symforge 8.1.0

SymForge — in-memory code intelligence MCP server for Claude Code, Codex, and Gemini CLI
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
//! Project conventions detection: analyzes the indexed codebase to infer
//! coding patterns, naming conventions, error handling style, test organization,
//! and file structure. Useful for LLMs writing code that fits the project.

use crate::live_index::store::LiveIndex;

/// Detected project conventions from static analysis of the index.
pub struct ProjectConventions {
    /// Dominant code language of the project (e.g. "Rust", "TypeScript"),
    /// with an optional note when a second language has a >25% share.
    pub language: String,
    pub error_handling: String,
    pub naming: String,
    pub test_patterns: String,
    pub common_imports: Vec<String>,
    pub file_organization: String,
    pub complexity: String,
}

/// Map an `IndexedFile.language` to a stable, human-readable code-language
/// bucket for the dominant-language vote, or `None` for config/markup files
/// that must be excluded from the vote.
///
/// JavaScript and TypeScript fold into a single `"TypeScript"` bucket so a
/// mixed `.js`/`.ts` codebase does not split its own vote. Config and markup
/// languages (Json/Toml/Yaml/Markdown/Env/Html/Css/Scss) are excluded by
/// `LanguageId` rather than by `FileClass` — every code path is classified
/// `FileClass::Code`, so a `FileClass`-based filter would let `.json` fixtures
/// win the vote.
fn code_language_bucket(language: &crate::domain::index::LanguageId) -> Option<&'static str> {
    use crate::domain::index::LanguageId;
    match language {
        LanguageId::Rust => Some("Rust"),
        LanguageId::Python => Some("Python"),
        // Fold JS + TS into one bucket so the vote is not split.
        LanguageId::JavaScript | LanguageId::TypeScript => Some("TypeScript"),
        LanguageId::Go => Some("Go"),
        LanguageId::Java => Some("Java"),
        LanguageId::C => Some("C"),
        LanguageId::Cpp => Some("C++"),
        LanguageId::CSharp => Some("C#"),
        LanguageId::Ruby => Some("Ruby"),
        LanguageId::Php => Some("PHP"),
        LanguageId::Swift => Some("Swift"),
        LanguageId::Kotlin => Some("Kotlin"),
        LanguageId::Dart => Some("Dart"),
        LanguageId::Perl => Some("Perl"),
        LanguageId::Elixir => Some("Elixir"),
        // Config / markup languages never vote on the dominant CODE language.
        LanguageId::Json
        | LanguageId::Toml
        | LanguageId::Yaml
        | LanguageId::Markdown
        | LanguageId::Env
        | LanguageId::Html
        | LanguageId::Css
        | LanguageId::Scss => None,
    }
}

/// Heuristic camelCase check for TS/JS function/method names: starts with a
/// lowercase ASCII letter and contains no underscores. Treats single-word
/// lowercase identifiers (e.g. `handle`) as camelCase, matching the JS/TS
/// convention. Names starting with `_` (private) or `$` are not counted.
fn is_camel_case(name: &str) -> bool {
    match name.chars().next() {
        Some(c) => c.is_ascii_lowercase() && !name.contains('_'),
        None => false,
    }
}

/// Analyze the index to detect project conventions.
///
/// The analysis is language-aware: it first votes on the project's dominant
/// code language by tallying `IndexedFile.language` over non-config code files,
/// then reports error-handling, naming, and test conventions framed for that
/// language. Without this, a TypeScript/NestJS project would be described in
/// Rust terms ("Result-based", "% snake_case", inline test modules) because the
/// substring heuristics fire on any `Result<T>` type and Rust-only symbol scans.
pub fn detect_conventions(index: &LiveIndex) -> ProjectConventions {
    use crate::domain::index::LanguageId;

    // ── Pass 1: dominant-language vote ───────────────────────────────────────
    // Tally over CODE files only, excluding config/markup by `LanguageId`
    // (NOT by `FileClass`, which is always `Code` — a `.json` fixture would
    // otherwise win the vote). JS and TS are folded into one bucket so a mixed
    // `.js`/`.ts` frontend does not split its own vote.
    let mut lang_votes: std::collections::HashMap<&'static str, u32> =
        std::collections::HashMap::new();
    // Per-member counts inside the folded JS/TS bucket, used ONLY to pick the
    // reported label (review finding 4, post-v7.19.0): the fold keeps a mixed
    // frontend from splitting its own vote, but a pure-JavaScript project must
    // not be LABELED "TypeScript".
    let mut js_file_votes = 0u32;
    let mut ts_file_votes = 0u32;
    for (_path, file) in index.all_files() {
        if let Some(bucket) = code_language_bucket(&file.language) {
            *lang_votes.entry(bucket).or_insert(0) += 1;
        }
        match file.language {
            LanguageId::JavaScript => js_file_votes += 1,
            LanguageId::TypeScript => ts_file_votes += 1,
            _ => {}
        }
    }
    let total_code_votes: u32 = lang_votes.values().copied().sum();
    let mut vote_vec: Vec<(&'static str, u32)> = lang_votes.into_iter().collect();
    // Deterministic ordering: by count desc, then name asc for tie-breaking.
    vote_vec.sort_by(|a, b| b.1.cmp(&a.1).then_with(|| a.0.cmp(b.0)));
    let primary_lang: &'static str = vote_vec.first().map(|(name, _)| *name).unwrap_or("Unknown");
    // Review finding 4 (post-v7.19.0): the folded JS/TS bucket keeps a mixed
    // frontend from splitting its own vote, but a pure-JavaScript project must
    // not be reported as "TypeScript". When the folded bucket wins with a
    // JavaScript-majority membership, relabel the primary to "JavaScript" —
    // every language-branched summary below matches "TypeScript" | "JavaScript",
    // so the summary framing is unchanged. Ties keep "TypeScript" (the
    // superset syntax).
    let primary_lang: &'static str =
        if primary_lang == "TypeScript" && js_file_votes > ts_file_votes {
            "JavaScript"
        } else {
            primary_lang
        };
    // Open decision (recorded): a mixed repo reports a SINGLE primary language
    // plus a one-line note when a SECOND language holds a >25% share.
    let secondary_note: Option<String> = vote_vec.get(1).and_then(|(name, count)| {
        if total_code_votes > 0 && (*count as u64) * 100 > (total_code_votes as u64) * 25 {
            let pct = (*count * 100).checked_div(total_code_votes).unwrap_or(0);
            Some(format!("{name} also {pct}%"))
        } else {
            None
        }
    });

