4".to_string(), "४".to_string());
    iso_to_deva_mappings.insert("5".to_string(), "५".to_string());
    iso_to_deva_mappings.insert("6".to_string(), "६".to_string());
    iso_to_deva_mappings.insert("7".to_string(), "७".to_string());
    iso_to_deva_mappings.insert("8".to_string(), "८".to_string());
    iso_to_deva_mappings.insert("9".to_string(), "९".to_string());

    // Special
    iso_to_deva_mappings.insert("ḷa".to_string(), "ळ".to_string());
    iso_to_deva_mappings.insert("।".to_string(), "।".to_string());
    iso_to_deva_mappings.insert("॥".to_string(), "॥".to_string());
    iso_to_deva_mappings.insert("kṣa".to_string(), "क्ष".to_string());
    iso_to_deva_mappings.insert("jña".to_string(), "ज्ञ".to_string());

    // Nukta consonants (for extended characters) - using precomposed forms
    iso_to_deva_mappings.insert("qa".to_string(), "\u{0958}".to_string()); // क़ precomposed
    iso_to_deva_mappings.insert("ḵẖa".to_string(), "\u{0959}".to_string()); // ख़ precomposed
    iso_to_deva_mappings.insert("ġa".to_string(), "\u{095A}".to_string()); // ग़ precomposed
    iso_to_deva_mappings.insert("za".to_string(), "\u{095B}".to_string()); // ज़ precomposed
    iso_to_deva_mappings.insert("ṛa".to_string(), "\u{095C}".to_string()); // ड़ precomposed
    iso_to_deva_mappings.insert("ṛha".to_string(), "\u{095D}".to_string()); // ढ़ precomposed
    iso_to_deva_mappings.insert("fa".to_string(), "\u{095E}".to_string()); // फ़ precomposed
    iso_to_deva_mappings.insert("ẏa".to_string(), "\u{095F}".to_string()); // य़ precomposed

    iso_to_deva_mappings
}

#[allow(dead_code)]
fn add_vowel_sign_mappings(
    roman_to_deva_mappings: &mut FxHashMap<String, String>,
    iso_to_deva_mappings: &FxHashMap<String, String>,
    vowel_mappings: &Option<FxHashMap<String, String>>,
) {
    if let Some(vowels) = vowel_mappings {
        for (roman_vowel, iso_vowel) in vowels.iter() {
            // Get the vowel sign by testing with a sample consonant
            if let (Some(ka_vowel), Some(ka_base)) = (
                iso_to_deva_mappings.get(&format!("k{iso_vowel}")),
                iso_to_deva_mappings.get("ka"),
            ) {
                // Extract vowel sign by comparing ka + vowel vs ka
                if ka_vowel != ka_base && ka_vowel.starts_with(ka_base) {
                    let vowel_sign = &ka_vowel[ka_base.len()..];
                    if !vowel_sign.is_empty() {
                        roman_to_deva_mappings.insert(
                            format!("__vowel_sign_{roman_vowel}"),
                            vowel_sign.to_string(),
                        );
                    }
                }
            }
        }
    }
}

fn generate_token_based_converter(
    handlebars: &Handlebars,
    schema: &ScriptSchema,
) -> Result<String, Box<dyn std::error::Error>> {
    let script_name = &schema.metadata.name;
    let struct_name = format!("{}Converter", capitalize_first(script_name));
    let is_alphabet = schema.target.as_deref() == Some("alphabet_tokens");

    // Collect all mappings with their categories
    let mut mappings = Vec::new();

    if let Some(ref vowels) = schema.mappings.vowels {
        let entries: Vec<_> = vowels
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                // Debug output to understand the issue
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Vowels",
            "entries": entries
        }));
    }

    if let Some(ref consonants) = schema.mappings.consonants {
        let entries: Vec<_> = consonants
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Consonants",
            "entries": entries
        }));
    }

    if let Some(ref vowel_signs) = schema.mappings.vowel_signs {
        let entries: Vec<_> = vowel_signs
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Vowel Signs",
            "entries": entries
        }));
    }

    if let Some(ref marks) = schema.mappings.marks {
        let entries: Vec<_> = marks
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Marks",
            "entries": entries
        }));
    }

    if let Some(ref special) = schema.mappings.special {
        let entries: Vec<_> = special
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Special",
            "entries": entries
        }));
    }

    if let Some(ref extended) = schema.mappings.extended {
        let entries: Vec<_> = extended
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Extended",
            "entries": entries
        }));
    }

    if let Some(ref vedic) = schema.mappings.vedic {
        let entries: Vec<_> = vedic
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Vedic",
            "entries": entries
        }));
    }

    if let Some(ref digits) = schema.mappings.digits {
        let entries: Vec<_> = digits
            .iter()
            .map(|(token, mapping)| {
                let (preferred, all_inputs) = match mapping {
                    TokenMapping::Single(s) => (s.clone(), vec![s.clone()]),
                    TokenMapping::Multiple(v) => (v[0].clone(), v.clone()),
                };
                json!({
                    "token": token,
                    "preferred": preferred,
                    "all_inputs": all_inputs
                })
            })
            .collect();
        mappings.push(json!({
            "category": "Digits",
            "entries": entries
        }));
    }

