File size: 110,984 Bytes
03c6ad0
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
2f7da2f
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
 
 
 
 
 
03c6ad0
 
 
 
b292d4f
03c6ad0
 
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
 
 
 
b292d4f
 
03c6ad0
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
 
 
 
 
b292d4f
 
 
 
 
 
03c6ad0
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
 
 
b292d4f
 
 
03c6ad0
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
b292d4f
03c6ad0
 
 
b292d4f
03c6ad0
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
 
 
03c6ad0
 
b292d4f
03c6ad0
 
 
 
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
b292d4f
 
03c6ad0
b292d4f
03c6ad0
 
 
 
 
 
 
b292d4f
 
 
 
 
 
 
 
03c6ad0
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
b292d4f
03c6ad0
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
b292d4f
 
03c6ad0
 
b292d4f
03c6ad0
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
 
 
 
 
 
 
b292d4f
 
 
03c6ad0
 
b292d4f
03c6ad0
 
 
 
 
2f7da2f
 
 
03c6ad0
b292d4f
 
 
 
03c6ad0
 
b292d4f
03c6ad0
 
 
 
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
 
 
b292d4f
 
03c6ad0
 
 
 
 
 
 
b292d4f
 
03c6ad0
 
 
 
b292d4f
 
03c6ad0
 
 
 
 
 
 
b292d4f
 
 
03c6ad0
 
b292d4f
03c6ad0
 
 
 
b292d4f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
03c6ad0
 
 
 
 
 
b292d4f
03c6ad0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b292d4f
 
03c6ad0
 
 
 
 
 
 
 
 
 
 
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
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"><head>

<meta charset="utf-8">
<meta name="generator" content="quarto-1.8.27">

<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">


<title>3&nbsp; Structured Information Extraction with Vision Language Models – AI Design Patterns for GLAM</title>
<style>
code{white-space: pre-wrap;}
span.smallcaps{font-variant: small-caps;}
div.columns{display: flex; gap: min(4vw, 1.5em);}
div.column{flex: auto; overflow-x: auto;}
div.hanging-indent{margin-left: 1.5em; text-indent: -1.5em;}
ul.task-list{list-style: none;}
ul.task-list li input[type="checkbox"] {
  width: 0.8em;
  margin: 0 0.8em 0.2em -1em; /* quarto-specific, see https://github.com/quarto-dev/quarto-cli/issues/4556 */ 
  vertical-align: middle;
}
/* CSS for syntax highlighting */
html { -webkit-text-size-adjust: 100%; }
pre > code.sourceCode { white-space: pre; position: relative; }
pre > code.sourceCode > span { display: inline-block; line-height: 1.25; }
pre > code.sourceCode > span:empty { height: 1.2em; }
.sourceCode { overflow: visible; }
code.sourceCode > span { color: inherit; text-decoration: inherit; }
div.sourceCode { margin: 1em 0; }
pre.sourceCode { margin: 0; }
@media screen {
div.sourceCode { overflow: auto; }
}
@media print {
pre > code.sourceCode { white-space: pre-wrap; }
pre > code.sourceCode > span { text-indent: -5em; padding-left: 5em; }
}
pre.numberSource code
  { counter-reset: source-line 0; }
pre.numberSource code > span
  { position: relative; left: -4em; counter-increment: source-line; }
pre.numberSource code > span > a:first-child::before
  { content: counter(source-line);
    position: relative; left: -1em; text-align: right; vertical-align: baseline;
    border: none; display: inline-block;
    -webkit-touch-callout: none; -webkit-user-select: none;
    -khtml-user-select: none; -moz-user-select: none;
    -ms-user-select: none; user-select: none;
    padding: 0 4px; width: 4em;
  }
pre.numberSource { margin-left: 3em;  padding-left: 4px; }
div.sourceCode
  {   }
@media screen {
pre > code.sourceCode > span > a:first-child::before { text-decoration: underline; }
}
</style>


<script src="https://cdn.jsdelivr.net/npm/jquery@3.5.1/dist/jquery.min.js" integrity="sha384-ZvpUoO/+PpLXR1lu4jmpXWu80pZlYUAfxl5NsBMWOEPSjUn/6Z/hRTt8+pR6L4N2" crossorigin="anonymous"></script><script src="../../site_libs/quarto-nav/quarto-nav.js"></script>
<script src="../../site_libs/quarto-nav/headroom.min.js"></script>
<script src="../../site_libs/clipboard/clipboard.min.js"></script>
<script src="../../site_libs/quarto-search/autocomplete.umd.js"></script>
<script src="../../site_libs/quarto-search/fuse.min.js"></script>
<script src="../../site_libs/quarto-search/quarto-search.js"></script>
<meta name="quarto:offset" content="../../">
<link href="../../patterns/structured-generation/intro.html" rel="prev">
<script src="../../site_libs/quarto-html/quarto.js" type="module"></script>
<script src="../../site_libs/quarto-html/tabsets/tabsets.js" type="module"></script>
<script src="../../site_libs/quarto-html/axe/axe-check.js" type="module"></script>
<script src="../../site_libs/quarto-html/popper.min.js"></script>
<script src="../../site_libs/quarto-html/tippy.umd.min.js"></script>
<script src="../../site_libs/quarto-html/anchor.min.js"></script>
<link href="../../site_libs/quarto-html/tippy.css" rel="stylesheet">
<link href="../../site_libs/quarto-html/quarto-syntax-highlighting-ed96de9b727972fe78a7b5d16c58bf87.css" rel="stylesheet" id="quarto-text-highlighting-styles">
<script src="../../site_libs/bootstrap/bootstrap.min.js"></script>
<link href="../../site_libs/bootstrap/bootstrap-icons.css" rel="stylesheet">
<link href="../../site_libs/bootstrap/bootstrap-27c261d06b905028a18691de25d09dde.min.css" rel="stylesheet" append-hash="true" id="quarto-bootstrap" data-mode="light">
<script id="quarto-search-options" type="application/json">{
  "location": "sidebar",
  "copy-button": false,
  "collapse-after": 3,
  "panel-placement": "start",
  "type": "textbox",
  "limit": 50,
  "keyboard-shortcut": [
    "f",
    "/",
    "s"
  ],
  "show-item-context": false,
  "language": {
    "search-no-results-text": "No results",
    "search-matching-documents-text": "matching documents",
    "search-copy-link-title": "Copy link to search",
    "search-hide-matches-text": "Hide additional matches",
    "search-more-match-text": "more match in this document",
    "search-more-matches-text": "more matches in this document",
    "search-clear-button-title": "Clear",
    "search-text-placeholder": "",
    "search-detached-cancel-button-title": "Cancel",
    "search-submit-button-title": "Submit",
    "search-label": "Search"
  }
}</script>
<script src="https://cdn.jsdelivr.net/npm/requirejs@2.3.6/require.min.js" integrity="sha384-c9c+LnTbwQ3aujuU7ULEPVvgLs+Fn6fJUvIGTsuu1ZcCf11fiEubah0ttpca4ntM sha384-6V1/AdqZRWk1KAlWbKBlGhN7VG4iE/yAZcO6NZPMF8od0vukrvr0tg4qY6NSrItx" crossorigin="anonymous"></script>

<script type="application/javascript">define('jquery', [],function() {return window.jQuery;})</script>
<script src="https://cdn.jsdelivr.net/npm/@jupyter-widgets/html-manager@*/dist/embed-amd.js" crossorigin="anonymous"></script>


</head>

<body class="nav-sidebar floating quarto-light">

<div id="quarto-search-results"></div>
  <header id="quarto-header" class="headroom fixed-top">
  <nav class="quarto-secondary-nav">
    <div class="container-fluid d-flex">
      <button type="button" class="quarto-btn-toggle btn" data-bs-toggle="collapse" role="button" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">
        <i class="bi bi-layout-text-sidebar-reverse"></i>
      </button>
        <nav class="quarto-page-breadcrumbs" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../../patterns/structured-generation/intro.html">Structured Information Extraction</a></li><li class="breadcrumb-item"><a href="../../patterns/structured-generation/vlm-structured-generation.html"><span class="chapter-number">3</span>&nbsp; <span class="chapter-title">Structured Information Extraction with Vision Language Models</span></a></li></ol></nav>
        <a class="flex-grow-1" role="navigation" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item" aria-controls="quarto-sidebar" aria-expanded="false" aria-label="Toggle sidebar navigation" onclick="if (window.quartoToggleHeadroom) { window.quartoToggleHeadroom(); }">      
        </a>
      <button type="button" class="btn quarto-search-button" aria-label="Search" onclick="window.quartoOpenSearch();">
        <i class="bi bi-search"></i>
      </button>
    </div>
  </nav>
</header>
<!-- content -->
<div id="quarto-content" class="quarto-container page-columns page-rows-contents page-layout-article">
<!-- sidebar -->
  <nav id="quarto-sidebar" class="sidebar collapse collapse-horizontal quarto-sidebar-collapse-item sidebar-navigation floating overflow-auto">
    <div class="pt-lg-2 mt-2 text-left sidebar-header">
    <div class="sidebar-title mb-0 py-0">
      <a href="../../">AI Design Patterns for GLAM</a> 
    </div>
      </div>
        <div class="mt-2 flex-shrink-0 align-items-center">
        <div class="sidebar-search">
        <div id="quarto-search" class="" title="Search"></div>
        </div>
        </div>
    <div class="sidebar-menu-container"> 
    <ul class="list-unstyled mt-1">
        <li class="sidebar-item">
  <div class="sidebar-item-container"> 
  <a href="../../index.html" class="sidebar-item-text sidebar-link">
 <span class="menu-text">Welcome</span></a>
  </div>
</li>
        <li class="sidebar-item sidebar-item-section">
      <span class="sidebar-item-text sidebar-link text-start">
 <span class="menu-text">Beyond Chat Interfaces to Collections?</span></span>
  </li>
        <li class="sidebar-item sidebar-item-section">
      <div class="sidebar-item-container"> 
            <a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-2" role="navigation" aria-expanded="true">
 <span class="menu-text">Design Patterns</span></a>
          <a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-2" role="navigation" aria-expanded="true" aria-label="Toggle section">
            <i class="bi bi-chevron-right ms-2"></i>
          </a> 
      </div>
      <ul id="quarto-sidebar-section-2" class="collapse list-unstyled sidebar-section depth1 show">  
          <li class="sidebar-item">
  <div class="sidebar-item-container"> 
  <a href="../../patterns/what-is-an-ai-pattern.html" class="sidebar-item-text sidebar-link">
 <span class="menu-text"><span class="chapter-number">1</span>&nbsp; <span class="chapter-title">What is an AI Pattern?</span></span></a>
  </div>
</li>
      </ul>
  </li>
        <li class="sidebar-item sidebar-item-section">
      <div class="sidebar-item-container"> 
            <a class="sidebar-item-text sidebar-link text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true">
 <span class="menu-text">Structured Information Extraction</span></a>
          <a class="sidebar-item-toggle text-start" data-bs-toggle="collapse" data-bs-target="#quarto-sidebar-section-3" role="navigation" aria-expanded="true" aria-label="Toggle section">
            <i class="bi bi-chevron-right ms-2"></i>
          </a> 
      </div>
      <ul id="quarto-sidebar-section-3" class="collapse list-unstyled sidebar-section depth1 show">  
          <li class="sidebar-item">
  <div class="sidebar-item-container"> 
  <a href="../../patterns/structured-generation/intro.html" class="sidebar-item-text sidebar-link">
 <span class="menu-text"><span class="chapter-number">2</span>&nbsp; <span class="chapter-title">Structured Document Processing</span></span></a>
  </div>
</li>
          <li class="sidebar-item">
  <div class="sidebar-item-container"> 
  <a href="../../patterns/structured-generation/vlm-structured-generation.html" class="sidebar-item-text sidebar-link active">
 <span class="menu-text"><span class="chapter-number">3</span>&nbsp; <span class="chapter-title">Structured Information Extraction with Vision Language Models</span></span></a>
  </div>
</li>
      </ul>
  </li>
    </ul>
    </div>
</nav>
<div id="quarto-sidebar-glass" class="quarto-sidebar-collapse-item" data-bs-toggle="collapse" data-bs-target=".quarto-sidebar-collapse-item"></div>
<!-- margin-sidebar -->
    <div id="quarto-margin-sidebar" class="sidebar margin-sidebar">
        <nav id="TOC" role="doc-toc" class="toc-active">
    <h2 id="toc-title">Table of contents</h2>
   