    // ── Pass 2: per-file scan (each counter gated by THAT file's language) ────
    // Rust counters.
    let mut error_result_count = 0u32;
    let mut error_anyhow_count = 0u32;
    let mut error_thiserror_count = 0u32;
    let mut unwrap_count = 0u32;
    let mut expect_count = 0u32;
    let mut snake_case_fns = 0u32;
    let mut camel_case_types = 0u32;
    let mut rust_total_fns = 0u32;
    let mut rust_total_types = 0u32;
    let mut inline_test_mod_count = 0u32;
    let mut test_fn_count = 0u32;

    // TS/JS counters.
    let mut try_catch_count = 0u32;
    let mut throw_new_count = 0u32;
    let mut http_exception_count = 0u32;
    let mut catch_error_count = 0u32;
    let mut camel_case_fns = 0u32;
    let mut pascal_case_types = 0u32;
    let mut ts_total_fns = 0u32;
    let mut ts_total_types = 0u32;
    let mut describe_block_count = 0u32;
    let mut decorator_files = 0u32;
    let mut dto_validator_files = 0u32;
    let mut signal_files = 0u32;

    // Python counters.
    let mut py_try_except_count = 0u32;
    let mut py_raise_count = 0u32;
    let mut py_suppress_count = 0u32;

    // Go counters.
    let mut go_err_check_count = 0u32;
    let mut go_err_construct_count = 0u32;
    let mut go_panic_count = 0u32;

    // Kotlin counters.
    let mut kt_try_catch_count = 0u32;
    let mut kt_runcatching_count = 0u32;
    let mut kt_throws_count = 0u32;

    // Elixir counters.
    let mut ex_ok_error_tuple_count = 0u32;
    let mut ex_with_else_count = 0u32;
    let mut ex_raise_rescue_count = 0u32;

    // Language-agnostic naming tally, computed over files whose language matches
    // the dominant bucket. Powers a non-empty Naming line for EVERY language
    // (snake_case vs camelCase functions, PascalCase types) — see SF-STRESS-021.
    let mut primary_snake_fns = 0u32;
    let mut primary_camel_fns = 0u32;
    let mut primary_total_fns = 0u32;
    let mut primary_pascal_types = 0u32;
    let mut primary_total_types = 0u32;

    // Language-agnostic counters.
    let mut test_file_count = 0u32;

    let mut import_counts: std::collections::HashMap<String, u32> =
        std::collections::HashMap::new();

    let mut total_files = 0u32;
    let mut total_symbols = 0u32;
    let mut max_symbols_per_file = 0u32;
    let mut total_file_bytes = 0u64;
    let mut max_file_bytes = 0u64;
    let mut code_file_count = 0u32;

    for (_path, file) in index.all_files() {
        total_files += 1;
        total_symbols += file.symbols.len() as u32;
        total_file_bytes += file.byte_len;

        if file.symbols.len() as u32 > max_symbols_per_file {
            max_symbols_per_file = file.symbols.len() as u32;
        }
        if file.byte_len > max_file_bytes {
            max_file_bytes = file.byte_len;
        }

        let is_code = matches!(
            file.classification.class,
            crate::domain::index::FileClass::Code
        );
        if is_code {
            code_file_count += 1;
        }

        // Test detection (language-agnostic: `is_test` already covers Rust
        // `test_`/`_test` and TS/JS `*.spec.ts`/`*.test.ts` naming).
        if file.classification.is_test {
            test_file_count += 1;
        }

        let content_str = std::str::from_utf8(&file.content).unwrap_or("");

        let is_rust = file.language == LanguageId::Rust;
        let is_ts_js = matches!(
            file.language,
            LanguageId::TypeScript | LanguageId::JavaScript
        );
        let is_python = file.language == LanguageId::Python;
        let is_go = file.language == LanguageId::Go;
        let is_kotlin = file.language == LanguageId::Kotlin;
        let is_elixir = file.language == LanguageId::Elixir;
        // A file counts toward the language-agnostic naming tally when its
        // language matches the reported dominant bucket (JS/TS folded).
        let is_primary_lang_file = code_language_bucket(&file.language)
            .map(|bucket| {
                bucket == primary_lang || (primary_lang == "JavaScript" && bucket == "TypeScript")
            })
            .unwrap_or(false);

        // Error-handling patterns — gated by the FILE's language, not just the
        // summary branch, so non-Rust files never pollute the `Result`/`anyhow`
        // counts in a Rust-majority mixed repo (and vice versa).
        if is_rust {
            if content_str.contains("Result<") || content_str.contains("-> Result") {
                error_result_count += 1;
            }
            if content_str.contains("anyhow") {
                error_anyhow_count += 1;
            }
            if content_str.contains("thiserror") {
                error_thiserror_count += 1;
            }
            unwrap_count += content_str.matches(".unwrap()").count() as u32;
            expect_count += content_str.matches(".expect(").count() as u32;

            // Inline test modules + `test_`-prefixed fns are Rust-only signals.
            for sym in &file.symbols {
                if sym.name == "tests"
                    && matches!(sym.kind, crate::domain::index::SymbolKind::Module)
                {
                    inline_test_mod_count += 1;
                }
                if sym.name.starts_with("test_")
                    && matches!(sym.kind, crate::domain::index::SymbolKind::Function)
                {
                    test_fn_count += 1;
                }
            }
        } else if is_ts_js {
            if content_str.contains("try {") || content_str.contains("} catch") {
                try_catch_count += 1;
            }
            if content_str.contains("throw new") {
                throw_new_count += 1;
            }
            if content_str.contains("HttpException") {
                http_exception_count += 1;
            }
            if content_str.contains("catchError") {
                catch_error_count += 1;
            }