    // Check if there are multi-character mappings
    let has_multi_char_mappings = schema
        .mappings
        .vowels
        .as_ref()
        .map(|m| m.keys().any(|k| k.len() > 1))
        .unwrap_or(false)
        || schema
            .mappings
            .consonants
            .as_ref()
            .map(|m| m.keys().any(|k| k.len() > 1))
            .unwrap_or(false)
        || schema
            .mappings
            .special
            .as_ref()
            .map(|m| m.keys().any(|k| k.len() > 1))
            .unwrap_or(false);

    let template_data = json!({
        "struct_name": struct_name,
        "script_name": script_name,
        "is_alphabet": is_alphabet,
        "target_type": schema.target.as_ref().unwrap_or(&"unknown".to_string()),
        "mappings": mappings,
        "has_multi_char_mappings": has_multi_char_mappings,
    });

    handlebars
        .render("token_based_converter", &template_data)
        .map_err(|e| format!("Template rendering failed: {e}").into())
}

/// Generate direct converters for common script pairs to bypass hub overhead
fn generate_direct_converters(
    handlebars: &Handlebars,
    schemas: &[ScriptSchema],
) -> Result<String, Box<dyn std::error::Error>> {
    let mut generated_code = String::new();

    // Add imports and header
    generated_code.push_str(
        r#"
// Auto-generated direct converters - bypass hub for maximum performance
// DO NOT EDIT - Generated by build.rs at compile time

#[allow(unreachable_patterns)]
#[allow(dead_code)]

use crate::modules::script_converter::ConverterError;
use aho_corasick::AhoCorasick;
use once_cell::sync::Lazy;

"#,
    );

    // Common high-performance conversion pairs
    let conversion_pairs = vec![
        // Roman ↔ Roman (very common)
        ("iast", "slp1"),
        ("slp1", "iast"),
        ("iast", "itrans"),
        ("itrans", "iast"),
        ("iast", "harvard_kyoto"),
        ("harvard_kyoto", "iast"),
        // Indic → Roman (performance critical)
        ("devanagari", "iast"),
        ("devanagari", "slp1"),
        ("telugu", "iast"),
        ("telugu", "slp1"),
        ("bengali", "iast"),
        ("tamil", "iast"),
        ("gujarati", "iast"),
        // Roman → Indic (also important)
        ("iast", "devanagari"),
        ("slp1", "devanagari"),
        ("iast", "telugu"),
        ("slp1", "telugu"),
        ("iast", "bengali"),
        ("iast", "tamil"),
        ("iast", "kannada"),
        ("iast", "gujarati"),
        // New Vedic scripts - high priority ones
        ("iast", "grantha"),
        ("iast", "sharada"),
        ("iast", "nandinagari"),
        ("iast", "newa"),
        ("iast", "siddham"),
        ("iast", "modi"),
        ("iast", "bhaiksuki"),
        ("iast", "kaithi"),
        ("iast", "takri"),
        ("iast", "dogra"),
    ];

    // Find schemas by name
    let schema_map: FxHashMap<_, _> = schemas
        .iter()
        .map(|s| (s.metadata.name.as_str(), s))
        .collect();

    for (from_script, to_script) in &conversion_pairs {
        if let (Some(from_schema), Some(to_schema)) =
            (schema_map.get(from_script), schema_map.get(to_script))
        {
            if let Ok(converter_code) =
                generate_single_direct_converter(handlebars, from_schema, to_schema)
            {
                generated_code.push_str(&converter_code);
                generated_code.push('\n');
            }
        }
    }

    // Add registry function to access all direct converters
    generated_code.push_str(r#"
/// Registry of all direct converters for fast lookup
pub struct DirectConverterRegistry {
    converters: std::collections::HashMap<(String, String), Box<dyn DirectConverter>>,
}

impl DirectConverterRegistry {
    pub fn new() -> Self {
        let mut converters: std::collections::HashMap<(String, String), Box<dyn DirectConverter>> = std::collections::HashMap::new();
        
        // Register all generated direct converters
"#);

    // Register each converter
    for (from_script, to_script) in &conversion_pairs {
        if schema_map.contains_key(from_script) && schema_map.contains_key(to_script) {
            let struct_name = format!(
                "{}To{}Converter",
                capitalize_first(from_script),
                capitalize_first(to_script)
            );
            generated_code.push_str(&format!(
                "        converters.insert((\"{}\".to_string(), \"{}\".to_string()), Box::new({}::new()));\n",
                from_script, to_script, struct_name
            ));
        }
    }

    generated_code.push_str(
        r#"        
        Self { converters }
    }
    
    pub fn get_converter(&self, from: &str, to: &str) -> Option<&dyn DirectConverter> {
        self.converters.get(&(from.to_string(), to.to_string()))
            .map(|c| c.as_ref())
    }
}

pub trait DirectConverter: Send + Sync {
    fn convert(&self, input: &str) -> Result<String, ConverterError>;
    fn from_script(&self) -> &'static str;
    fn to_script(&self) -> &'static str;
}
"#,
    );