  <ul>
  <li><a href="#introduction" id="toc-introduction" class="nav-link active" data-scroll-target="#introduction"><span class="header-section-number">3.1</span> Introduction</a></li>
  <li><a href="#the-sloane-index-cards-dataset" id="toc-the-sloane-index-cards-dataset" class="nav-link" data-scroll-target="#the-sloane-index-cards-dataset"><span class="header-section-number">3.2</span> The Sloane Index Cards Dataset</a></li>
  <li><a href="#setup" id="toc-setup" class="nav-link" data-scroll-target="#setup"><span class="header-section-number">3.3</span> Setup</a>
  <ul class="collapse">
  <li><a href="#connecting-to-a-vision-language-model" id="toc-connecting-to-a-vision-language-model" class="nav-link" data-scroll-target="#connecting-to-a-vision-language-model"><span class="header-section-number">3.3.1</span> Connecting to a Vision Language Model</a></li>
  </ul></li>
  <li><a href="#basic-vlm-query" id="toc-basic-vlm-query" class="nav-link" data-scroll-target="#basic-vlm-query"><span class="header-section-number">3.4</span> Basic VLM Query</a></li>
  <li><a href="#simple-vlm-query-example" id="toc-simple-vlm-query-example" class="nav-link" data-scroll-target="#simple-vlm-query-example"><span class="header-section-number">3.5</span> 102: Simple VLM Query Example</a></li>
  <li><a href="#classification" id="toc-classification" class="nav-link" data-scroll-target="#classification"><span class="header-section-number">3.6</span> Classification</a>
  <ul class="collapse">
  <li><a href="#classifying-with-structured-labels" id="toc-classifying-with-structured-labels" class="nav-link" data-scroll-target="#classifying-with-structured-labels"><span class="header-section-number">3.6.1</span> Classifying with structured labels</a></li>
  </ul></li>
  <li><a href="#beyond-classifying---extracting-structured-information" id="toc-beyond-classifying---extracting-structured-information" class="nav-link" data-scroll-target="#beyond-classifying---extracting-structured-information"><span class="header-section-number">3.7</span> Beyond classifying - Extracting structured information</a></li>
  <li><a href="#appendix-using-a-local-model" id="toc-appendix-using-a-local-model" class="nav-link" data-scroll-target="#appendix-using-a-local-model"><span class="header-section-number">3.8</span> Appendix: Using a Local Model</a></li>
  </ul>
</nav>
    </div>
<!-- main -->
<main class="content" id="quarto-document-content">

<header id="title-block-header" class="quarto-title-block default"><nav class="quarto-page-breadcrumbs quarto-title-breadcrumbs d-none d-lg-block" aria-label="breadcrumb"><ol class="breadcrumb"><li class="breadcrumb-item"><a href="../../patterns/structured-generation/intro.html">Structured Information Extraction</a></li><li class="breadcrumb-item"><a href="../../patterns/structured-generation/vlm-structured-generation.html"><span class="chapter-number">3</span>&nbsp; <span class="chapter-title">Structured Information Extraction with Vision Language Models</span></a></li></ol></nav>
<div class="quarto-title">
<h1 class="title"><span class="chapter-number">3</span>&nbsp; <span class="chapter-title">Structured Information Extraction with Vision Language Models</span></h1>
</div>



<div class="quarto-title-meta">

    
  