            // Test framework calls (`describe(`/`it(`/`test(`) — `SymbolRecord`
            // does not store these, so scan the source.
            if content_str.contains("describe(")
                || content_str.contains("it(")
                || content_str.contains("test(")
            {
                describe_block_count += 1;
            }

            // Decorators / DTO validators / signals — `SymbolRecord` does NOT
            // store decorators, so detect them via content scan.
            if content_str.contains("@Controller")
                || content_str.contains("@Injectable")
                || content_str.contains("@Module")
                || content_str.contains("@Component")
            {
                decorator_files += 1;
            }
            if content_str.contains("@IsString")
                || content_str.contains("@IsNumber")
                || content_str.contains("@IsNotEmpty")
            {
                dto_validator_files += 1;
            }
            if content_str.contains("signal(") || content_str.contains("inject(") {
                signal_files += 1;
            }
        } else if is_python {
            // Python: try/except, raise X, contextlib.suppress.
            if content_str.contains("try:")
                && (content_str.contains("except ") || content_str.contains("except:"))
            {
                py_try_except_count += 1;
            }
            if content_str.contains("raise ") {
                py_raise_count += 1;
            }
            if content_str.contains("suppress(") || content_str.contains("contextlib.suppress") {
                py_suppress_count += 1;
            }
        } else if is_go {
            // Go: `if err != nil` checks, error constructors, panic.
            go_err_check_count += content_str.matches("if err != nil").count() as u32;
            if content_str.contains("errors.New")
                || content_str.contains("fmt.Errorf")
                || content_str.contains("errors.Wrap")
            {
                go_err_construct_count += 1;
            }
            go_panic_count += content_str.matches("panic(").count() as u32;
        } else if is_kotlin {
            // Kotlin: try/catch, Result/runCatching, @Throws.
            if content_str.contains("try {") || content_str.contains("} catch") {
                kt_try_catch_count += 1;
            }
            if content_str.contains("runCatching") || content_str.contains("Result<") {
                kt_runcatching_count += 1;
            }
            if content_str.contains("@Throws") {
                kt_throws_count += 1;
            }
        } else if is_elixir {
            // Elixir: {:ok, _}/{:error, _} tuple returns, with/else, raise/rescue.
            if content_str.contains("{:ok,") || content_str.contains("{:error,") {
                ex_ok_error_tuple_count += 1;
            }
            if content_str.contains("with ") && content_str.contains("else") {
                ex_with_else_count += 1;
            }
            if content_str.contains("raise ") || content_str.contains("rescue") {
                ex_raise_rescue_count += 1;
            }
        }

        // Language-agnostic naming tally over dominant-language files: powers a
        // non-empty Naming line for EVERY language (not just Rust / TS-JS).
        if is_primary_lang_file {
            for sym in &file.symbols {
                match sym.kind {
                    crate::domain::index::SymbolKind::Function
                    | crate::domain::index::SymbolKind::Method => {
                        primary_total_fns += 1;
                        let name = &sym.name;
                        // snake_case wins when it contains `_` and is all-lower;
                        // otherwise a lowercase-initial name with no `_` is camelCase.
                        if name.contains('_') && *name == name.to_ascii_lowercase() {
                            primary_snake_fns += 1;
                        } else if name.chars().next().is_some_and(|c| c.is_ascii_lowercase())
                            && !name.contains('_')
                        {
                            primary_camel_fns += 1;
                        }
                    }
                    crate::domain::index::SymbolKind::Struct
                    | crate::domain::index::SymbolKind::Class
                    | crate::domain::index::SymbolKind::Enum
                    | crate::domain::index::SymbolKind::Trait
                    | crate::domain::index::SymbolKind::Interface
                    | crate::domain::index::SymbolKind::Type => {
                        primary_total_types += 1;
                        if sym.name.chars().next().is_some_and(|c| c.is_uppercase()) {
                            primary_pascal_types += 1;
                        }
                    }
                    _ => {}
                }
            }
        }

        // Naming conventions — gated by the file's language so the Rust
        // snake_case tally and the TS/JS camelCase tally never cross-pollute.
        for sym in &file.symbols {
            match sym.kind {
                crate::domain::index::SymbolKind::Function
                | crate::domain::index::SymbolKind::Method => {
                    if is_rust {
                        rust_total_fns += 1;
                        if sym.name.contains('_') && sym.name == sym.name.to_ascii_lowercase() {
                            snake_case_fns += 1;
                        }
                    } else if is_ts_js {
                        ts_total_fns += 1;
                        if is_camel_case(&sym.name) {
                            camel_case_fns += 1;
                        }
                    }
                }
                crate::domain::index::SymbolKind::Struct
                | crate::domain::index::SymbolKind::Class
                | crate::domain::index::SymbolKind::Enum
                | crate::domain::index::SymbolKind::Trait
                | crate::domain::index::SymbolKind::Interface
                | crate::domain::index::SymbolKind::Type => {
                    if is_rust {
                        rust_total_types += 1;
                        if sym.name.chars().next().is_some_and(|c| c.is_uppercase()) {
                            camel_case_types += 1;
                        }
                    } else if is_ts_js {
                        ts_total_types += 1;
                        if sym.name.chars().next().is_some_and(|c| c.is_uppercase()) {
                            pascal_case_types += 1;
                        }
                    }
                }
                _ => {}
            }
        }

        // Import tracking (language-agnostic).
        for reference in &file.references {
            if matches!(reference.kind, crate::domain::index::ReferenceKind::Import) {
                let import_name = reference
                    .qualified_name
                    .as_deref()
                    .unwrap_or(&reference.name);
                // Extract the crate/module root
                let root = import_name.split("::").next().unwrap_or(import_name);
                if !root.is_empty() && root.len() > 1 {
                    *import_counts.entry(root.to_string()).or_insert(0) += 1;
                }
            }
        }
    }

