Skip to content
Navigation Menu
{{ message }}
forked from ggml-org/llama.cpp
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathtest-grammar-integration.cpp
More file actions
1493 lines (1398 loc) · 41.4 KB
/
Copy pathtest-grammar-integration.cpp
File metadata and controls
1493 lines (1398 loc) · 41.4 KB
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
#ifdef NDEBUG
#undef NDEBUG
#endif
#include "json-schema-to-grammar.h"
#include "../src/unicode.h"
#include "../src/llama-grammar.h"
#include <nlohmann/json.hpp>
#include <cassert>
#include <string>
#include <vector>
using json = nlohmann::ordered_json;
static llama_grammar * build_grammar_with_root(const std::string & grammar_str, const char * grammar_root) {
return llama_grammar_init_impl(nullptr, grammar_str.c_str(), grammar_root, false, nullptr, 0, nullptr, 0);
}
static llama_grammar * build_grammar(const std::string & grammar_str) {
return build_grammar_with_root(grammar_str, "root");
}
static bool test_build_grammar_fails(const std::string & grammar_str) {
fprintf(stderr, "⚫ Testing failure for grammar: %s\n", grammar_str.c_str());
bool grammar_fails = false;
llama_grammar * grammar = build_grammar(grammar_str);
if (grammar != nullptr) {
fprintf(stderr, " ❌ Expected build failure, but succeeded\n");
} else {
grammar_fails = true;
fprintf(stdout, " ✅︎\n");
}
return grammar_fails;
}
struct token_and_piece {
llama_token token;
std::string piece;
};
// token() encodes a 32-bit ID as 5 bytes: a 0xff marker followed by the ID in big-endian order.
static std::string token(llama_token id) {
return std::string{
static_cast<char>(0xff),
static_cast<char>((id >> 24) & 0xff),
static_cast<char>((id >> 16) & 0xff),
static_cast<char>((id >> 8) & 0xff),
static_cast<char>(id & 0xff)
};
}
// parse_tokens() parses the token encodes above and UTF-8 text.
static std::vector<token_and_piece> parse_tokens(const std::string & input) {
std::vector<token_and_piece> result;
result.reserve(input.size());
size_t offset = 0;
while (offset < input.size()) {
try {
if (static_cast<unsigned char>(input[offset]) == 0xff) {
if (offset + 5 > input.size()) {
throw std::runtime_error("not enough bytes for token id");
}
uint32_t val =
(static_cast<unsigned char>(input[offset + 1]) << 24) |
(static_cast<unsigned char>(input[offset + 2]) << 16) |
(static_cast<unsigned char>(input[offset + 3]) << 8) |
(static_cast<unsigned char>(input[offset + 4]));
auto piece = "<[" + std::to_string(val) + "]>";
result.push_back({static_cast<llama_token>(val), piece});
offset += 5;
} else {
uint32_t cpt = unicode_cpt_from_utf8(input, offset);
result.push_back({0, unicode_cpt_to_utf8(cpt)});
}
} catch (const std::invalid_argument & /*ex*/) {
// Silently ignore invalid UTF-8 input to avoid leaking the exception beyond llama_tokenize
++offset;
result.push_back({0, unicode_cpt_to_utf8(0xFFFD)}); // replacement character
}
}
return result;
}
static bool match_string(const std::string & input, llama_grammar * grammar) {
const auto parsed = parse_tokens(input);
auto & stacks_cur = llama_grammar_get_stacks(grammar);
for (const auto & in : parsed) {
try {
llama_grammar_accept_token(*grammar, in.token, in.piece);
} catch (const std::runtime_error & /*e*/) {
// normally this shouldn't get hit because of llama_grammar_apply
return false;
}
if (stacks_cur.empty()) {
// no stacks means that the grammar failed to match at this point
return false;
}
}
for (const auto & stack : stacks_cur) {
if (stack.empty()) {
// An empty stack means that the grammar has been completed
return true;
}
}
return false;
}
static void test(const std::string & test_desc, const std::string & grammar_str, const std::vector<std::string> & passing_strings, const std::vector<std::string> & failing_strings) {
fprintf(stderr, "⚫ Testing %s\n%s\n", test_desc.c_str(), grammar_str.c_str());
fflush(stderr);
auto * grammar = build_grammar(grammar_str);
// Save the original grammar stacks so that we can reset after every new string we want to test
const llama_grammar_stacks stacks_org = llama_grammar_get_stacks(grammar); // copy
llama_grammar_stacks & stacks_cur = llama_grammar_get_stacks(grammar);
fprintf(stderr, " 🔵 Valid strings:\n");
// Passing strings
for (const auto & test_string : passing_strings) {
fprintf(stderr, " \"%s\" ", test_string.c_str());
fflush(stderr);
bool matched = match_string(test_string, grammar);
if (!matched) {
fprintf(stderr, "❌ (failed to match)\n");
// DEBUG: Write strings to files so that we can analyze more easily with gbnf-validator program to see exactly where things failed.