    
  </div>
  


</header>


<section id="introduction" class="level2" data-number="3.1">
<h2 data-number="3.1" class="anchored" data-anchor-id="introduction"><span class="header-section-number">3.1</span> Introduction</h2>
<p>In this chapter we’ll start to look at how we can use Visual Language Models (VLMs) to extract structured information from images of documents.</p>
<p>We already saw what this looked like at a conceptual level in the previous chapter. In this chapter we’ll get hands on with some code examples to illustrate how this can be done in practice. To start we’ll focus on some relatively simple documents and tasks. This allows us to focus on the core concepts without getting bogged down in too many complexities. We’ll use open source models accessed via the Hugging Face Inference API (you can also run them locally — see the <a href="#appendix-using-a-local-model">appendix</a>).</p>
</section>
<section id="the-sloane-index-cards-dataset" class="level2" data-number="3.2">
<h2 data-number="3.2" class="anchored" data-anchor-id="the-sloane-index-cards-dataset"><span class="header-section-number">3.2</span> The Sloane Index Cards Dataset</h2>
<p>We’ll use the <a href="https://huggingface.co/datasets/biglam/sloane-index-cards">Sloane Index Cards Dataset</a> from Hugging Face for our examples. This is a publicly available dataset that is well suited to demonstrating structured information extraction with VLMs.</p>
<blockquote class="blockquote">
<p>The files in this dataset are derived from microfilm copies of the original library catalogue of Sir Hans Sloane, now presented across 9 volumes, Sloane MS 3972 C 1-8, and the name index to the Sloane library catalogue, Sloane MS 3972 D. The catalogues are crucial for understanding the development of Sloane’s collections, the present-day collections of the British Library, British Museum and Natural History Museum, and to identifying collection items which are now dispersed across a number of institutions.</p>
</blockquote>
<p>The dataset is available in Parquet format on <a href="https://huggingface.co/datasets/biglam/biglam/sloane-catalogues">Hugging Face</a> so it can be easily loaded using the <code>datasets</code> library.</p>
<p>Let’s load the dataset and take a look at one row.</p>
<div id="7d5760cc" class="cell" data-cache="true" data-execution_count="1">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb1"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb1-1"><a href="#cb1-1" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> datasets <span class="im">import</span> load_dataset</span>
<span id="cb1-2"><a href="#cb1-2" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb1-3"><a href="#cb1-3" aria-hidden="true" tabindex="-1"></a>ds <span class="op">=</span> load_dataset(<span class="st">"biglam/sloane-catalogues"</span>, split<span class="op">=</span><span class="st">"train"</span>)</span>
<span id="cb1-4"><a href="#cb1-4" aria-hidden="true" tabindex="-1"></a>ds[<span class="dv">0</span>]</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="1">
<pre><code>{'image': &lt;PIL.JpegImagePlugin.JpegImageFile image mode=L size=3144x2267&gt;,
 'filename': 'sloane_ms_3972_c!1_001.jpg',
 'collection': 'sloane_ms_3972_c!1_jpegs',
 'page_number': 1,
 'page_index_in_directory': 0,
 'source': 'British Library Sloane Manuscripts'}</code></pre>
</div>
</div>
<p>We can see that we have a dictionary that contains an <code>image</code> as well as some additional metadata fields.</p>
<p>Let’s take a look at an actual example image from the dataset.</p>
<div id="902a544b" class="cell" data-cache="true" data-execution_count="2">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb3"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb3-1"><a href="#cb3-1" aria-hidden="true" tabindex="-1"></a>ds[<span class="dv">0</span>][<span class="st">"image"</span>]</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="2">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-3-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
<p>Let’s look at a couple more examples to get a sense of the variety in the dataset.</p>
<div id="b1e3fc05" class="cell" data-execution_count="3">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb4"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb4-1"><a href="#cb4-1" aria-hidden="true" tabindex="-1"></a>ds[<span class="dv">2</span>][<span class="st">"image"</span>]</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="3">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-4-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
<p>one more from later in the dataset</p>
<div id="b23fa754" class="cell" data-execution_count="4">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb5"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb5-1"><a href="#cb5-1" aria-hidden="true" tabindex="-1"></a>ds[<span class="dv">50</span>][<span class="st">"image"</span>]</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="4">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-5-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
<p>We can see we have a mixture of different types of digitised content here including index cards from the original microfilm as well as the actual handwritten manuscript pages from Sloane’s collection.</p>
<p>We’ll look at how we can use VLMs to work with kind of collection.</p>
</section>
<section id="setup" class="level2" data-number="3.3">
<h2 data-number="3.3" class="anchored" data-anchor-id="setup"><span class="header-section-number">3.3</span> Setup</h2>
<section id="connecting-to-a-vision-language-model" class="level3" data-number="3.3.1">
<h3 data-number="3.3.1" class="anchored" data-anchor-id="connecting-to-a-vision-language-model"><span class="header-section-number">3.3.1</span> Connecting to a Vision Language Model</h3>
<p>We’ll use <a href="https://huggingface.co/docs/api-inference/">Hugging Face Inference Providers</a> to access VLMs via an API. This means we don’t need to install or run any models locally — we just need a free Hugging Face account and an API token.</p>
<p>Since the Hugging Face Inference API is compatible with the OpenAI Python client, all the code in this chapter will also work with local model servers (like LM Studio, Ollama, or vLLM) with just a one-line change to the client setup. See the <a href="#appendix-using-a-local-model">appendix</a> at the end of this chapter for details.</p>
<div class="callout callout-style-default callout-tip callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
<span class="screen-reader-only">Tip</span>Getting a Hugging Face Token
</div>
</div>
<div class="callout-body-container callout-body">
<ol type="1">
<li>Create a free account at <a href="https://huggingface.co/join">huggingface.co</a></li>
<li>Go to <a href="https://huggingface.co/settings/tokens">Settings → Access Tokens</a></li>
<li>Create a new token with <strong>Read</strong> access</li>
<li>Set it as an environment variable: <code>export HF_TOKEN=hf_...</code> or add it to a <code>.env</code> file</li>
</ol>
</div>
</div>
<div id="4fc5f6da" class="cell" data-cache="true" data-execution_count="5">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb6"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb6-1"><a href="#cb6-1" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> os</span>
<span id="cb6-2"><a href="#cb6-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> openai <span class="im">import</span> OpenAI</span>
<span id="cb6-3"><a href="#cb6-3" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> rich <span class="im">import</span> <span class="bu">print</span> <span class="im">as</span> rprint</span>
<span id="cb6-4"><a href="#cb6-4" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> dotenv <span class="im">import</span> load_dotenv</span>
<span id="cb6-5"><a href="#cb6-5" aria-hidden="true" tabindex="-1"></a>load_dotenv()</span>
<span id="cb6-6"><a href="#cb6-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb6-7"><a href="#cb6-7" aria-hidden="true" tabindex="-1"></a>client <span class="op">=</span> OpenAI(</span>
<span id="cb6-8"><a href="#cb6-8" aria-hidden="true" tabindex="-1"></a>    base_url<span class="op">=</span><span class="st">"https://router.huggingface.co/v1"</span>,</span>
<span id="cb6-9"><a href="#cb6-9" aria-hidden="true" tabindex="-1"></a>    api_key<span class="op">=</span>os.environ.get(<span class="st">"HF_TOKEN"</span>),</span>
<span id="cb6-10"><a href="#cb6-10" aria-hidden="true" tabindex="-1"></a>)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
</div>
<p>We’ll use <code>Qwen/Qwen3-VL-8B-Instruct</code> throughout this chapter — an 8 billion parameter vision-language model that offers a good balance of quality and speed for document understanding tasks.</p>
</section>
</section>
<section id="basic-vlm-query" class="level2" data-number="3.4">
<h2 data-number="3.4" class="anchored" data-anchor-id="basic-vlm-query"><span class="header-section-number">3.4</span> Basic VLM Query</h2>
<p>Let’s start by defining a simple function that we can use to query a VLM with an image and a prompt. This function will handle converting the image to base64 and sending the request to the model.</p>
<p>We’ll default to using the <code>Qwen/Qwen3-VL-8B-Instruct</code> model via HF Inference Providers. You could experiment with different models later on — the code works with any OpenAI-compatible VLM endpoint.</p>
<div id="36489552" class="cell" data-cache="true" data-execution_count="6">
<details class="code-fold">
<summary>Code</summary>
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb7"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb7-1"><a href="#cb7-1" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> base64</span>
<span id="cb7-2"><a href="#cb7-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> PIL.Image <span class="im">import</span> Image <span class="im">as</span> PILImage</span>
<span id="cb7-3"><a href="#cb7-3" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> io <span class="im">import</span> BytesIO</span>
<span id="cb7-4"><a href="#cb7-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb7-5"><a href="#cb7-5" aria-hidden="true" tabindex="-1"></a><span class="kw">def</span> query_image(image: <span class="bu">str</span> <span class="op">|</span> PILImage, prompt: <span class="bu">str</span>, model: <span class="bu">str</span><span class="op">=</span><span class="st">'Qwen/Qwen3-VL-8B-Instruct'</span>, max_image_size: <span class="bu">int</span><span class="op">=</span><span class="dv">1024</span>, client<span class="op">=</span>client) <span class="op">-&gt;</span> <span class="bu">str</span>:</span>
<span id="cb7-6"><a href="#cb7-6" aria-hidden="true" tabindex="-1"></a>    <span class="co">"""Query VLM with an image."""</span></span>
<span id="cb7-7"><a href="#cb7-7" aria-hidden="true" tabindex="-1"></a>    <span class="cf">if</span> <span class="bu">isinstance</span>(image, PILImage):</span>
<span id="cb7-8"><a href="#cb7-8" aria-hidden="true" tabindex="-1"></a>        <span class="co"># Convert PIL Image to bytes and encode to base64</span></span>
<span id="cb7-9"><a href="#cb7-9" aria-hidden="true" tabindex="-1"></a>        buffered <span class="op">=</span> BytesIO()</span>
<span id="cb7-10"><a href="#cb7-10" aria-hidden="true" tabindex="-1"></a>        <span class="co"># ensure image is not too big</span></span>
<span id="cb7-11"><a href="#cb7-11" aria-hidden="true" tabindex="-1"></a>        <span class="cf">if</span> image.size <span class="op">&gt;</span> (max_image_size, max_image_size):</span>
<span id="cb7-12"><a href="#cb7-12" aria-hidden="true" tabindex="-1"></a>            image.thumbnail((max_image_size, max_image_size))   </span>
<span id="cb7-13"><a href="#cb7-13" aria-hidden="true" tabindex="-1"></a>        image.save(buffered, <span class="bu">format</span><span class="op">=</span><span class="st">"JPEG"</span>)</span>
<span id="cb7-14"><a href="#cb7-14" aria-hidden="true" tabindex="-1"></a>        image_base64 <span class="op">=</span> base64.b64encode(buffered.getvalue()).decode(<span class="st">'utf-8'</span>)</span>
<span id="cb7-15"><a href="#cb7-15" aria-hidden="true" tabindex="-1"></a>    <span class="cf">else</span>:</span>
<span id="cb7-16"><a href="#cb7-16" aria-hidden="true" tabindex="-1"></a>        <span class="co"># Assume image is a file path</span></span>
<span id="cb7-17"><a href="#cb7-17" aria-hidden="true" tabindex="-1"></a>        <span class="cf">with</span> <span class="bu">open</span>(image, <span class="st">"rb"</span>) <span class="im">as</span> f:</span>
<span id="cb7-18"><a href="#cb7-18" aria-hidden="true" tabindex="-1"></a>            image_base64 <span class="op">=</span> base64.b64encode(f.read()).decode(<span class="st">'utf-8'</span>)</span>
<span id="cb7-19"><a href="#cb7-19" aria-hidden="true" tabindex="-1"></a>    <span class="co">#</span></span>
<span id="cb7-20"><a href="#cb7-20" aria-hidden="true" tabindex="-1"></a>    <span class="co"># Query</span></span>
<span id="cb7-21"><a href="#cb7-21" aria-hidden="true" tabindex="-1"></a>    response <span class="op">=</span> client.chat.completions.create(</span>
<span id="cb7-22"><a href="#cb7-22" aria-hidden="true" tabindex="-1"></a>        model<span class="op">=</span>model,</span>
<span id="cb7-23"><a href="#cb7-23" aria-hidden="true" tabindex="-1"></a>        messages<span class="op">=</span>[{</span>
<span id="cb7-24"><a href="#cb7-24" aria-hidden="true" tabindex="-1"></a>            <span class="st">"role"</span>: <span class="st">"user"</span>,</span>
<span id="cb7-25"><a href="#cb7-25" aria-hidden="true" tabindex="-1"></a>            <span class="st">"content"</span>: [</span>
<span id="cb7-26"><a href="#cb7-26" aria-hidden="true" tabindex="-1"></a>                {<span class="st">"type"</span>: <span class="st">"text"</span>, <span class="st">"text"</span>: prompt},</span>
<span id="cb7-27"><a href="#cb7-27" aria-hidden="true" tabindex="-1"></a>                {<span class="st">"type"</span>: <span class="st">"image_url"</span>, <span class="st">"image_url"</span>: {<span class="st">"url"</span>: <span class="ss">f"data:image/jpeg;base64,</span><span class="sc">{</span>image_base64<span class="sc">}</span><span class="ss">"</span>}}</span>
<span id="cb7-28"><a href="#cb7-28" aria-hidden="true" tabindex="-1"></a>            ]</span>
<span id="cb7-29"><a href="#cb7-29" aria-hidden="true" tabindex="-1"></a>        }]</span>
<span id="cb7-30"><a href="#cb7-30" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb7-31"><a href="#cb7-31" aria-hidden="true" tabindex="-1"></a>    <span class="cf">return</span> response.choices[<span class="dv">0</span>].message.content</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
</details>
</div>
</section>
<section id="simple-vlm-query-example" class="level2" data-number="3.5">
<h2 data-number="3.5" class="anchored" data-anchor-id="simple-vlm-query-example"><span class="header-section-number">3.5</span> 102: Simple VLM Query Example</h2>
<p>To get started let’s do a simple query to describe an image from the dataset.</p>
<div id="9b70b918" class="cell" data-cache="true" data-execution_count="7">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb8"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb8-1"><a href="#cb8-1" aria-hidden="true" tabindex="-1"></a>image <span class="op">=</span> ds[<span class="dv">0</span>][<span class="st">"image"</span>]</span>
<span id="cb8-2"><a href="#cb8-2" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb8-3"><a href="#cb8-3" aria-hidden="true" tabindex="-1"></a><span class="co"># Query the VLM to describe the image</span></span>
<span id="cb8-4"><a href="#cb8-4" aria-hidden="true" tabindex="-1"></a>description <span class="op">=</span> query_image(image, <span class="st">"Describe this image."</span>, model<span class="op">=</span><span class="st">'Qwen/Qwen3-VL-8B-Instruct'</span>)</span>
<span id="cb8-5"><a href="#cb8-5" aria-hidden="true" tabindex="-1"></a>rprint(description)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display">
<pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">This is a black-and-white image of a form from The British Library's Reprographic Section, used to request a copy 
of a manuscript or archival material.

Here is a breakdown of the information on the form:

*   **Institution:** The British Library, Reference Division, Reprographic Section.
*   **Address:** Great Russell Street, London WC1B 3DG.
*   **Department:** Manuscripts.
*   **Shelfmark:** SLOANE <span style="color: #008080; text-decoration-color: #008080; font-weight: bold">3972.</span>C. <span style="font-weight: bold">(</span>Vol. <span style="color: #008080; text-decoration-color: #008080; font-weight: bold">1</span><span style="font-weight: bold">)</span>
*   **Order Number:** SCH NO <span style="color: #008080; text-decoration-color: #008080; font-weight: bold">98876</span>
*   **Author:** This field is blank.
*   **Title:** CATALOGUE OF SIR HANS SLOANES LIBRARY
*   **Place and date of publication:** This field is blank.
*   **Scale/Reduction:** The form indicates a reduction of <span style="color: #008080; text-decoration-color: #008080; font-weight: bold">12</span>, meaning the reproduction will be scaled down to 
<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">1</span>/12th of the original size. A ruler scale is provided for reference in centimetres and inches.