    // ── Language header ──────────────────────────────────────────────────────
    let language = match &secondary_note {
        Some(note) => format!("{primary_lang} (primary; {note})"),
        None => primary_lang.to_string(),
    };

    // ── Error handling summary (language-branched) ───────────────────────────
    let error_handling = match primary_lang {
        "Rust" => {
            if error_anyhow_count > 2 && error_thiserror_count > 2 {
                format!(
                    "Mixed: anyhow ({error_anyhow_count} files) + thiserror ({error_thiserror_count} files), Result<> in {error_result_count} files, {unwrap_count} unwrap()s, {expect_count} expect()s"
                )
            } else if error_anyhow_count > 2 {
                format!(
                    "anyhow-based: {error_anyhow_count} files use anyhow, Result<> in {error_result_count} files, {unwrap_count} unwrap()s, {expect_count} expect()s"
                )
            } else if error_thiserror_count > 2 {
                format!(
                    "thiserror-based: {error_thiserror_count} files use thiserror, Result<> in {error_result_count} files, {unwrap_count} unwrap()s, {expect_count} expect()s"
                )
            } else if error_result_count > 0 {
                format!(
                    "Result-based: {error_result_count} files return Result, {unwrap_count} unwrap()s, {expect_count} expect()s"
                )
            } else {
                format!(
                    "Minimal error handling detected. {unwrap_count} unwrap()s, {expect_count} expect()s"
                )
            }
        }
        "TypeScript" | "JavaScript" => {
            if try_catch_count == 0
                && throw_new_count == 0
                && http_exception_count == 0
                && catch_error_count == 0
            {
                "Minimal explicit error handling detected (no try/catch or throw found)."
                    .to_string()
            } else {
                let mut parts: Vec<String> = Vec::new();
                if try_catch_count > 0 {
                    parts.push(format!("try/catch in {try_catch_count} files"));
                }
                if throw_new_count > 0 {
                    parts.push(format!("`throw new` in {throw_new_count} files"));
                }
                if http_exception_count > 0 {
                    parts.push(format!(
                        "NestJS HttpException in {http_exception_count} files"
                    ));
                }
                if catch_error_count > 0 {
                    parts.push(format!("RxJS catchError in {catch_error_count} files"));
                }
                format!("Exception-based: {}", parts.join(", "))
            }
        }
        "Python" => {
            if py_try_except_count == 0 && py_raise_count == 0 && py_suppress_count == 0 {
                "Minimal explicit error handling detected (no try/except or raise found)."
                    .to_string()
            } else {
                let mut parts: Vec<String> = Vec::new();
                if py_try_except_count > 0 {
                    parts.push(format!("try/except in {py_try_except_count} files"));
                }
                if py_raise_count > 0 {
                    parts.push(format!("`raise` in {py_raise_count} files"));
                }
                if py_suppress_count > 0 {
                    parts.push(format!("contextlib.suppress in {py_suppress_count} files"));
                }
                format!("Exception-based: {}", parts.join(", "))
            }
        }
        "Go" => {
            if go_err_check_count == 0 && go_err_construct_count == 0 && go_panic_count == 0 {
                "Minimal explicit error handling detected (no `if err != nil` found).".to_string()
            } else {
                format!(
                    "Value-based: {go_err_check_count} `if err != nil` checks, errors.New/fmt.Errorf in {go_err_construct_count} files, {go_panic_count} panic()s"
                )
            }
        }
        "Kotlin" => {
            if kt_try_catch_count == 0 && kt_runcatching_count == 0 && kt_throws_count == 0 {
                "Minimal explicit error handling detected (no try/catch or runCatching found)."
                    .to_string()
            } else {
                let mut parts: Vec<String> = Vec::new();
                if kt_try_catch_count > 0 {
                    parts.push(format!("try/catch in {kt_try_catch_count} files"));
                }
                if kt_runcatching_count > 0 {
                    parts.push(format!(
                        "Result/runCatching in {kt_runcatching_count} files"
                    ));
                }
                if kt_throws_count > 0 {
                    parts.push(format!("@Throws in {kt_throws_count} files"));
                }
                format!("Exception-based: {}", parts.join(", "))
            }
        }
        "Elixir" => {
            if ex_ok_error_tuple_count == 0 && ex_with_else_count == 0 && ex_raise_rescue_count == 0
            {
                "Minimal explicit error handling detected (no {:ok,_}/{:error,_} tuples found)."
                    .to_string()
            } else {
                let mut parts: Vec<String> = Vec::new();
                if ex_ok_error_tuple_count > 0 {
                    parts.push(format!(
                        "{{:ok,_}}/{{:error,_}} tuples in {ex_ok_error_tuple_count} files"
                    ));
                }
                if ex_with_else_count > 0 {
                    parts.push(format!("with/else in {ex_with_else_count} files"));
                }
                if ex_raise_rescue_count > 0 {
                    parts.push(format!("raise/rescue in {ex_raise_rescue_count} files"));
                }
                format!("Tuple-based: {}", parts.join(", "))
            }
        }
        // No language-specific heuristics yet for this dominant language. The
        // label is prepended by `format_conventions`, so do NOT embed it here
        // (that produced doubled "Error handling: Error handling:" output).
        _ => "language-specific heuristics not yet implemented for this dominant language."
            .to_string(),
    };