// DEBUG: Write the grammar_str to test-grammar-integration.grammar.gbnf
FILE* grammar_file = fopen("test-grammar-integration.grammar.gbnf", "w");
if (grammar_file) {
fprintf(grammar_file, "%s", grammar_str.c_str());
fclose(grammar_file);
}
// DEBUG: Write the test string to test-grammar-integration.string.txt
FILE* string_file = fopen("test-grammar-integration.string.txt", "w");
if (string_file) {
fprintf(string_file, "%s", test_string.c_str());
fclose(string_file);
}
fprintf(stderr, "\n NOTE: Debug grammar file generated. To analyze this failure in detail, run the following command: ./llama-gbnf-validator test-grammar-integration.grammar.gbnf test-grammar-integration.string.txt\n\n");
} else {
fprintf(stdout, "✅︎\n");
}
assert(matched);
// Reset the grammar stacks
stacks_cur = stacks_org;
}
fprintf(stderr, " 🟠 Invalid strings:\n");
// Failing strings
for (const auto & test_string : failing_strings) {
fprintf(stderr, " \"%s\" ", test_string.c_str());
fflush(stderr);
bool matched = match_string(test_string, grammar);
if (matched) {
fprintf(stderr, "❌ (incorrectly matched)\n");
} else {
fprintf(stdout, "✅︎\n");
}
assert(!matched);
// Reset the grammar stacks
stacks_cur = stacks_org;
}
// Clean up allocated memory
llama_grammar_free_impl(grammar);
}
static void test_grammar(const std::string & test_desc, const std::string & grammar_str, const std::vector<std::string> & passing_strings, const std::vector<std::string> & failing_strings) {
test(test_desc + ". Grammar: " + grammar_str, grammar_str, passing_strings, failing_strings);
}
static void test_schema(const std::string & test_desc, const std::string & schema_str, const std::vector<std::string> & passing_strings, const std::vector<std::string> & failing_strings) {
test(test_desc + ". Schema: " + schema_str, json_schema_to_grammar(json::parse(schema_str), true), passing_strings, failing_strings);
}
static void test_simple_grammar() {
test_schema(
"min 0",
R"""({
"type": "integer",
"minimum": 0
})""",
// Passing strings
{
"0",
"10",
"12",
"10000",
},
// Failing strings
{
"-1",
"-10",
"-10000",
"-100000000000000000000000000000000",
"100000000000000000000000000000000",
"00",
"01",
"-0",
}
);
test_schema(
"min 2",
// Schema
R"""({
"type": "integer",
"minimum": 2
})""",
// Passing strings
{
"2",
"3",
"4",
"10",
"20",
"1234567890000000",
},
// Failing strings
{
"0",
"1",
"-1",
"-100",
"0",
"1",
"01",
"02",
"12345678900000000",
}
);
test_schema(
"min 456",
R"""({
"type": "integer",
"minimum": 456
})""",
// Passing strings
{
"456",
"4560",
"457",
"460",
"500",
},
// Failing strings
{
"455",
"356",
"50",
"050",
"-1",
"-456",
}
);
test_schema(
"min -123",
R"""({
"type": "integer",
"minimum": -123
})""",