The form appears to be filled out by hand for a specific item: Volume <span style="color: #008080; text-decoration-color: #008080; font-weight: bold">1</span> of the <span style="color: #008000; text-decoration-color: #008000">"Catalogue of Sir Hans Sloane's </span>
<span style="color: #008000; text-decoration-color: #008000">Library,"</span> which is held in the Manuscripts department under the shelfmark SLOANE <span style="color: #008080; text-decoration-color: #008080; font-weight: bold">3972.</span>C. This catalogue was 
compiled by Sir Hans Sloane himself and is a significant historical document detailing his extensive collection, 
which later formed part of the foundation of the British Museum and now resides at the British Library.
</pre>
</div>
</div>
<p>We can see we get a fairly useful description of the card. If we compare against the image we can see most of the details it mentions appear to be largely correct.</p>
<div id="47752997" class="cell" data-execution_count="8">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb9"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb9-1"><a href="#cb9-1" aria-hidden="true" tabindex="-1"></a>image</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="8">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-9-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
<p>There are workflows where open ended description like this could be useful but this isn’t usually the kind of format we want if we want to take some action or do something based on the predictions of the model. In these cases it’s usually nice to have some more controlled output, for example, a label.</p>
</section>
<section id="classification" class="level2" data-number="3.6">
<h2 data-number="3.6" class="anchored" data-anchor-id="classification"><span class="header-section-number">3.6</span> Classification</h2>
<!-- TODO maybe we have a "chapter" or "section" with tasks which we can then embedded across other parts of the book? -->
<p>We’ll define a fairly simple prompt that asks the VLM to decide if a page is one of three categories. We describe each of these categopries and then ask the model to only return one of these as the output. We’ll do this for ten examples and we’ll also log how long it’s taking.</p>
<div id="588f0dd9" class="cell" data-cache="true" data-execution_count="9">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb10"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb10-1"><a href="#cb10-1" aria-hidden="true" tabindex="-1"></a><span class="im">import</span> time</span>
<span id="cb10-2"><a href="#cb10-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> tqdm.auto <span class="im">import</span> tqdm</span>
<span id="cb10-3"><a href="#cb10-3" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> rich <span class="im">import</span> <span class="bu">print</span> <span class="im">as</span> rprint</span>
<span id="cb10-4"><a href="#cb10-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-5"><a href="#cb10-5" aria-hidden="true" tabindex="-1"></a>sample_size <span class="op">=</span> <span class="dv">10</span></span>
<span id="cb10-6"><a href="#cb10-6" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-7"><a href="#cb10-7" aria-hidden="true" tabindex="-1"></a>sample <span class="op">=</span> ds.take(sample_size)</span>
<span id="cb10-8"><a href="#cb10-8" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-9"><a href="#cb10-9" aria-hidden="true" tabindex="-1"></a>prompt <span class="op">=</span> <span class="st">"""Classify this image into one of the following categories:</span></span>
<span id="cb10-10"><a href="#cb10-10" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-11"><a href="#cb10-11" aria-hidden="true" tabindex="-1"></a><span class="st">1. **Index/Reference Card**: A library catalog or reference card</span></span>
<span id="cb10-12"><a href="#cb10-12" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-13"><a href="#cb10-13" aria-hidden="true" tabindex="-1"></a><span class="st">2. **Manuscript Page**: A handwritten or historical document page</span></span>
<span id="cb10-14"><a href="#cb10-14" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-15"><a href="#cb10-15" aria-hidden="true" tabindex="-1"></a><span class="st">3. **Other**: Any document that doesn't fit the above categories</span></span>
<span id="cb10-16"><a href="#cb10-16" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-17"><a href="#cb10-17" aria-hidden="true" tabindex="-1"></a><span class="st">Examine the overall structure, layout, and content type to determine the classification. Focus on whether the document is a structured catalog/reference tool (Index Card) or a historical manuscript with continuous text (Manuscript Page).</span></span>
<span id="cb10-18"><a href="#cb10-18" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-19"><a href="#cb10-19" aria-hidden="true" tabindex="-1"></a><span class="st">Return only the category name: "Index/Reference Card", "Manuscript Page", or "Other"</span></span>
<span id="cb10-20"><a href="#cb10-20" aria-hidden="true" tabindex="-1"></a><span class="st">"""</span></span>
<span id="cb10-21"><a href="#cb10-21" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb10-22"><a href="#cb10-22" aria-hidden="true" tabindex="-1"></a>results <span class="op">=</span> []</span>
<span id="cb10-23"><a href="#cb10-23" aria-hidden="true" tabindex="-1"></a><span class="co"># Time the execution using standard Python</span></span>
<span id="cb10-24"><a href="#cb10-24" aria-hidden="true" tabindex="-1"></a>start_time <span class="op">=</span> time.time()</span>
<span id="cb10-25"><a href="#cb10-25" aria-hidden="true" tabindex="-1"></a><span class="cf">for</span> row <span class="kw">in</span> tqdm(sample):</span>
<span id="cb10-26"><a href="#cb10-26" aria-hidden="true" tabindex="-1"></a>    image <span class="op">=</span> row[<span class="st">'image'</span>]</span>
<span id="cb10-27"><a href="#cb10-27" aria-hidden="true" tabindex="-1"></a>    results.append(query_image(image, prompt, model<span class="op">=</span><span class="st">'Qwen/Qwen3-VL-8B-Instruct'</span>))</span>
<span id="cb10-28"><a href="#cb10-28" aria-hidden="true" tabindex="-1"></a>elapsed_time <span class="op">=</span> time.time() <span class="op">-</span> start_time</span>
<span id="cb10-29"><a href="#cb10-29" aria-hidden="true" tabindex="-1"></a><span class="bu">print</span>(<span class="ss">f"Execution time: </span><span class="sc">{</span>elapsed_time<span class="sc">:.2f}</span><span class="ss"> seconds"</span>)</span>
<span id="cb10-30"><a href="#cb10-30" aria-hidden="true" tabindex="-1"></a>rprint(results)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display">
<script type="application/vnd.jupyter.widget-view+json">
{"model_id":"7c6477b5384d4da3837c3ae504d452ba","version_major":2,"version_minor":0,"quarto_mimetype":"application/vnd.jupyter.widget-view+json"}
</script>
</div>
<div class="cell-output cell-output-stdout">
<pre><code>Execution time: 18.31 seconds</code></pre>
</div>
<div class="cell-output cell-output-display">
<pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"><span style="font-weight: bold">[</span>
    <span style="color: #008000; text-decoration-color: #008000">'Index/Reference Card'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Other'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>,
    <span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span>
<span style="font-weight: bold">]</span>
</pre>
</div>
</div>
<p>Let’s check the result that was predicted as “index/reference card”</p>
<div id="25e8df82" class="cell" data-execution_count="10">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb12"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb12-1"><a href="#cb12-1" aria-hidden="true" tabindex="-1"></a>sample[<span class="dv">0</span>][<span class="st">'image'</span>]</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="10">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-11-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
<p>We can extrapolate how long this would take for the full dataset</p>
<div id="c4891945" class="cell" data-cache="true" data-execution_count="11">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb13"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb13-1"><a href="#cb13-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Calculate average time per image</span></span>
<span id="cb13-2"><a href="#cb13-2" aria-hidden="true" tabindex="-1"></a>avg_time_per_image <span class="op">=</span> elapsed_time <span class="op">/</span> sample_size</span>
<span id="cb13-3"><a href="#cb13-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb13-4"><a href="#cb13-4" aria-hidden="true" tabindex="-1"></a><span class="co"># Project time for full dataset</span></span>
<span id="cb13-5"><a href="#cb13-5" aria-hidden="true" tabindex="-1"></a>total_images <span class="op">=</span> <span class="bu">len</span>(ds)</span>
<span id="cb13-6"><a href="#cb13-6" aria-hidden="true" tabindex="-1"></a>projected_time <span class="op">=</span> avg_time_per_image <span class="op">*</span> total_images</span>
<span id="cb13-7"><a href="#cb13-7" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb13-8"><a href="#cb13-8" aria-hidden="true" tabindex="-1"></a><span class="bu">print</span>(<span class="ss">f"Sample processing time: </span><span class="sc">{</span>elapsed_time<span class="sc">:.2f}</span><span class="ss"> seconds (</span><span class="sc">{</span>elapsed_time<span class="op">/</span><span class="dv">60</span><span class="sc">:.2f}</span><span class="ss"> minutes)"</span>)</span>
<span id="cb13-9"><a href="#cb13-9" aria-hidden="true" tabindex="-1"></a><span class="bu">print</span>(<span class="ss">f"Average time per image: </span><span class="sc">{</span>avg_time_per_image<span class="sc">:.2f}</span><span class="ss"> seconds"</span>)</span>
<span id="cb13-10"><a href="#cb13-10" aria-hidden="true" tabindex="-1"></a><span class="bu">print</span>(<span class="ss">f"Total images in dataset: </span><span class="sc">{</span>total_images<span class="sc">}</span><span class="ss">"</span>)</span>
<span id="cb13-11"><a href="#cb13-11" aria-hidden="true" tabindex="-1"></a><span class="bu">print</span>(<span class="ss">f"Projected time for full dataset: </span><span class="sc">{</span>projected_time<span class="op">/</span><span class="dv">60</span><span class="sc">:.2f}</span><span class="ss"> minutes (</span><span class="sc">{</span>projected_time<span class="op">/</span><span class="dv">3600</span><span class="sc">:.2f}</span><span class="ss"> hours)"</span>)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-stdout">
<pre><code>Sample processing time: 18.31 seconds (0.31 minutes)
Average time per image: 1.83 seconds
Total images in dataset: 2734
Projected time for full dataset: 83.45 minutes (1.39 hours)</code></pre>
</div>
</div>
<section id="classifying-with-structured-labels" class="level3" data-number="3.6.1">
<h3 data-number="3.6.1" class="anchored" data-anchor-id="classifying-with-structured-labels"><span class="header-section-number">3.6.1</span> Classifying with structured labels</h3>
<p>In the previous example, we relied on the model to return the label in the correct format. While this often works, it can sometimes lead to inconsistencies in the output. To address this, we can use Pydantic models to define a structured output format. This way, we can ensure that the output adheres to a specific schema.</p>
<p>In this example, we’ll define a Pydantic model for our classification task. The model will have a single field <code>category</code> which can take one of three literal values <code>"Index/Reference Card"</code>, <code>"Manuscript Page"</code>, or <code>"other"</code>.</p>
<p>What this means in practice is that the model will only be able to return one of these three values for the <code>category</code> field.</p>