    // ── Naming summary (language-branched) ───────────────────────────────────
    let naming = match primary_lang {
        "Rust" => {
            let fn_pct = (snake_case_fns * 100)
                .checked_div(rust_total_fns)
                .unwrap_or(0);
            let type_pct = (camel_case_types * 100)
                .checked_div(rust_total_types)
                .unwrap_or(0);
            format!(
                "Functions: {fn_pct}% snake_case ({snake_case_fns}/{rust_total_fns}). Types: {type_pct}% CamelCase ({camel_case_types}/{rust_total_types})."
            )
        }
        "TypeScript" | "JavaScript" => {
            let fn_pct = (camel_case_fns * 100)
                .checked_div(ts_total_fns)
                .unwrap_or(0);
            let type_pct = (pascal_case_types * 100)
                .checked_div(ts_total_types)
                .unwrap_or(0);
            format!(
                "Functions: {fn_pct}% camelCase ({camel_case_fns}/{ts_total_fns}). Types: {type_pct}% PascalCase ({pascal_case_types}/{ts_total_types})."
            )
        }
        // Language-agnostic fallback for every other dominant language: report
        // the observed snake_case/camelCase function ratio and PascalCase type
        // ratio from the index instead of an empty "unavailable" line. The label
        // is prepended by `format_conventions`, so it is NOT embedded here.
        _ => {
            if primary_total_fns == 0 && primary_total_types == 0 {
                "no functions or types indexed for the dominant language.".to_string()
            } else {
                let snake_pct = (primary_snake_fns * 100)
                    .checked_div(primary_total_fns)
                    .unwrap_or(0);
                let camel_pct = (primary_camel_fns * 100)
                    .checked_div(primary_total_fns)
                    .unwrap_or(0);
                let type_pct = (primary_pascal_types * 100)
                    .checked_div(primary_total_types)
                    .unwrap_or(0);
                format!(
                    "Functions: {snake_pct}% snake_case, {camel_pct}% camelCase ({primary_total_fns} total). Types: {type_pct}% PascalCase ({primary_pascal_types}/{primary_total_types})."
                )
            }
        }
    };

    // ── Test patterns (language-branched; `test_file_count` is shared) ───────
    let test_patterns = match primary_lang {
        "Rust" => format!(
            "{test_file_count} test files, {inline_test_mod_count} inline test modules, {test_fn_count} test functions"
        ),
        "TypeScript" | "JavaScript" => {
            let mut extras: Vec<String> = Vec::new();
            extras.push(format!(
                "{describe_block_count} files with describe/it/test"
            ));
            if decorator_files > 0 {
                extras.push(format!("{decorator_files} files with decorators"));
            }
            if dto_validator_files > 0 {
                extras.push(format!(
                    "{dto_validator_files} files with class-validator DTOs"
                ));
            }
            if signal_files > 0 {
                extras.push(format!("{signal_files} files with signals/inject"));
            }
            let framework = if describe_block_count > 0 {
                " (Jest/Mocha-style)"
            } else {
                ""
            };
            format!(
                "{test_file_count} test files{framework}, {extras}",
                extras = extras.join(", ")
            )
        }
        _ => format!("{test_file_count} test files"),
    };

    // Common imports (top 10)
    let mut import_vec: Vec<(String, u32)> = import_counts.into_iter().collect();
    import_vec.sort_by_key(|(_, count)| std::cmp::Reverse(*count));
    let common_imports: Vec<String> = import_vec
        .iter()
        .take(10)
        .map(|(name, count)| format!("{name} ({count} uses)"))
        .collect();

    // File organization (language-agnostic)
    let avg_symbols = total_symbols.checked_div(code_file_count).unwrap_or(0);
    let avg_size = if code_file_count > 0 {
        total_file_bytes / code_file_count as u64
    } else {
        0
    };
    let file_organization = format!(
        "{total_files} files ({code_file_count} code), avg {avg_symbols} symbols/file, avg {avg_kb}KB/file, largest {max_kb}KB ({max_symbols_per_file} symbols)",
        avg_kb = avg_size / 1024,
        max_kb = max_file_bytes / 1024,
    );

    // Complexity (language-agnostic)
    let complexity = if max_symbols_per_file > 200 {
        format!("High: largest file has {max_symbols_per_file} symbols")
    } else if max_symbols_per_file > 100 {
        format!("Medium: largest file has {max_symbols_per_file} symbols")
    } else {
        format!("Low: largest file has {max_symbols_per_file} symbols")
    };

    ProjectConventions {
        language,
        error_handling,
        naming,
        test_patterns,
        common_imports,
        file_organization,
        complexity,
    }
}

/// Extract the top N import root names from the index (cheap — single pass, no formatting).
/// Returns lowercase crate/module roots like `["serde", "tokio", "anyhow"]`.
///
/// Sources (unioned):
/// 1. `use`/`import` references extracted by tree-sitter (covers explicit imports).
/// 2. Manifest dependencies from `Cargo.toml` / `package.json` (covers crates used
///    only via derive macros, path-qualified syntax, or re-exports that tree-sitter
///    does not classify as imports — e.g. `thiserror`).
pub fn extract_top_import_roots(index: &LiveIndex, limit: usize) -> Vec<String> {
    let mut counts: std::collections::HashMap<String, u32> = std::collections::HashMap::new();

    // Source 1: explicit import references from the tree-sitter index.
    for (_path, file) in index.all_files() {
        for reference in &file.references {
            if matches!(reference.kind, crate::domain::index::ReferenceKind::Import) {
                let import_name = reference
                    .qualified_name
                    .as_deref()
                    .unwrap_or(&reference.name);
                let root = import_name.split("::").next().unwrap_or(import_name);
                if root.len() > 1 {
                    *counts.entry(root.to_ascii_lowercase()).or_insert(0) += 1;
                }
            }
        }
    }

    // Source 2: manifest dependency names (Cargo.toml / package.json).
    // These ensure crates used only via derive macros or path-qualified syntax
    // (e.g. `#[derive(thiserror::Error)]`) still appear in the import list.
    for (path, file) in index.all_files() {
        let is_cargo = path.ends_with("Cargo.toml");
        let is_package_json = path.ends_with("package.json");
        if !is_cargo && !is_package_json {
            continue;
        }
        for sym in &file.symbols {
            // Cargo.toml: "dependencies.thiserror", "dev-dependencies.once_cell"
            // package.json: "dependencies.express", "devDependencies.lodash"
            let dep_name = if is_cargo {
                sym.name
                    .strip_prefix("dependencies.")
                    .or_else(|| sym.name.strip_prefix("dev-dependencies."))
            } else {
                sym.name
                    .strip_prefix("dependencies.")
                    .or_else(|| sym.name.strip_prefix("devDependencies."))
            };
            if let Some(raw) = dep_name {
                // Skip nested sub-keys like "dependencies.serde.version".
                if raw.contains('.') {
                    continue;
                }
                // Normalize: Cargo crate hyphens → underscores (e.g. tree-sitter → tree_sitter).
                let normalized = if is_cargo {
                    raw.replace('-', "_").to_ascii_lowercase()
                } else {
                    raw.to_ascii_lowercase()
                };
                if normalized.len() > 1 {
                    counts.entry(normalized).or_insert(1);
                }
            }
        }
    }

    let mut vec: Vec<(String, u32)> = counts.into_iter().collect();
    vec.sort_by_key(|(_, count)| std::cmp::Reverse(*count));
    vec.into_iter().take(limit).map(|(name, _)| name).collect()
}