// Passing strings
{
"-123",
"-122",
"-11",
"-1",
"0",
"1",
"123",
"1234",
"2345",
},
// Failing strings
{
"-1234",
"-124",
}
);
test_schema(
"max 9999",
// Schema
R"""({
"type": "integer",
"maximum": 9999
})""",
// Passing strings
{
"-99999",
"0",
"9999",
},
// Failing strings
{
"10000",
"99991",
}
);
test_schema(
"max -9999",
// Schema
R"""({
"type": "integer",
"maximum": -9999
})""",
// Passing strings
{
"-10000",
"-9999",
},
// Failing strings
{
"-9998",
"0",
"9999",
}
);
test_schema(
"min 5 max 30",
// Schema
R"""({
"type": "integer",
"minimum": 5,
"maximum": 30
})""",
// Passing strings
{
"5",
"10",
"30",
},
// Failing strings
{
"05",
"4",
"-1",
"31",
"123",
"0123",
}
);
test_schema(
"min 1 max 900719925474091",
// Schema
R"""({
"type": "integer",
"exclusiveMinimum": 0,
"maximum": 900719925474091
})""",
// Passing strings
{
"1",
"2",
"10",
"900719925474090",
"900719925474091",
},
// Failing strings
{
"0",
"01",
"900719925474092",
"9007199254740910",
}
);
test_schema(
"min -1 max 1",
R"""({
"type": "integer",
"minimum": -1,
"maximum": 1
})""",
// Passing strings
{
"-1",
"0",
"1",
},
// Failing strings
{
"-11",
"-10",
"-2",
"2",
"10",
"11",
}
);
test_schema(
"min -123 max 42",
R"""({
"type": "integer",
"minimum": -123,
"maximum": 42
})""",
// Passing strings
{
"-123",
"-122",
"-13",
"-11",
"-2",
"-1",
"0",
"1",
"5",
"10",
"39",
"40",
"42",
},
// Failing strings
{
"-0123",
"-124",
"-1123",
"-200",
"43",
"123",
"0123",
}
);
test_schema(
"exclusive min / max",
// Schema
R"""({
"type": "integer",
"exclusiveMinimum": 0,
"exclusiveMaximum": 10000
})""",
// Passing strings
{
"1",
"9999",
},
// Failing strings
{
"0",
"01",
"10000",
"99999",
}
);
// Test case for a simple grammar
test_grammar(
"simple grammar",
R"""(
root ::= expr
expr ::= term ("+" term)*
term ::= number
number ::= [0-9]+)""",
// Passing strings
{
"42",
"1+2+3+4+5",
"123+456",
},
// Failing strings
{
"+",
"/ 3",
"1+2+3+4+5+",
"12a45",
}
);
// Test case for a simple grammar with tokens
test_grammar(
"simple grammar with tokens",
R"""(
root ::= <[10]> content <[11]>
content ::= (!<[11]>)*)""",
// Passing strings
{
token(10) + "hello world" + token(11),
token(10) + "text with " + token(12) + " other tokens " + token(13) + " mixed in" + token(11),
token(10) + token(11),
token(10) + token(12) + token(13) + token(14) + token(15) + token(11),
token(10) + "a" + token(11),
},
// Failing strings
{
token(10) + "missing end token",
token(10),
"missing start token" + token(11),
token(10) + token(11) + token(11), // double end token
token(11) + "wrong order" + token(10),
}
);
}
static void test_complex_grammar() {
// Test case for a more complex grammar, with both failure strings and success strings