<div id="907e5dd9" class="cell" data-execution_count="12">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb15"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb15-1"><a href="#cb15-1" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> pydantic <span class="im">import</span> BaseModel, Field</span>
<span id="cb15-2"><a href="#cb15-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> typing <span class="im">import</span> Literal</span>
<span id="cb15-3"><a href="#cb15-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb15-4"><a href="#cb15-4" aria-hidden="true" tabindex="-1"></a><span class="kw">class</span> PageCategory(BaseModel):</span>
<span id="cb15-5"><a href="#cb15-5" aria-hidden="true" tabindex="-1"></a>    category: Literal[<span class="st">"Index/Reference Card"</span>, <span class="st">"Manuscript Page"</span>, <span class="st">"other"</span>] <span class="op">=</span> Field(</span>
<span id="cb15-6"><a href="#cb15-6" aria-hidden="true" tabindex="-1"></a>        ..., description<span class="op">=</span><span class="st">"The category of the image"</span></span>
<span id="cb15-7"><a href="#cb15-7" aria-hidden="true" tabindex="-1"></a>    )</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
</div>
<p>When using the OpenAI client we can specify this Pydantic model as the <code>response_format</code> when making the request. This tells the model to return the output in a format that can be parsed into the Pydantic model (the APIs for this are still evolving so may change slightly over time).</p>
<div id="166e0d6b" class="cell" data-execution_count="13">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb16"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb16-1"><a href="#cb16-1" aria-hidden="true" tabindex="-1"></a>buffered <span class="op">=</span> BytesIO()</span>
<span id="cb16-2"><a href="#cb16-2" aria-hidden="true" tabindex="-1"></a>image.save(buffered, <span class="bu">format</span><span class="op">=</span><span class="st">"JPEG"</span>)</span>
<span id="cb16-3"><a href="#cb16-3" aria-hidden="true" tabindex="-1"></a>image_base64 <span class="op">=</span> base64.b64encode(buffered.getvalue()).decode(<span class="st">'utf-8'</span>)</span>
<span id="cb16-4"><a href="#cb16-4" aria-hidden="true" tabindex="-1"></a>completion <span class="op">=</span> client.beta.chat.completions.parse(</span>
<span id="cb16-5"><a href="#cb16-5" aria-hidden="true" tabindex="-1"></a>    model<span class="op">=</span><span class="st">"Qwen/Qwen3-VL-8B-Instruct"</span>,</span>
<span id="cb16-6"><a href="#cb16-6" aria-hidden="true" tabindex="-1"></a>    messages<span class="op">=</span>[</span>
<span id="cb16-7"><a href="#cb16-7" aria-hidden="true" tabindex="-1"></a>         {</span>
<span id="cb16-8"><a href="#cb16-8" aria-hidden="true" tabindex="-1"></a>            <span class="st">"role"</span>: <span class="st">"user"</span>,</span>
<span id="cb16-9"><a href="#cb16-9" aria-hidden="true" tabindex="-1"></a>            <span class="st">"content"</span>: [</span>
<span id="cb16-10"><a href="#cb16-10" aria-hidden="true" tabindex="-1"></a>                {</span>
<span id="cb16-11"><a href="#cb16-11" aria-hidden="true" tabindex="-1"></a>                    <span class="st">"type"</span>: <span class="st">"text"</span>,</span>
<span id="cb16-12"><a href="#cb16-12" aria-hidden="true" tabindex="-1"></a>                    <span class="st">"text"</span>: prompt,</span>
<span id="cb16-13"><a href="#cb16-13" aria-hidden="true" tabindex="-1"></a>                },</span>
<span id="cb16-14"><a href="#cb16-14" aria-hidden="true" tabindex="-1"></a>                {</span>
<span id="cb16-15"><a href="#cb16-15" aria-hidden="true" tabindex="-1"></a>                    <span class="st">"type"</span>: <span class="st">"image_url"</span>,</span>
<span id="cb16-16"><a href="#cb16-16" aria-hidden="true" tabindex="-1"></a>                    <span class="st">"image_url"</span>: {<span class="st">"url"</span>: <span class="ss">f"data:image/jpeg;base64,</span><span class="sc">{</span>image_base64<span class="sc">}</span><span class="ss">"</span>},</span>
<span id="cb16-17"><a href="#cb16-17" aria-hidden="true" tabindex="-1"></a>                },</span>
<span id="cb16-18"><a href="#cb16-18" aria-hidden="true" tabindex="-1"></a>            ],</span>
<span id="cb16-19"><a href="#cb16-19" aria-hidden="true" tabindex="-1"></a>        },</span>
<span id="cb16-20"><a href="#cb16-20" aria-hidden="true" tabindex="-1"></a>    ],</span>
<span id="cb16-21"><a href="#cb16-21" aria-hidden="true" tabindex="-1"></a>    max_tokens<span class="op">=</span><span class="dv">200</span>,</span>
<span id="cb16-22"><a href="#cb16-22" aria-hidden="true" tabindex="-1"></a>    temperature<span class="op">=</span><span class="fl">0.7</span>,</span>
<span id="cb16-23"><a href="#cb16-23" aria-hidden="true" tabindex="-1"></a>    response_format<span class="op">=</span>PageCategory,</span>
<span id="cb16-24"><a href="#cb16-24" aria-hidden="true" tabindex="-1"></a>)</span>
<span id="cb16-25"><a href="#cb16-25" aria-hidden="true" tabindex="-1"></a>rprint(completion)</span>
<span id="cb16-26"><a href="#cb16-26" aria-hidden="true" tabindex="-1"></a>rprint(completion.choices[<span class="dv">0</span>].message.parsed)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display">
<pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">ParsedChatCompletion<span style="font-weight: bold">[</span>PageCategory<span style="font-weight: bold">](</span>
    <span style="color: #808000; text-decoration-color: #808000">id</span>=<span style="color: #008000; text-decoration-color: #008000">'47c1a911111046291f6bbec65ebb1ead'</span>,
    <span style="color: #808000; text-decoration-color: #808000">choices</span>=<span style="font-weight: bold">[</span>
        ParsedChoice<span style="font-weight: bold">[</span>PageCategory<span style="font-weight: bold">](</span>
            <span style="color: #808000; text-decoration-color: #808000">finish_reason</span>=<span style="color: #008000; text-decoration-color: #008000">'stop'</span>,
            <span style="color: #808000; text-decoration-color: #808000">index</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">logprobs</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
            <span style="color: #808000; text-decoration-color: #808000">message</span>=<span style="color: #800080; text-decoration-color: #800080">ParsedChatCompletionMessage</span><span style="font-weight: bold">[</span>PageCategory<span style="font-weight: bold">](</span>
                <span style="color: #808000; text-decoration-color: #808000">content</span>=<span style="color: #008000; text-decoration-color: #008000">'{\n  "category": "Manuscript Page"\n}'</span>,
                <span style="color: #808000; text-decoration-color: #808000">refusal</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
                <span style="color: #808000; text-decoration-color: #808000">role</span>=<span style="color: #008000; text-decoration-color: #008000">'assistant'</span>,
                <span style="color: #808000; text-decoration-color: #808000">annotations</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
                <span style="color: #808000; text-decoration-color: #808000">audio</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
                <span style="color: #808000; text-decoration-color: #808000">function_call</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
                <span style="color: #808000; text-decoration-color: #808000">tool_calls</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
                <span style="color: #808000; text-decoration-color: #808000">parsed</span>=<span style="color: #800080; text-decoration-color: #800080; font-weight: bold">PageCategory</span><span style="font-weight: bold">(</span><span style="color: #808000; text-decoration-color: #808000">category</span>=<span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span><span style="font-weight: bold">)</span>
            <span style="font-weight: bold">)</span>
        <span style="font-weight: bold">)</span>
    <span style="font-weight: bold">]</span>,
    <span style="color: #808000; text-decoration-color: #808000">created</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">1771255803</span>,
    <span style="color: #808000; text-decoration-color: #808000">model</span>=<span style="color: #008000; text-decoration-color: #008000">'qwen/qwen3-vl-8b-instruct'</span>,
    <span style="color: #808000; text-decoration-color: #808000">object</span>=<span style="color: #008000; text-decoration-color: #008000">'chat.completion'</span>,
    <span style="color: #808000; text-decoration-color: #808000">service_tier</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
    <span style="color: #808000; text-decoration-color: #808000">system_fingerprint</span>=<span style="color: #008000; text-decoration-color: #008000">''</span>,
    <span style="color: #808000; text-decoration-color: #808000">usage</span>=<span style="color: #800080; text-decoration-color: #800080; font-weight: bold">CompletionUsage</span><span style="font-weight: bold">(</span>
        <span style="color: #808000; text-decoration-color: #808000">completion_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">13</span>,
        <span style="color: #808000; text-decoration-color: #808000">prompt_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">966</span>,
        <span style="color: #808000; text-decoration-color: #808000">total_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">979</span>,
        <span style="color: #808000; text-decoration-color: #808000">completion_tokens_details</span>=<span style="color: #800080; text-decoration-color: #800080; font-weight: bold">CompletionTokensDetails</span><span style="font-weight: bold">(</span>
            <span style="color: #808000; text-decoration-color: #808000">accepted_prediction_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">audio_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">reasoning_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">rejected_prediction_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">text_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">13</span>,
            <span style="color: #808000; text-decoration-color: #808000">image_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">video_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>
        <span style="font-weight: bold">)</span>,
        <span style="color: #808000; text-decoration-color: #808000">prompt_tokens_details</span>=<span style="color: #800080; text-decoration-color: #800080; font-weight: bold">PromptTokensDetails</span><span style="font-weight: bold">(</span>
            <span style="color: #808000; text-decoration-color: #808000">audio_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">cached_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">cache_creation_input_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">cache_read_input_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>,
            <span style="color: #808000; text-decoration-color: #808000">text_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">228</span>,
            <span style="color: #808000; text-decoration-color: #808000">image_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">738</span>,
            <span style="color: #808000; text-decoration-color: #808000">video_tokens</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">0</span>
        <span style="font-weight: bold">)</span>
    <span style="font-weight: bold">)</span>
<span style="font-weight: bold">)</span>
</pre>
</div>
<div class="cell-output cell-output-display">
<pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"><span style="color: #800080; text-decoration-color: #800080; font-weight: bold">PageCategory</span><span style="font-weight: bold">(</span><span style="color: #808000; text-decoration-color: #808000">category</span>=<span style="color: #008000; text-decoration-color: #008000">'Manuscript Page'</span><span style="font-weight: bold">)</span>
</pre>
</div>
</div>
<div id="16d63ec2" class="cell" data-execution_count="14">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb17"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb17-1"><a href="#cb17-1" aria-hidden="true" tabindex="-1"></a>image</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="14">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-15-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
</section>
</section>
<section id="beyond-classifying---extracting-structured-information" class="level2" data-number="3.7">
<h2 data-number="3.7" class="anchored" data-anchor-id="beyond-classifying---extracting-structured-information"><span class="header-section-number">3.7</span> Beyond classifying - Extracting structured information</h2>
<p>So far we’ve focused on classifying images but what if we want to extract information from the images? Let’s take the first example from the dataset again.</p>
<div id="bba089db" class="cell" data-cache="true" data-execution_count="15">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb18"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb18-1"><a href="#cb18-1" aria-hidden="true" tabindex="-1"></a>index_image <span class="op">=</span> ds[<span class="dv">0</span>][<span class="st">'image'</span>]</span>
<span id="cb18-2"><a href="#cb18-2" aria-hidden="true" tabindex="-1"></a>index_image</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="15">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-16-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
<p>If we have an image like this we don’t just want to assign a label from it (we may do this as a first step) we actually want to extract the various fields from the card in a structured way. We can again use a Pydantic model to define the structure of the data we want to extract.</p>
<div id="93efb7a1" class="cell" data-execution_count="16">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb19"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb19-1"><a href="#cb19-1" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> pydantic <span class="im">import</span> BaseModel, Field</span>
<span id="cb19-2"><a href="#cb19-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> typing <span class="im">import</span> Optional</span>
<span id="cb19-3"><a href="#cb19-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb19-4"><a href="#cb19-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb19-5"><a href="#cb19-5" aria-hidden="true" tabindex="-1"></a><span class="kw">class</span> BritishLibraryReprographicCard(BaseModel):</span>
<span id="cb19-6"><a href="#cb19-6" aria-hidden="true" tabindex="-1"></a>    <span class="co">"""</span></span>
<span id="cb19-7"><a href="#cb19-7" aria-hidden="true" tabindex="-1"></a><span class="co">    Pydantic model for extracting information from British Library Reference Division </span></span>
<span id="cb19-8"><a href="#cb19-8" aria-hidden="true" tabindex="-1"></a><span class="co">    reprographic cards used to document manuscripts and other materials.</span></span>
<span id="cb19-9"><a href="#cb19-9" aria-hidden="true" tabindex="-1"></a><span class="co">    """</span></span>
<span id="cb19-10"><a href="#cb19-10" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb19-11"><a href="#cb19-11" aria-hidden="true" tabindex="-1"></a>    department: <span class="bu">str</span> <span class="op">=</span> Field(</span>
<span id="cb19-12"><a href="#cb19-12" aria-hidden="true" tabindex="-1"></a>        ..., </span>
<span id="cb19-13"><a href="#cb19-13" aria-hidden="true" tabindex="-1"></a>        description<span class="op">=</span><span class="st">"The division that holds the material (e.g., 'MANUSCRIPTS')"</span></span>
<span id="cb19-14"><a href="#cb19-14" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb19-15"><a href="#cb19-15" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb19-16"><a href="#cb19-16" aria-hidden="true" tabindex="-1"></a>    shelfmark: <span class="bu">str</span> <span class="op">=</span> Field(</span>
<span id="cb19-17"><a href="#cb19-17" aria-hidden="true" tabindex="-1"></a>        ..., </span>
<span id="cb19-18"><a href="#cb19-18" aria-hidden="true" tabindex="-1"></a>        description<span class="op">=</span><span class="st">"The library's classification/location code (e.g., 'SLOANE 3972.C. (VOL 1)')"</span></span>
<span id="cb19-19"><a href="#cb19-19" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb19-20"><a href="#cb19-20" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb19-21"><a href="#cb19-21" aria-hidden="true" tabindex="-1"></a>    order: <span class="bu">str</span> <span class="op">=</span> Field(</span>
<span id="cb19-22"><a href="#cb19-22" aria-hidden="true" tabindex="-1"></a>        ..., </span>
<span id="cb19-23"><a href="#cb19-23" aria-hidden="true" tabindex="-1"></a>        description<span class="op">=</span><span class="st">"Order reference, typically starting with 'SCH NO' followed by numbers"</span></span>
<span id="cb19-24"><a href="#cb19-24" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb19-25"><a href="#cb19-25" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb19-26"><a href="#cb19-26" aria-hidden="true" tabindex="-1"></a>    author: Optional[<span class="bu">str</span>] <span class="op">=</span> Field(</span>
<span id="cb19-27"><a href="#cb19-27" aria-hidden="true" tabindex="-1"></a>        <span class="va">None</span>, </span>
<span id="cb19-28"><a href="#cb19-28" aria-hidden="true" tabindex="-1"></a>        description<span class="op">=</span><span class="st">"Author name if present, null if blank or marked with diagonal line"</span></span>
<span id="cb19-29"><a href="#cb19-29" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb19-30"><a href="#cb19-30" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb19-31"><a href="#cb19-31" aria-hidden="true" tabindex="-1"></a>    title: <span class="bu">str</span> <span class="op">=</span> Field(</span>
<span id="cb19-32"><a href="#cb19-32" aria-hidden="true" tabindex="-1"></a>        ..., </span>
<span id="cb19-33"><a href="#cb19-33" aria-hidden="true" tabindex="-1"></a>        description<span class="op">=</span><span class="st">"The name of the work or manuscript"</span></span>
<span id="cb19-34"><a href="#cb19-34" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb19-35"><a href="#cb19-35" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb19-36"><a href="#cb19-36" aria-hidden="true" tabindex="-1"></a>    place_and_date_of_publication: Optional[<span class="bu">str</span>] <span class="op">=</span> Field(</span>
<span id="cb19-37"><a href="#cb19-37" aria-hidden="true" tabindex="-1"></a>        <span class="va">None</span>, </span>
<span id="cb19-38"><a href="#cb19-38" aria-hidden="true" tabindex="-1"></a>        description<span class="op">=</span><span class="st">"Place and date of publication if present, null if blank"</span></span>
<span id="cb19-39"><a href="#cb19-39" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb19-40"><a href="#cb19-40" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb19-41"><a href="#cb19-41" aria-hidden="true" tabindex="-1"></a>    reduction: <span class="bu">int</span> <span class="op">=</span> Field(</span>
<span id="cb19-42"><a href="#cb19-42" aria-hidden="true" tabindex="-1"></a>        ..., </span>
<span id="cb19-43"><a href="#cb19-43" aria-hidden="true" tabindex="-1"></a>        description<span class="op">=</span><span class="st">"The reduction number shown at the bottom of the card"</span></span>
<span id="cb19-44"><a href="#cb19-44" aria-hidden="true" tabindex="-1"></a>    )</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
</div>
<p>We’ll now create a function to handle the querying process using this structured schema.</p>
<div id="617aedff" class="cell" data-execution_count="17">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb20"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb20-1"><a href="#cb20-1" aria-hidden="true" tabindex="-1"></a><span class="kw">def</span> query_image_structured(image, prompt, schema, model<span class="op">=</span><span class="st">'Qwen/Qwen3-VL-8B-Instruct'</span>):</span>
<span id="cb20-2"><a href="#cb20-2" aria-hidden="true" tabindex="-1"></a>    <span class="co">"""</span></span>
<span id="cb20-3"><a href="#cb20-3" aria-hidden="true" tabindex="-1"></a><span class="co">    Query VLM with an image and get structured output based on a Pydantic schema.</span></span>
<span id="cb20-4"><a href="#cb20-4" aria-hidden="true" tabindex="-1"></a><span class="co">    </span></span>
<span id="cb20-5"><a href="#cb20-5" aria-hidden="true" tabindex="-1"></a><span class="co">    Args:</span></span>
<span id="cb20-6"><a href="#cb20-6" aria-hidden="true" tabindex="-1"></a><span class="co">        image: PIL Image or file path to the image</span></span>
<span id="cb20-7"><a href="#cb20-7" aria-hidden="true" tabindex="-1"></a><span class="co">        prompt: Text prompt describing what to extract</span></span>
<span id="cb20-8"><a href="#cb20-8" aria-hidden="true" tabindex="-1"></a><span class="co">        schema: Pydantic model class defining the expected output structure</span></span>
<span id="cb20-9"><a href="#cb20-9" aria-hidden="true" tabindex="-1"></a><span class="co">        model: Model ID to use for the query</span></span>
<span id="cb20-10"><a href="#cb20-10" aria-hidden="true" tabindex="-1"></a><span class="co">    </span></span>
<span id="cb20-11"><a href="#cb20-11" aria-hidden="true" tabindex="-1"></a><span class="co">    Returns:</span></span>
<span id="cb20-12"><a href="#cb20-12" aria-hidden="true" tabindex="-1"></a><span class="co">        Parsed Pydantic model instance with the extracted data</span></span>
<span id="cb20-13"><a href="#cb20-13" aria-hidden="true" tabindex="-1"></a><span class="co">    """</span></span>
<span id="cb20-14"><a href="#cb20-14" aria-hidden="true" tabindex="-1"></a>    <span class="co"># Convert image to base64</span></span>
<span id="cb20-15"><a href="#cb20-15" aria-hidden="true" tabindex="-1"></a>    <span class="cf">if</span> <span class="bu">isinstance</span>(image, PILImage):</span>
<span id="cb20-16"><a href="#cb20-16" aria-hidden="true" tabindex="-1"></a>        buffered <span class="op">=</span> BytesIO()</span>
<span id="cb20-17"><a href="#cb20-17" aria-hidden="true" tabindex="-1"></a>        image.save(buffered, <span class="bu">format</span><span class="op">=</span><span class="st">"JPEG"</span>)</span>
<span id="cb20-18"><a href="#cb20-18" aria-hidden="true" tabindex="-1"></a>        image_base64 <span class="op">=</span> base64.b64encode(buffered.getvalue()).decode(<span class="st">'utf-8'</span>)</span>
<span id="cb20-19"><a href="#cb20-19" aria-hidden="true" tabindex="-1"></a>    <span class="cf">else</span>:</span>
<span id="cb20-20"><a href="#cb20-20" aria-hidden="true" tabindex="-1"></a>        <span class="cf">with</span> <span class="bu">open</span>(image, <span class="st">"rb"</span>) <span class="im">as</span> f:</span>
<span id="cb20-21"><a href="#cb20-21" aria-hidden="true" tabindex="-1"></a>            image_base64 <span class="op">=</span> base64.b64encode(f.read()).decode(<span class="st">'utf-8'</span>)</span>
<span id="cb20-22"><a href="#cb20-22" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb20-23"><a href="#cb20-23" aria-hidden="true" tabindex="-1"></a>    <span class="co"># Query with structured output</span></span>
<span id="cb20-24"><a href="#cb20-24" aria-hidden="true" tabindex="-1"></a>    completion <span class="op">=</span> client.beta.chat.completions.parse(</span>