/// Format conventions for display.
pub fn format_conventions(conv: &ProjectConventions) -> String {
    let mut lines = vec![
        "── Project Conventions ──".to_string(),
        String::new(),
        format!("Language: {}", conv.language),
        format!("Error handling: {}", conv.error_handling),
        format!("Naming: {}", conv.naming),
        format!("Tests: {}", conv.test_patterns),
        format!("File organization: {}", conv.file_organization),
        format!("Complexity: {}", conv.complexity),
    ];

    if !conv.common_imports.is_empty() {
        lines.push(String::new());
        lines.push("Common imports:".to_string());
        for import in &conv.common_imports {
            lines.push(format!("  {import}"));
        }
    }

    lines.join("\n")
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::domain::index::{
        FileClassification, LanguageId, ReferenceKind, ReferenceRecord, SymbolKind, SymbolRecord,
    };
    use crate::live_index::store::IndexedFile;

    /// Build a `SymbolRecord` with the given name and kind. Byte/line ranges are
    /// placeholders — `detect_conventions` only reads `name` and `kind`.
    fn sym(name: &str, kind: SymbolKind) -> SymbolRecord {
        SymbolRecord {
            name: name.to_string(),
            kind,
            depth: 0,
            sort_order: 0,
            byte_range: (0, 0),
            line_range: (0, 0),
            doc_byte_range: None,
            item_byte_range: None,
        }
    }

    fn import_ref(name: &str) -> ReferenceRecord {
        ReferenceRecord {
            name: name.to_string(),
            qualified_name: None,
            kind: ReferenceKind::Import,
            byte_range: (0, 0),
            line_range: (0, 0),
            enclosing_symbol_index: None,
        }
    }

    /// Build an `IndexedFile` with the language wired EXPLICITLY (per verifier:
    /// not left defaulted) so a test that asserts language-aware behavior cannot
    /// pass for the wrong reason.
    fn make_file(
        path: &str,
        language: LanguageId,
        content: &str,
        symbols: Vec<SymbolRecord>,
        references: Vec<ReferenceRecord>,
    ) -> IndexedFile {
        let bytes = content.as_bytes().to_vec();
        let byte_len = bytes.len() as u64;
        IndexedFile {
            relative_path: path.to_string(),
            language,
            classification: FileClassification::for_code_path(path),
            content: bytes,
            symbols,
            parse_status: crate::live_index::store::ParseStatus::Parsed,
            parse_diagnostic: None,
            byte_len,
            content_hash: "hash".to_string(),
            references,
            alias_map: std::collections::HashMap::new(),
            mtime_secs: 0,
        }
    }

    /// Assemble a `LiveIndex` from fixture files via the public shared-index
    /// API and run `detect_conventions`.
    fn conventions_for(files: Vec<(&str, IndexedFile)>) -> ProjectConventions {
        let shared = LiveIndex::empty();
        {
            let mut guard = shared.write();
            for (path, file) in files {
                guard.add_file(path.to_string(), file);
            }
        }
        let guard = shared.read();
        detect_conventions(&guard)
    }

    #[test]
    fn rust_majority_project_reports_result_snake_case_and_test_modules() {
        let lib = make_file(
            "src/lib.rs",
            LanguageId::Rust,
            "use anyhow::Result;\nfn do_thing() -> Result<()> { foo().unwrap(); Ok(()) }\n",
            vec![
                sym("do_thing", SymbolKind::Function),
                sym("MyStruct", SymbolKind::Struct),
            ],
            vec![import_ref("anyhow")],
        );
        let store = make_file(
            "src/store.rs",
            LanguageId::Rust,
            "fn load_store() -> Result<u8> { Ok(0) }\n",
            vec![sym("load_store", SymbolKind::Function)],
            vec![],
        );
        // A Rust test file with an inline `tests` module + a `test_` fn.
        let test_file = make_file(
            "src/store_tests.rs",
            LanguageId::Rust,
            "#[cfg(test)] mod tests { fn test_load() {} }\n",
            vec![
                sym("tests", SymbolKind::Module),
                sym("test_load", SymbolKind::Function),
            ],
            vec![],
        );
        // Config noise must NOT win the vote or pollute the Rust pass.
        let pkg = make_file(
            "config.json",
            LanguageId::Json,
            "{ \"name\": \"x\", \"result\": \"Result<T>\" }\n",
            vec![],
            vec![],
        );

        let conv = conventions_for(vec![
            ("src/lib.rs", lib),
            ("src/store.rs", store),
            ("src/store_tests.rs", test_file),
            ("config.json", pkg),
        ]);

        assert_eq!(conv.language, "Rust", "Rust must win the dominant vote");
        assert!(
            conv.error_handling.contains("Result"),
            "Rust error handling should mention Result, got: {}",
            conv.error_handling
        );
        assert!(
            conv.naming.contains("snake_case"),
            "Rust naming should mention snake_case, got: {}",
            conv.naming
        );
        assert!(
            conv.test_patterns.contains("inline test module")
                && !conv.test_patterns.contains("0 inline test modules"),
            "Rust test patterns should report nonzero inline test modules, got: {}",
            conv.test_patterns
        );
    }