test_grammar(
"medium complexity grammar",
// Grammar
R"""(
root ::= expression
expression ::= term ws (("+"|"-") ws term)*
term ::= factor ws (("*"|"/") ws factor)*
factor ::= number | variable | "(" expression ")" | function-call
number ::= [0-9]+
variable ::= [a-zA-Z_][a-zA-Z0-9_]*
function-call ::= variable ws "(" (expression ("," ws expression)*)? ")"
ws ::= [ \t\n\r]?)""",
// Passing strings
{
"42",
"1*2*3*4*5",
"x",
"x+10",
"x1+y2",
"(a+b)*(c-d)",
"func()",
"func(x,y+2)",
"a*(b+c)-d/e",
"f(g(x),h(y,z))",
"x + 10",
"x1 + y2",
"(a + b) * (c - d)",
"func()",
"func(x, y + 2)",
"a * (b + c) - d / e",
"f(g(x), h(y, z))",
"123+456",
"123*456*789-123/456+789*123",
"123+456*789-123/456+789*123-456/789+123*456-789/123+456*789-123/456+789*123-456"
},
// Failing strings
{
"+",
"/ 3x",
"x + + y",
"a * / b",
"func(,)",
"func(x y)",
"(a + b",
"x + y)",
"a + b * (c - d",
"42 +",
"x +",
"x + 10 +",
"(a + b) * (c - d",
"func(",
"func(x, y + 2",
"a * (b + c) - d /",
"f(g(x), h(y, z)",
"123+456*789-123/456+789*123-456/789+123*456-789/123+456*789-123/456+789*123-456/",
}
);
// Test case for a more complex grammar with tokens
test_grammar(
"complex grammar with tokens",
R"""(
root ::= reasoning+ content tool-call*
reasoning ::= <[10]> (!<[11]>)* <[11]>
content ::= <[20]> (!<[21]>)* <[21]>
tool-call ::= <[12]> name <[13]> args <[14]>
name ::= (!<[13]>)+
args ::= (!<[14]>)*)""",
// Passing strings
{
token(10) + "I am thinking" + token(11) + token(20) + "hello world!" + token(21) + token(12) + "search" + token(13) + "query=test" + token(14),
token(10) + "reasoning 1" + token(11) + token(10) + "reasoning 2" + token(11) + token(20) + token(21) + token(12) + "tool" + token(13) + token(14),
token(10) + token(11) + token(20) + "content" + token(21),
token(10) + "think" + token(12) + " nested" + token(11) + token(20) + token(10) + "more content" + token(21) + token(12) + "fn" + token(13) + "x=1,y=2" + token(14) + token(12) + "fn2" + token(13) + token(14),
token(10) + "reasoning" + token(11) + token(10) + "more" + token(11) + token(10) + "even more" + token(11) + token(20) + "text" + token(21) + token(12) + "a" + token(13) + "b" + token(14) + token(12) + "c" + token(13) + "d" + token(14),
},
// Failing strings
{
token(20) + "content only" + token(21),
token(10) + "no closing reasoning",
token(10) + token(11) + token(20) + "no closing content",
token(10) + token(11) + token(20) + token(21) + token(12) + "incomplete tool",
token(10) + token(11) + token(11) + token(20) + token(21),
}
);
}
static void test_special_chars() {
// A collection of tests to exercise special characters such as "."
test_grammar(
"special characters",
// Grammar
R"""(
root ::= ... "abc" ...