<span id="cb20-25"><a href="#cb20-25" aria-hidden="true" tabindex="-1"></a>        model<span class="op">=</span>model,</span>
<span id="cb20-26"><a href="#cb20-26" aria-hidden="true" tabindex="-1"></a>        messages<span class="op">=</span>[{</span>
<span id="cb20-27"><a href="#cb20-27" aria-hidden="true" tabindex="-1"></a>            <span class="st">"role"</span>: <span class="st">"user"</span>,</span>
<span id="cb20-28"><a href="#cb20-28" aria-hidden="true" tabindex="-1"></a>            <span class="st">"content"</span>: [</span>
<span id="cb20-29"><a href="#cb20-29" aria-hidden="true" tabindex="-1"></a>                {<span class="st">"type"</span>: <span class="st">"text"</span>, <span class="st">"text"</span>: prompt},</span>
<span id="cb20-30"><a href="#cb20-30" aria-hidden="true" tabindex="-1"></a>                {<span class="st">"type"</span>: <span class="st">"image_url"</span>, <span class="st">"image_url"</span>: {<span class="st">"url"</span>: <span class="ss">f"data:image/jpeg;base64,</span><span class="sc">{</span>image_base64<span class="sc">}</span><span class="ss">"</span>}}</span>
<span id="cb20-31"><a href="#cb20-31" aria-hidden="true" tabindex="-1"></a>            ]</span>
<span id="cb20-32"><a href="#cb20-32" aria-hidden="true" tabindex="-1"></a>        }],</span>
<span id="cb20-33"><a href="#cb20-33" aria-hidden="true" tabindex="-1"></a>        response_format<span class="op">=</span>schema,</span>
<span id="cb20-34"><a href="#cb20-34" aria-hidden="true" tabindex="-1"></a>        temperature<span class="op">=</span><span class="fl">0.3</span>  <span class="co"># Lower temperature for more consistent extraction</span></span>
<span id="cb20-35"><a href="#cb20-35" aria-hidden="true" tabindex="-1"></a>    )</span>
<span id="cb20-36"><a href="#cb20-36" aria-hidden="true" tabindex="-1"></a>    </span>
<span id="cb20-37"><a href="#cb20-37" aria-hidden="true" tabindex="-1"></a>    <span class="co"># Return the parsed structured data</span></span>
<span id="cb20-38"><a href="#cb20-38" aria-hidden="true" tabindex="-1"></a>    <span class="cf">return</span> completion.choices[<span class="dv">0</span>].message.parsed</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
</div>
<p>We also need to define a prompt that describes what information we want to extract from the card.</p>
<div id="6cb8e664" class="cell" data-execution_count="18">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb21"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb21-1"><a href="#cb21-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Example usage</span></span>
<span id="cb21-2"><a href="#cb21-2" aria-hidden="true" tabindex="-1"></a>extraction_prompt <span class="op">=</span> <span class="st">"""</span></span>
<span id="cb21-3"><a href="#cb21-3" aria-hidden="true" tabindex="-1"></a><span class="st">Extract the information from this British Library card into structured data (JSON format).</span></span>
<span id="cb21-4"><a href="#cb21-4" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb21-5"><a href="#cb21-5" aria-hidden="true" tabindex="-1"></a><span class="st">Read each field on the card and extract the following information:</span></span>
<span id="cb21-6"><a href="#cb21-6" aria-hidden="true" tabindex="-1"></a><span class="st">- department: The division name (e.g., "MANUSCRIPTS")</span></span>
<span id="cb21-7"><a href="#cb21-7" aria-hidden="true" tabindex="-1"></a><span class="st">- shelfmark: The catalog number (e.g., "SLOANE 3972.C. (VOL 1)")</span></span>
<span id="cb21-8"><a href="#cb21-8" aria-hidden="true" tabindex="-1"></a><span class="st">- order: The SCH NO reference number</span></span>
<span id="cb21-9"><a href="#cb21-9" aria-hidden="true" tabindex="-1"></a><span class="st">- author: The author name, or null if blank</span></span>
<span id="cb21-10"><a href="#cb21-10" aria-hidden="true" tabindex="-1"></a><span class="st">- title: The full title of the work</span></span>
<span id="cb21-11"><a href="#cb21-11" aria-hidden="true" tabindex="-1"></a><span class="st">- place_and_date_of_publication: Publication info, or null if blank</span></span>
<span id="cb21-12"><a href="#cb21-12" aria-hidden="true" tabindex="-1"></a><span class="st">- reduction: The reduction number (as integer) at bottom of card</span></span>
<span id="cb21-13"><a href="#cb21-13" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb21-14"><a href="#cb21-14" aria-hidden="true" tabindex="-1"></a><span class="st">Return the exact text as shown on the card. For empty fields with diagonal lines or no text, use null.</span></span>
<span id="cb21-15"><a href="#cb21-15" aria-hidden="true" tabindex="-1"></a><span class="st">"""</span></span>
<span id="cb21-16"><a href="#cb21-16" aria-hidden="true" tabindex="-1"></a>result <span class="op">=</span> query_image_structured(index_image, extraction_prompt, BritishLibraryReprographicCard)</span>
<span id="cb21-17"><a href="#cb21-17" aria-hidden="true" tabindex="-1"></a>rprint(result)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display">
<pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"><span style="color: #800080; text-decoration-color: #800080; font-weight: bold">BritishLibraryReprographicCard</span><span style="font-weight: bold">(</span>
    <span style="color: #808000; text-decoration-color: #808000">department</span>=<span style="color: #008000; text-decoration-color: #008000">'MANUSCRIPTS'</span>,
    <span style="color: #808000; text-decoration-color: #808000">shelfmark</span>=<span style="color: #008000; text-decoration-color: #008000">'SLOANE 3972.C. (VOL 1)'</span>,
    <span style="color: #808000; text-decoration-color: #808000">order</span>=<span style="color: #008000; text-decoration-color: #008000">'SCH NO 98876'</span>,
    <span style="color: #808000; text-decoration-color: #808000">author</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
    <span style="color: #808000; text-decoration-color: #808000">title</span>=<span style="color: #008000; text-decoration-color: #008000">'CATALOGUE OF SIR HANS SLOANES LIBRARY'</span>,
    <span style="color: #808000; text-decoration-color: #808000">place_and_date_of_publication</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
    <span style="color: #808000; text-decoration-color: #808000">reduction</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">12</span>
<span style="font-weight: bold">)</span>
</pre>
</div>
</div>
<div id="b8ef1639" class="cell" data-execution_count="19">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb22"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb22-1"><a href="#cb22-1" aria-hidden="true" tabindex="-1"></a>rprint(result)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display">
<pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"><span style="color: #800080; text-decoration-color: #800080; font-weight: bold">BritishLibraryReprographicCard</span><span style="font-weight: bold">(</span>
    <span style="color: #808000; text-decoration-color: #808000">department</span>=<span style="color: #008000; text-decoration-color: #008000">'MANUSCRIPTS'</span>,
    <span style="color: #808000; text-decoration-color: #808000">shelfmark</span>=<span style="color: #008000; text-decoration-color: #008000">'SLOANE 3972.C. (VOL 1)'</span>,
    <span style="color: #808000; text-decoration-color: #808000">order</span>=<span style="color: #008000; text-decoration-color: #008000">'SCH NO 98876'</span>,
    <span style="color: #808000; text-decoration-color: #808000">author</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
    <span style="color: #808000; text-decoration-color: #808000">title</span>=<span style="color: #008000; text-decoration-color: #008000">'CATALOGUE OF SIR HANS SLOANES LIBRARY'</span>,
    <span style="color: #808000; text-decoration-color: #808000">place_and_date_of_publication</span>=<span style="color: #800080; text-decoration-color: #800080; font-style: italic">None</span>,
    <span style="color: #808000; text-decoration-color: #808000">reduction</span>=<span style="color: #008080; text-decoration-color: #008080; font-weight: bold">12</span>
<span style="font-weight: bold">)</span>
</pre>
</div>
</div>
<div id="6008255a" class="cell" data-execution_count="20">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb23"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb23-1"><a href="#cb23-1" aria-hidden="true" tabindex="-1"></a>index_image</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
<div class="cell-output cell-output-display" data-execution_count="20">
<div>
<figure class="figure">
<p><img src="vlm-structured-generation_files/figure-html/cell-21-output-1.png" class="img-fluid figure-img"></p>
</figure>
</div>
</div>
</div>
</section>
<section id="appendix-using-a-local-model" class="level2" data-number="3.8">
<h2 data-number="3.8" class="anchored" data-anchor-id="appendix-using-a-local-model"><span class="header-section-number">3.8</span> Appendix: Using a Local Model</h2>
<p>All the code in this chapter uses the OpenAI-compatible API, which means you can swap in a local model server with a single change to the client setup. Everything else — schemas, prompts, <code>.parse()</code> calls — works identically.</p>
<div id="91cc2f2d" class="cell" data-execution_count="21">
<div class="code-copy-outer-scaffold"><div class="sourceCode cell-code" id="cb24"><pre class="sourceCode python code-with-copy"><code class="sourceCode python"><span id="cb24-1"><a href="#cb24-1" aria-hidden="true" tabindex="-1"></a><span class="co"># Replace the HF Inference client with a local server</span></span>
<span id="cb24-2"><a href="#cb24-2" aria-hidden="true" tabindex="-1"></a><span class="im">from</span> openai <span class="im">import</span> OpenAI</span>
<span id="cb24-3"><a href="#cb24-3" aria-hidden="true" tabindex="-1"></a></span>
<span id="cb24-4"><a href="#cb24-4" aria-hidden="true" tabindex="-1"></a>client <span class="op">=</span> OpenAI(</span>
<span id="cb24-5"><a href="#cb24-5" aria-hidden="true" tabindex="-1"></a>    base_url<span class="op">=</span><span class="st">"http://localhost:1234/v1"</span>,  <span class="co"># LM Studio default port</span></span>
<span id="cb24-6"><a href="#cb24-6" aria-hidden="true" tabindex="-1"></a>    api_key<span class="op">=</span><span class="st">"lm-studio"</span>                   <span class="co"># Default API key</span></span>
<span id="cb24-7"><a href="#cb24-7" aria-hidden="true" tabindex="-1"></a>)</span></code></pre></div><button title="Copy to Clipboard" class="code-copy-button"><i class="bi"></i></button></div>
</div>
<p><strong>Popular local options:</strong></p>
<table class="caption-top table">
<thead>
<tr class="header">
<th>Tool</th>
<th>Best for</th>
<th>Notes</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><a href="https://lmstudio.ai/">LM Studio</a></td>
<td>Getting started quickly</td>
<td>GUI-based, MLX acceleration on Mac, built-in model browser</td>
</tr>
<tr class="even">
<td><a href="https://ollama.com/">Ollama</a></td>
<td>CLI workflows</td>
<td>Simple <code>ollama run</code> commands, runs on port 11434</td>
</tr>
<tr class="odd">
<td><a href="https://docs.vllm.ai/">vLLM</a></td>
<td>Production &amp; batch processing</td>
<td>GPU-optimized, highest throughput, best for large-scale extraction</td>
</tr>
</tbody>
</table>
<div class="callout callout-style-default callout-note callout-titled">
<div class="callout-header d-flex align-content-center">
<div class="callout-icon-container">
<i class="callout-icon"></i>
</div>
<div class="callout-title-container flex-fill">
Note
</div>
</div>
<div class="callout-body-container callout-body">
<p>Smaller local models (2B-4B parameters) work well for simpler tasks like classification, but for accurate structured extraction you’ll generally want 8B+ parameter models. The trade-off is between running costs/speed (local, smaller models) and extraction quality (API or larger models).</p>
</div>
</div>