    #[test]
    fn typescript_majority_project_reports_exceptions_camelcase_and_no_result_wording() {
        // NestJS controller with decorators, exceptions, and a `Result<T>` type
        // that must NOT trigger Rust "Result-based" wording.
        let controller = make_file(
            "src/users/users.controller.ts",
            LanguageId::TypeScript,
            "@Controller('users')\nexport class UsersController {\n  getUser() {\n    if (!ok) { throw new HttpException('no', 404); }\n  }\n}\ninterface Result<T> { ok: boolean; value: T; }\n",
            vec![
                sym("UsersController", SymbolKind::Class),
                sym("getUser", SymbolKind::Method),
                sym("Result", SymbolKind::Interface),
            ],
            vec![import_ref("@nestjs/common")],
        );
        let service = make_file(
            "src/users/users.service.ts",
            LanguageId::TypeScript,
            "@Injectable()\nexport class UsersService {\n  findAll() {\n    try { return this.repo.find(); } catch (e) { throw new Error('x'); }\n  }\n}\n",
            vec![
                sym("UsersService", SymbolKind::Class),
                sym("findAll", SymbolKind::Method),
            ],
            vec![],
        );
        let dto = make_file(
            "src/users/create-user.dto.ts",
            LanguageId::TypeScript,
            "export class CreateUserDto {\n  @IsString()\n  name: string;\n}\n",
            vec![sym("CreateUserDto", SymbolKind::Class)],
            vec![],
        );
        // A spec test with describe(/it(.
        let spec = make_file(
            "src/users/users.controller.spec.ts",
            LanguageId::TypeScript,
            "describe('UsersController', () => {\n  it('returns a user', () => { expect(1).toBe(1); });\n});\n",
            vec![],
            vec![],
        );
        // Config noise must NOT win the vote.
        let pkg = make_file(
            "package.json",
            LanguageId::Json,
            "{ \"name\": \"app\" }\n",
            vec![],
            vec![],
        );

        let conv = conventions_for(vec![
            ("src/users/users.controller.ts", controller),
            ("src/users/users.service.ts", service),
            ("src/users/create-user.dto.ts", dto),
            ("src/users/users.controller.spec.ts", spec),
            ("package.json", pkg),
        ]);

        assert_eq!(
            conv.language, "TypeScript",
            "TypeScript must win the dominant vote"
        );
        assert!(
            !conv.error_handling.contains("Result-based"),
            "TS error handling must not say Result-based, got: {}",
            conv.error_handling
        );
        assert!(
            conv.error_handling.contains("Exception-based")
                || conv.error_handling.to_lowercase().contains("exception")
                || conv.error_handling.contains("throw new"),
            "TS error handling should mention exceptions/throw, got: {}",
            conv.error_handling
        );
        assert!(
            conv.naming.contains("camelCase"),
            "TS naming should mention camelCase, got: {}",
            conv.naming
        );
        assert!(
            !conv.test_patterns.starts_with("0 test files"),
            "TS test count should be nonzero, got: {}",
            conv.test_patterns
        );
        assert!(
            conv.test_patterns.contains("describe/it/test"),
            "TS test patterns should mention describe/it/test scan, got: {}",
            conv.test_patterns
        );
    }

    /// Review finding 4 (post-v7.19.0): a pure-JavaScript project folds into
    /// the JS/TS vote bucket (so the vote is not split) but must be REPORTED
    /// as "JavaScript", not "TypeScript" — while keeping the TS/JS-framed
    /// summaries (exception-based error handling, camelCase), never Rust
    /// wording.
    #[test]
    fn javascript_majority_project_is_labeled_javascript() {
        let service = make_file(
            "src/service.js",
            LanguageId::JavaScript,
            "function handleRequest(req) {\n  try { run(req); } catch (err) { throw new Error('boom'); }\n}\n",
            vec![sym("handleRequest", SymbolKind::Function)],
            vec![],
        );
        let util = make_file(
            "src/util.js",
            LanguageId::JavaScript,
            "function buildPayload() { return {}; }\n",
            vec![sym("buildPayload", SymbolKind::Function)],
            vec![],
        );
        // One stray .ts file: JS still holds the bucket majority (2 > 1).
        let typed = make_file(
            "src/types.ts",
            LanguageId::TypeScript,
            "export interface Payload { ok: boolean }\n",
            vec![sym("Payload", SymbolKind::Interface)],
            vec![],
        );

        let conv = conventions_for(vec![
            ("src/service.js", service),
            ("src/util.js", util),
            ("src/types.ts", typed),
        ]);

        assert!(
            conv.language.starts_with("JavaScript"),
            "JS-majority project must be labeled JavaScript, got: {}",
            conv.language
        );
        assert!(
            conv.error_handling.contains("try/catch") || conv.error_handling.contains("throw new"),
            "JS project keeps the exception-based summary framing, got: {}",
            conv.error_handling
        );
        assert!(
            !conv.error_handling.contains("unwrap"),
            "JS project must not get Rust error-handling wording, got: {}",
            conv.error_handling
        );
        assert!(
            conv.naming.contains("camelCase"),
            "JS naming keeps the camelCase framing, got: {}",
            conv.naming
        );
    }

    #[test]
    fn mixed_repo_records_secondary_language_note_over_25pct() {
        // 3 Rust + 2 TS code files -> TS share 40% (>25%) -> primary Rust + note.
        let files = vec![
            (
                "a.rs",
                make_file("a.rs", LanguageId::Rust, "fn a() {}\n", vec![], vec![]),
            ),
            (
                "b.rs",
                make_file("b.rs", LanguageId::Rust, "fn b() {}\n", vec![], vec![]),
            ),
            (
                "c.rs",
                make_file("c.rs", LanguageId::Rust, "fn c() {}\n", vec![], vec![]),
            ),
            (
                "d.ts",
                make_file(
                    "d.ts",
                    LanguageId::TypeScript,
                    "function d() {}\n",
                    vec![],
                    vec![],
                ),
            ),
            (
                "e.ts",
                make_file(
                    "e.ts",
                    LanguageId::TypeScript,
                    "function e() {}\n",
                    vec![],
                    vec![],
                ),
            ),
        ];
        let conv = conventions_for(files);
        assert!(
            conv.language.starts_with("Rust"),
            "primary language should be Rust, got: {}",
            conv.language
        );
        assert!(
            conv.language.contains("TypeScript") && conv.language.contains('%'),
            "secondary language note (>25%) should mention TypeScript with a percentage, got: {}",
            conv.language
        );
    }