)""",
// Passing strings
{
"abcabcabc",
"aaaabcccc",
// NOTE: Also ensures that multi-byte characters still count as a single character
"🔵🟠✅abc❌🟠🔵"
},
// Failing strings
{
"aaabcccc",
"aaaaabcccc",
"aaaabccc",
"aaaabccccc",
"🔵🟠✅❌abc❌✅🟠🔵",
"🔵🟠abc🟠🔵"
}
);
}
static void test_quantifiers() {
// A collection of tests to exercise * + and ? quantifiers
test_grammar(
"* quantifier",
// Grammar
R"""(root ::= "a"*)""",
// Passing strings
{
"",
"a",
"aaaaa",
"aaaaaaaaaaaaaaaaaa",
"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
},
// Failing strings
{
"b",
"ab",
"aab",
"ba",
"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab"
}
);
test_grammar(
"+ quantifier",
// Grammar
R"""(root ::= "a"+)""",
// Passing strings
{
"a",
"aaaaa",
"aaaaaaaaaaaaaaaaaa",
"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
},
// Failing strings
{
"",
"b",
"ab",
"aab",
"ba",
"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab"
}
);
test_grammar(
"? quantifier",
// Grammar
R"""(root ::= "a"?)""",
// Passing strings
{
"",
"a"
},
// Failing strings
{
"b",
"ab",
"aa",
"ba",
}
);
test_grammar(
"mixed quantifiers",
// Grammar
R"""(
root ::= cons+ vowel* cons? (vowel cons)*
vowel ::= [aeiouy]
cons ::= [bcdfghjklmnpqrstvwxyz]
)""",
// Passing strings
{
"yes",
"no",
"noyes",
"crwth",
"four",
"bryyyy",
},
// Failing strings
{
"yess",
"yesno",
"forty",
"catyyy",
}
);
test_grammar(
"simple exact repetition",
// Grammar
R"""(
root ::= [ab]{4}
)""",
// Passing strings
{
"aaaa",
"bbbb",
"abab",
},
// Failing strings
{
"a",
"b",
"aaaaa",
}
);
test_grammar(
"simple min repetition",
// Grammar
R"""(
root ::= [ab]{4,}
)""",
// Passing strings
{
"aaaa",
"aaaaab",
"bbbb",
"ababab",
},
// Failing strings
{
"",
"aba",
}
);
test_grammar(
"simple max repetition",
// Grammar
R"""(
root ::= [ab]{0,4}
)""",
// Passing strings
{
"",
"a",
"aa",
"aaa",
"aaab",
},
// Failing strings
{
"aaaaa",
}
);
test_grammar(
"min / max repetition",
// Grammar
R"""(
root ::= ("0x" [A-F0-9]{2} " "?){3,5}
)""",
// Passing strings
{
"0xFF 0x12 0xAB",
"0xFF 0x12 0xAB 0x00 0x00",
},
// Failing strings
{
"",
"0xFF",
"0xFF 0x12",
"0xFF 0x12 0xAB 0x00 0x00 0x00",
}
);
test_grammar(
"segfault",
// Grammar
R"""(
root ::= ( [x]* )*
)""",
// Passing strings
{
"",
"x",
"xx"
},
// Failing strings
{
"y",
"yy"
}
);
}
static void test_failure_missing_root() {
fprintf(stderr, "⚫ Testing missing root node:\n");
// Test case for a grammar that is missing a root rule
const std::string grammar_str = R"""(
rot ::= expr
expr ::= term ("+" term)*
term ::= number
number ::= [0-9]+)""";
llama_grammar_parser parsed_grammar;
parsed_grammar.parse(grammar_str.c_str());
// Ensure we parsed correctly
assert(!parsed_grammar.rules.empty());
// Ensure we do NOT have a root node
assert(parsed_grammar.symbol_ids.find("root") == parsed_grammar.symbol_ids.end());
fprintf(stderr, " ✅︎ Passed\n");
}
static void test_failure_missing_reference() {
fprintf(stderr, "⚫ Testing missing reference node:\n");
// Test case for a grammar that is missing a referenced rule
const std::string grammar_str =
R"""(root ::= expr
expr ::= term ("+" term)*
term ::= numero
number ::= [0-9]+)""";
fprintf(stderr, " Expected error: ");
llama_grammar_parser parsed_grammar;
parsed_grammar.parse(grammar_str.c_str());
// Ensure we did NOT parsed correctly
assert(parsed_grammar.rules.empty());
fprintf(stderr, " End of expected error.\n");
fprintf(stderr, " ✅︎ Passed\n");
}
static void test_failure_left_recursion() {