</section>

</main> <!-- /main -->
<script type="application/vnd.jupyter.widget-state+json">
{"state":{"075d2973f7704bbea4fb038e6bbc38e7":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"17e3be666ff34554be6e0467208b49bb":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"28a32a0d0ce74f03b03c78d258e833c7":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"ProgressStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"296e56f128d74b4a9a3220376ae8e408":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_17e3be666ff34554be6e0467208b49bb","placeholder":"​","style":"IPY_MODEL_9b22327111634402993093389cb6ffd9","tabbable":null,"tooltip":null,"value":" 10/10 [00:18&lt;00:00,  2.15s/it]"}},"7c6477b5384d4da3837c3ae504d452ba":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HBoxModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_88cf551dd9ee4bfca638f5f035536e70","IPY_MODEL_af3005f932a343828a0633f15f02fb38","IPY_MODEL_296e56f128d74b4a9a3220376ae8e408"],"layout":"IPY_MODEL_a9b589f05da7421ea81affbdf6dbe09b","tabbable":null,"tooltip":null}},"81bcff1e98ad4aa580f8f28285245d97":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"88cf551dd9ee4bfca638f5f035536e70":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"HTMLView","description":"","description_allow_html":false,"layout":"IPY_MODEL_9871cddb874d4820803d85d7df53e33c","placeholder":"​","style":"IPY_MODEL_075d2973f7704bbea4fb038e6bbc38e7","tabbable":null,"tooltip":null,"value":"100%"}},"9871cddb874d4820803d85d7df53e33c":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"9b22327111634402993093389cb6ffd9":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"HTMLStyleModel","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"HTMLStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"StyleView","background":null,"description_width":"","font_size":null,"text_color":null}},"a9b589f05da7421ea81affbdf6dbe09b":{"model_module":"@jupyter-widgets/base","model_module_version":"2.0.0","model_name":"LayoutModel","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"2.0.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"2.0.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border_bottom":null,"border_left":null,"border_right":null,"border_top":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"af3005f932a343828a0633f15f02fb38":{"model_module":"@jupyter-widgets/controls","model_module_version":"2.0.0","model_name":"FloatProgressModel","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"2.0.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"2.0.0","_view_name":"ProgressView","bar_style":"success","description":"","description_allow_html":false,"layout":"IPY_MODEL_81bcff1e98ad4aa580f8f28285245d97","max":10,"min":0,"orientation":"horizontal","style":"IPY_MODEL_28a32a0d0ce74f03b03c78d258e833c7","tabbable":null,"tooltip":null,"value":10}}},"version_major":2,"version_minor":0}
</script>
<script id="quarto-html-after-body" type="application/javascript">
  window.document.addEventListener("DOMContentLoaded", function (event) {
    const icon = "";
    const anchorJS = new window.AnchorJS();
    anchorJS.options = {
      placement: 'right',
      icon: icon
    };
    anchorJS.add('.anchored');
    const isCodeAnnotation = (el) => {
      for (const clz of el.classList) {
        if (clz.startsWith('code-annotation-')) {                     
          return true;
        }
      }
      return false;
    }
    const onCopySuccess = function(e) {
      // button target
      const button = e.trigger;
      // don't keep focus
      button.blur();
      // flash "checked"
      button.classList.add('code-copy-button-checked');
      var currentTitle = button.getAttribute("title");
      button.setAttribute("title", "Copied!");
      let tooltip;
      if (window.bootstrap) {
        button.setAttribute("data-bs-toggle", "tooltip");
        button.setAttribute("data-bs-placement", "left");
        button.setAttribute("data-bs-title", "Copied!");
        tooltip = new bootstrap.Tooltip(button, 
          { trigger: "manual", 
            customClass: "code-copy-button-tooltip",
            offset: [0, -8]});
        tooltip.show();    
      }
      setTimeout(function() {
        if (tooltip) {
          tooltip.hide();
          button.removeAttribute("data-bs-title");
          button.removeAttribute("data-bs-toggle");
          button.removeAttribute("data-bs-placement");
        }
        button.setAttribute("title", currentTitle);
        button.classList.remove('code-copy-button-checked');
      }, 1000);
      // clear code selection
      e.clearSelection();
    }
    const getTextToCopy = function(trigger) {
      const outerScaffold = trigger.parentElement.cloneNode(true);
      const codeEl = outerScaffold.querySelector('code');
      for (const childEl of codeEl.children) {
        if (isCodeAnnotation(childEl)) {
          childEl.remove();
        }
      }
      return codeEl.innerText;
    }
    const clipboard = new window.ClipboardJS('.code-copy-button:not([data-in-quarto-modal])', {
      text: getTextToCopy
    });
    clipboard.on('success', onCopySuccess);
    if (window.document.getElementById('quarto-embedded-source-code-modal')) {
      const clipboardModal = new window.ClipboardJS('.code-copy-button[data-in-quarto-modal]', {
        text: getTextToCopy,
        container: window.document.getElementById('quarto-embedded-source-code-modal')
      });
      clipboardModal.on('success', onCopySuccess);
    }
      var localhostRegex = new RegExp(/^(?:http|https):\/\/localhost\:?[0-9]*\//);
      var mailtoRegex = new RegExp(/^mailto:/);
        var filterRegex = new RegExp('/' + window.location.host + '/');
      var isInternal = (href) => {
          return filterRegex.test(href) || localhostRegex.test(href) || mailtoRegex.test(href);
      }
      // Inspect non-navigation links and adorn them if external
     var links = window.document.querySelectorAll('a[href]:not(.nav-link):not(.navbar-brand):not(.toc-action):not(.sidebar-link):not(.sidebar-item-toggle):not(.pagination-link):not(.no-external):not([aria-hidden]):not(.dropdown-item):not(.quarto-navigation-tool):not(.about-link)');
      for (var i=0; i<links.length; i++) {
        const link = links[i];
        if (!isInternal(link.href)) {
          // undo the damage that might have been done by quarto-nav.js in the case of
          // links that we want to consider external
          if (link.dataset.originalHref !== undefined) {
            link.href = link.dataset.originalHref;
          }
        }
      }
    function tippyHover(el, contentFn, onTriggerFn, onUntriggerFn) {
      const config = {
        allowHTML: true,
        maxWidth: 500,
        delay: 100,
        arrow: false,
        appendTo: function(el) {
            return el.parentElement;
        },
        interactive: true,
        interactiveBorder: 10,
        theme: 'quarto',
        placement: 'bottom-start',
      };
      if (contentFn) {
        config.content = contentFn;
      }
      if (onTriggerFn) {
        config.onTrigger = onTriggerFn;
      }
      if (onUntriggerFn) {
        config.onUntrigger = onUntriggerFn;
      }
      window.tippy(el, config); 
    }
    const noterefs = window.document.querySelectorAll('a[role="doc-noteref"]');
    for (var i=0; i<noterefs.length; i++) {
      const ref = noterefs[i];
      tippyHover(ref, function() {
        // use id or data attribute instead here
        let href = ref.getAttribute('data-footnote-href') || ref.getAttribute('href');
        try { href = new URL(href).hash; } catch {}
        const id = href.replace(/^#\/?/, "");
        const note = window.document.getElementById(id);
        if (note) {
          return note.innerHTML;
        } else {
          return "";
        }
      });
    }
    const xrefs = window.document.querySelectorAll('a.quarto-xref');
    const processXRef = (id, note) => {
      // Strip column container classes
      const stripColumnClz = (el) => {
        el.classList.remove("page-full", "page-columns");
        if (el.children) {
          for (const child of el.children) {
            stripColumnClz(child);
          }
        }
      }
      stripColumnClz(note)
      if (id === null || id.startsWith('sec-')) {
        // Special case sections, only their first couple elements
        const container = document.createElement("div");
        if (note.children && note.children.length > 2) {
          container.appendChild(note.children[0].cloneNode(true));
          for (let i = 1; i < note.children.length; i++) {
            const child = note.children[i];
            if (child.tagName === "P" && child.innerText === "") {
              continue;
            } else {
              container.appendChild(child.cloneNode(true));
              break;
            }
          }
          if (window.Quarto?.typesetMath) {
            window.Quarto.typesetMath(container);
          }
          return container.innerHTML
        } else {
          if (window.Quarto?.typesetMath) {
            window.Quarto.typesetMath(note);
          }
          return note.innerHTML;
        }
      } else {
        // Remove any anchor links if they are present
        const anchorLink = note.querySelector('a.anchorjs-link');
        if (anchorLink) {
          anchorLink.remove();
        }
        if (window.Quarto?.typesetMath) {
          window.Quarto.typesetMath(note);
        }
        if (note.classList.contains("callout")) {
          return note.outerHTML;
        } else {
          return note.innerHTML;
        }
      }
    }
    for (var i=0; i<xrefs.length; i++) {
      const xref = xrefs[i];
      tippyHover(xref, undefined, function(instance) {
        instance.disable();
        let url = xref.getAttribute('href');
        let hash = undefined; 
        if (url.startsWith('#')) {
          hash = url;
        } else {
          try { hash = new URL(url).hash; } catch {}
        }
        if (hash) {
          const id = hash.replace(/^#\/?/, "");
          const note = window.document.getElementById(id);
          if (note !== null) {
            try {
              const html = processXRef(id, note.cloneNode(true));
              instance.setContent(html);
            } finally {
              instance.enable();
              instance.show();
            }
          } else {
            // See if we can fetch this
            fetch(url.split('#')[0])
            .then(res => res.text())
            .then(html => {
              const parser = new DOMParser();
              const htmlDoc = parser.parseFromString(html, "text/html");
              const note = htmlDoc.getElementById(id);
              if (note !== null) {
                const html = processXRef(id, note);
                instance.setContent(html);
              } 
            }).finally(() => {
              instance.enable();
              instance.show();
            });
          }
        } else {
          // See if we can fetch a full url (with no hash to target)
          // This is a special case and we should probably do some content thinning / targeting
          fetch(url)
          .then(res => res.text())
          .then(html => {
            const parser = new DOMParser();
            const htmlDoc = parser.parseFromString(html, "text/html");
            const note = htmlDoc.querySelector('main.content');
            if (note !== null) {
              // This should only happen for chapter cross references
              // (since there is no id in the URL)
              // remove the first header
              if (note.children.length > 0 && note.children[0].tagName === "HEADER") {
                note.children[0].remove();
              }
              const html = processXRef(null, note);
              instance.setContent(html);
            } 
          }).finally(() => {
            instance.enable();
            instance.show();
          });
        }
      }, function(instance) {
      });
    }
        let selectedAnnoteEl;
        const selectorForAnnotation = ( cell, annotation) => {
          let cellAttr = 'data-code-cell="' + cell + '"';
          let lineAttr = 'data-code-annotation="' +  annotation + '"';
          const selector = 'span[' + cellAttr + '][' + lineAttr + ']';
          return selector;
        }
        const selectCodeLines = (annoteEl) => {
          const doc = window.document;
          const targetCell = annoteEl.getAttribute("data-target-cell");
          const targetAnnotation = annoteEl.getAttribute("data-target-annotation");
          const annoteSpan = window.document.querySelector(selectorForAnnotation(targetCell, targetAnnotation));
          const lines = annoteSpan.getAttribute("data-code-lines").split(",");
          const lineIds = lines.map((line) => {
            return targetCell + "-" + line;
          })
          let top = null;
          let height = null;
          let parent = null;
          if (lineIds.length > 0) {
              //compute the position of the single el (top and bottom and make a div)
              const el = window.document.getElementById(lineIds[0]);
              top = el.offsetTop;
              height = el.offsetHeight;
              parent = el.parentElement.parentElement;
            if (lineIds.length > 1) {
              const lastEl = window.document.getElementById(lineIds[lineIds.length - 1]);
              const bottom = lastEl.offsetTop + lastEl.offsetHeight;
              height = bottom - top;
            }
            if (top !== null && height !== null && parent !== null) {
              // cook up a div (if necessary) and position it 
              let div = window.document.getElementById("code-annotation-line-highlight");
              if (div === null) {
                div = window.document.createElement("div");
                div.setAttribute("id", "code-annotation-line-highlight");
                div.style.position = 'absolute';
                parent.appendChild(div);
              }
              div.style.top = top - 2 + "px";
              div.style.height = height + 4 + "px";
              div.style.left = 0;
              let gutterDiv = window.document.getElementById("code-annotation-line-highlight-gutter");
              if (gutterDiv === null) {
                gutterDiv = window.document.createElement("div");
                gutterDiv.setAttribute("id", "code-annotation-line-highlight-gutter");
                gutterDiv.style.position = 'absolute';
                const codeCell = window.document.getElementById(targetCell);
                const gutter = codeCell.querySelector('.code-annotation-gutter');
                gutter.appendChild(gutterDiv);
              }
              gutterDiv.style.top = top - 2 + "px";
              gutterDiv.style.height = height + 4 + "px";
            }
            selectedAnnoteEl = annoteEl;
          }
        };
        const unselectCodeLines = () => {
          const elementsIds = ["code-annotation-line-highlight", "code-annotation-line-highlight-gutter"];
          elementsIds.forEach((elId) => {
            const div = window.document.getElementById(elId);
            if (div) {
              div.remove();
            }
          });
          selectedAnnoteEl = undefined;
        };
          // Handle positioning of the toggle
      window.addEventListener(
        "resize",
        throttle(() => {
          elRect = undefined;
          if (selectedAnnoteEl) {
            selectCodeLines(selectedAnnoteEl);
          }
        }, 10)
      );
      function throttle(fn, ms) {
      let throttle = false;
      let timer;
        return (...args) => {
          if(!throttle) { // first call gets through
              fn.apply(this, args);
              throttle = true;
          } else { // all the others get throttled
              if(timer) clearTimeout(timer); // cancel #2
              timer = setTimeout(() => {
                fn.apply(this, args);
                timer = throttle = false;
              }, ms);
          }
        };
      }
        // Attach click handler to the DT
        const annoteDls = window.document.querySelectorAll('dt[data-target-cell]');
        for (const annoteDlNode of annoteDls) {
          annoteDlNode.addEventListener('click', (event) => {
            const clickedEl = event.target;
            if (clickedEl !== selectedAnnoteEl) {
              unselectCodeLines();
              const activeEl = window.document.querySelector('dt[data-target-cell].code-annotation-active');
              if (activeEl) {
                activeEl.classList.remove('code-annotation-active');
              }
              selectCodeLines(clickedEl);
              clickedEl.classList.add('code-annotation-active');
            } else {
              // Unselect the line
              unselectCodeLines();
              clickedEl.classList.remove('code-annotation-active');
            }
          });
        }
    const findCites = (el) => {
      const parentEl = el.parentElement;
      if (parentEl) {
        const cites = parentEl.dataset.cites;
        if (cites) {
          return {
            el,
            cites: cites.split(' ')
          };
        } else {
          return findCites(el.parentElement)
        }
      } else {
        return undefined;
      }
    };
    var bibliorefs = window.document.querySelectorAll('a[role="doc-biblioref"]');
    for (var i=0; i<bibliorefs.length; i++) {
      const ref = bibliorefs[i];
      const citeInfo = findCites(ref);
      if (citeInfo) {
        tippyHover(citeInfo.el, function() {
          var popup = window.document.createElement('div');
          citeInfo.cites.forEach(function(cite) {
            var citeDiv = window.document.createElement('div');
            citeDiv.classList.add('hanging-indent');
            citeDiv.classList.add('csl-entry');
            var biblioDiv = window.document.getElementById('ref-' + cite);
            if (biblioDiv) {
              citeDiv.innerHTML = biblioDiv.innerHTML;
            }
            popup.appendChild(citeDiv);
          });
          return popup.innerHTML;
        });
      }
    }
  });
  </script>
<nav class="page-navigation">
  <div class="nav-page nav-page-previous">
      <a href="../../patterns/structured-generation/intro.html" class="pagination-link" aria-label="Structured Document Processing">
        <i class="bi bi-arrow-left-short"></i> <span class="nav-page-text"><span class="chapter-number">2</span>&nbsp; <span class="chapter-title">Structured Document Processing</span></span>
      </a>          
  </div>
  <div class="nav-page nav-page-next">
  </div>
</nav>
</div> <!-- /content -->




</body></html>