    #[test]
    fn js_and_ts_fold_into_one_bucket() {
        // 2 JS + 2 TS would split 2/2 if not folded; folded they total 4 and win
        // over a single Rust file.
        let files = vec![
            (
                "x.rs",
                make_file("x.rs", LanguageId::Rust, "fn x() {}\n", vec![], vec![]),
            ),
            (
                "a.js",
                make_file(
                    "a.js",
                    LanguageId::JavaScript,
                    "function a() {}\n",
                    vec![],
                    vec![],
                ),
            ),
            (
                "b.js",
                make_file(
                    "b.js",
                    LanguageId::JavaScript,
                    "function b() {}\n",
                    vec![],
                    vec![],
                ),
            ),
            (
                "c.ts",
                make_file(
                    "c.ts",
                    LanguageId::TypeScript,
                    "function c() {}\n",
                    vec![],
                    vec![],
                ),
            ),
            (
                "d.ts",
                make_file(
                    "d.ts",
                    LanguageId::TypeScript,
                    "function d() {}\n",
                    vec![],
                    vec![],
                ),
            ),
        ];
        let conv = conventions_for(files);
        assert_eq!(
            conv.language, "TypeScript",
            "folded JS+TS (4) should beat Rust (1); got: {}",
            conv.language
        );
    }

    #[test]
    fn config_only_index_does_not_panic_and_reports_unknown() {
        let files = vec![
            (
                "a.json",
                make_file("a.json", LanguageId::Json, "{}\n", vec![], vec![]),
            ),
            (
                "b.toml",
                make_file("b.toml", LanguageId::Toml, "x = 1\n", vec![], vec![]),
            ),
        ];
        let conv = conventions_for(files);
        assert_eq!(
            conv.language, "Unknown",
            "config-only index should report Unknown dominant language"
        );
        // Generic branch must not use Rust-specific wording.
        assert!(!conv.error_handling.contains("Result-based"));
        assert!(!conv.naming.contains("snake_case"));
    }

    #[test]
    fn format_conventions_includes_language_header() {
        let conv = ProjectConventions {
            language: "TypeScript".to_string(),
            error_handling: "Exception-based: throw new in 2 files".to_string(),
            naming: "Functions: 100% camelCase (4/4).".to_string(),
            test_patterns: "3 test files".to_string(),
            common_imports: vec![],
            file_organization: "5 files".to_string(),
            complexity: "Low".to_string(),
        };
        let out = format_conventions(&conv);
        assert!(
            out.contains("Language: TypeScript"),
            "format output should include a Language header, got:\n{out}"
        );
    }

    /// SF-STRESS-021 regression: a Python-dominant repo must produce a non-empty
    /// error-handling AND naming headline (previously both fell to the empty
    /// "heuristics unavailable" fallback), and the rendered output must not
    /// double the field labels.
    #[test]
    fn python_majority_project_has_nonempty_headlines_and_no_doubled_labels() {
        let app = make_file(
            "app.py",
            LanguageId::Python,
            "import os\n\n\ndef do_thing():\n    try:\n        risky()\n    except ValueError:\n        raise RuntimeError('boom')\n",
            vec![
                sym("do_thing", SymbolKind::Function),
                sym("MyModel", SymbolKind::Class),
            ],
            vec![],
        );
        let util = make_file(
            "util.py",
            LanguageId::Python,
            "def load_config():\n    raise NotImplementedError\n",
            vec![sym("load_config", SymbolKind::Function)],
            vec![],
        );

        let conv = conventions_for(vec![("app.py", app), ("util.py", util)]);

        assert_eq!(conv.language, "Python", "Python must win the dominant vote");
        assert!(
            conv.error_handling.contains("try/except") || conv.error_handling.contains("raise"),
            "Python error handling must be populated, got: {}",
            conv.error_handling
        );
        assert!(
            !conv.error_handling.contains("heuristics unavailable"),
            "Python must not fall to the empty fallback, got: {}",
            conv.error_handling
        );
        assert!(
            conv.naming.contains("snake_case") || conv.naming.contains("PascalCase"),
            "Python naming must be populated from the language-agnostic tally, got: {}",
            conv.naming
        );

        // No doubled labels in the rendered output (the fallback strings must not
        // embed the label that `format_conventions` already prepends).
        let out = format_conventions(&conv);
        assert!(
            !out.contains("Error handling: Error handling:"),
            "error-handling label must not be doubled, got:\n{out}"
        );
        assert!(
            !out.contains("Naming: Naming:"),
            "naming label must not be doubled, got:\n{out}"
        );
    }

    /// A Go-dominant repo reports value-based error handling (the highest-corpus
    /// presence language), proving the new per-language heuristic fires.
    #[test]
    fn go_majority_project_reports_value_based_error_handling() {
        let main = make_file(
            "main.go",
            LanguageId::Go,
            "package main\n\nfunc run() error {\n    if err != nil {\n        return fmt.Errorf(\"wrap: %w\", err)\n    }\n    return nil\n}\n",
            vec![
                sym("run", SymbolKind::Function),
                sym("Server", SymbolKind::Struct),
            ],
            vec![],
        );

        let conv = conventions_for(vec![("main.go", main)]);
        assert_eq!(conv.language, "Go");
        assert!(
            conv.error_handling.contains("if err != nil")
                || conv.error_handling.contains("Value-based"),
            "Go error handling must mention value-based err checks, got: {}",
            conv.error_handling
        );
        assert!(
            !conv.naming.contains("heuristics unavailable"),
            "Go naming must be populated, got: {}",
            conv.naming
        );
    }
}