    Ok(generated_code)
}

/// Generate a single direct converter between two schemas
fn generate_single_direct_converter(
    handlebars: &Handlebars,
    from_schema: &ScriptSchema,
    to_schema: &ScriptSchema,
) -> Result<String, Box<dyn std::error::Error>> {
    let from_script = &from_schema.metadata.name;
    let to_script = &to_schema.metadata.name;

    // Build direct mapping by computing token-to-token conversion at build time
    let mut direct_mappings = Vec::new();

    // Get all mappings from both schemas
    let from_mappings = collect_all_mappings(from_schema);
    let to_mappings = collect_all_mappings(to_schema);

    // Create reverse mapping for target schema (string -> token)
    let mut to_reverse_map = FxHashMap::default();
    for (token, strings) in &to_mappings {
        for string in strings {
            to_reverse_map.insert(token.clone(), string.clone());
        }
    }

    // For each source pattern, find the corresponding target pattern
    for (token, from_strings) in &from_mappings {
        if let Some(to_string) = to_reverse_map.get(token) {
            for from_string in from_strings {
                direct_mappings.push(json!({
                    "from_pattern": from_string,
                    "to_pattern": to_string
                }));
            }
        }
    }

    // Sort by length (longest first) for proper matching
    direct_mappings.sort_by(|a, b| {
        let a_len = a["from_pattern"].as_str().unwrap().len();
        let b_len = b["from_pattern"].as_str().unwrap().len();
        b_len.cmp(&a_len)
    });

    let struct_name = format!(
        "{}To{}Converter",
        capitalize_first(from_script),
        capitalize_first(to_script)
    );

    let template_data = json!({
        "struct_name": struct_name,
        "from_script": from_script,
        "to_script": to_script,
        "direct_mappings": direct_mappings,
    });

    let mut converter_code = handlebars.render("direct_converter", &template_data)?;

    // Add trait implementation
    converter_code.push_str(&format!(
        r#"

impl DirectConverter for {} {{
    fn convert(&self, input: &str) -> Result<String, ConverterError> {{
        self.convert(input)
    }}
    
    fn from_script(&self) -> &'static str {{
        "{}"
    }}
    
    fn to_script(&self) -> &'static str {{
        "{}"
    }}
}}
"#,
        struct_name, from_script, to_script
    ));

    Ok(converter_code)
}

/// Collect all mappings from a schema (token -> [strings])
fn collect_all_mappings(schema: &ScriptSchema) -> FxHashMap<String, Vec<String>> {
    let mut mappings = FxHashMap::default();

    // Process each mapping category
    if let Some(ref vowels) = schema.mappings.vowels {
        for (token, mapping) in vowels {
            let strings = match mapping {
                TokenMapping::Single(s) => vec![s.clone()],
                TokenMapping::Multiple(v) => v.clone(),
            };
            mappings.insert(token.clone(), strings);
        }
    }

    if let Some(ref consonants) = schema.mappings.consonants {
        for (token, mapping) in consonants {
            let strings = match mapping {
                TokenMapping::Single(s) => vec![s.clone()],
                TokenMapping::Multiple(v) => v.clone(),
            };
            mappings.insert(token.clone(), strings);
        }
    }

    if let Some(ref marks) = schema.mappings.marks {
        for (token, mapping) in marks {
            let strings = match mapping {
                TokenMapping::Single(s) => vec![s.clone()],
                TokenMapping::Multiple(v) => v.clone(),
            };
            mappings.insert(token.clone(), strings);
        }
    }

    if let Some(ref digits) = schema.mappings.digits {
        for (token, mapping) in digits {
            let strings = match mapping {
                TokenMapping::Single(s) => vec![s.clone()],
                TokenMapping::Multiple(v) => v.clone(),
            };
            mappings.insert(token.clone(), strings);
        }
    }

    if let Some(ref special) = schema.mappings.special {
        for (token, mapping) in special {
            let strings = match mapping {
                TokenMapping::Single(s) => vec![s.clone()],
                TokenMapping::Multiple(v) => v.clone(),
            };
            mappings.insert(token.clone(), strings);
        }
    }

    if let Some(ref extended) = schema.mappings.extended {
        for (token, mapping) in extended {
            let strings = match mapping {
                TokenMapping::Single(s) => vec![s.clone()],
                TokenMapping::Multiple(v) => v.clone(),
            };
            mappings.insert(token.clone(), strings);
        }
    }

    mappings
}