fprintf(stderr, "⚫ Testing left recursion detection:\n");
// Test simple left recursion detection
const std::string simple_str = R"""(root ::= "a" | root "a")""";
assert(test_build_grammar_fails(simple_str));
// Test more complicated left recursion detection
const std::string medium_str = R"""(
root ::= asdf
asdf ::= "a" | asdf "a"
)""";
assert(test_build_grammar_fails(medium_str));
// Test even more complicated left recursion detection
const std::string hard_str = R"""(
root ::= asdf
asdf ::= "a" | foo "b"
foo ::= "c" | asdf "d" | "e")""";
assert(test_build_grammar_fails(hard_str));
// Test yet even more complicated left recursion detection
const std::string hardest_str = R"""(
root ::= asdf
asdf ::= "a" | foo "b"
foo ::= "c" | empty asdf "d" | "e"
empty ::= "blah" | )""";
assert(test_build_grammar_fails(hardest_str));
fprintf(stderr, " ✅︎ Passed\n");
}
static void test_failure_missing_root_symbol() {
fprintf(stderr, "⚫ Testing missing root symbol:\n");
const std::string grammar_str = R"""(
root ::= "foobar"
)""";
llama_grammar * failure_result = build_grammar_with_root(grammar_str, "nonexistent");
assert(failure_result == nullptr);
fprintf(stderr, " ✅︎ Passed\n");
}
static void test_custom_root_symbol_check() {
fprintf(stderr, "⚫ Testing custom root symbol check:\n");
const std::string custom_root_grammar_str = R"""(
foobar ::= "foobar"
)""";
llama_grammar * failure_result = build_grammar_with_root(custom_root_grammar_str, "root");
assert(failure_result == nullptr);
llama_grammar * success_result = build_grammar_with_root(custom_root_grammar_str, "foobar");
assert(success_result != nullptr);
llama_grammar_free_impl(success_result);
fprintf(stderr, " ✅︎ Passed\n");
}
static void test_json_schema() {
// Note that this is similar to the regular grammar tests,
// but we convert each json schema to a grammar before parsing.
// Otherwise, this test structure is the same.
test_schema(
"empty schema (object)",
// Schema
R"""(
{}
)""",
// Passing strings
{
R"""({})""",
R"""({"foo": "bar"})""",
},
// Failing strings
{
"",
"[]",
"null",
R"""("")""",
"true",
}
);
test_schema(
"exotic formats (list)",
// Schema
R"""({
"items": [
{ "format": "date" },
{ "format": "uuid" },
{ "format": "time" },
{ "format": "date-time" }
]
})""",
// Passing strings
{
// "{}", // NOTE: This string passes for this schema on https://www.jsonschemavalidator.net/ -- should it?
// "[]", // NOTE: This string passes for this schema on https://www.jsonschemavalidator.net/ -- should it?
R"""(["2012-04-23", "12345678-1234-1234-1234-1234567890ab", "18:25:43.511Z", "2012-04-23T18:25:43.511Z"])""",
//R"""(["2012-04-23","12345678-1234-1234-1234-1234567890ab"])""", // NOTE: This string passes for this schema on https://www.jsonschemavalidator.net/ -- should it?
//R"""({"foo": "bar"})""", // NOTE: This string passes for this schema on https://www.jsonschemavalidator.net/ -- should it?
},
// Failing strings
{
R"""(["foo", "bar"])""",
R"""(["12345678-1234-1234-1234-1234567890ab"])""",
}
);
test_schema(
"string",
// Schema
R"""({
"type": "string"
})""",
// Passing strings
{
R"""("foo")""",
R"""("bar")""",
R"""("")""",
},
// Failing strings
{
R"""({})""",
R"""("foo": "bar")""",
}
);
test_schema(
"string w/ min length 1",
// Schema
R"""({
"type": "string",
"minLength": 1
})""",
// Passing strings
{
R"""("foo")""",
R"""("bar")""",
},
// Failing strings
{
R"""("")""",
You can’t perform that action at this time.