@geotechcli/core 0.4.26 → 0.4.28

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (57) hide show
  1. package/dist/index.d.ts +2 -0
  2. package/dist/index.d.ts.map +1 -1
  3. package/dist/index.js +3 -0
  4. package/dist/index.js.map +1 -1
  5. package/dist/ingest/job-store.d.ts.map +1 -1
  6. package/dist/ingest/job-store.js +15 -5
  7. package/dist/ingest/job-store.js.map +1 -1
  8. package/dist/ingest/job-worker.d.ts +2 -1
  9. package/dist/ingest/job-worker.d.ts.map +1 -1
  10. package/dist/ingest/job-worker.js +186 -32
  11. package/dist/ingest/job-worker.js.map +1 -1
  12. package/dist/meta/metadata.json +1 -1
  13. package/dist/tabular/csv.d.ts +17 -0
  14. package/dist/tabular/csv.d.ts.map +1 -0
  15. package/dist/tabular/csv.js +123 -0
  16. package/dist/tabular/csv.js.map +1 -0
  17. package/dist/tabular/index.d.ts +4 -0
  18. package/dist/tabular/index.d.ts.map +1 -0
  19. package/dist/tabular/index.js +4 -0
  20. package/dist/tabular/index.js.map +1 -0
  21. package/dist/tabular/schema-inference.d.ts +45 -0
  22. package/dist/tabular/schema-inference.d.ts.map +1 -0
  23. package/dist/tabular/schema-inference.js +272 -0
  24. package/dist/tabular/schema-inference.js.map +1 -0
  25. package/dist/tabular/xlsx.d.ts +15 -0
  26. package/dist/tabular/xlsx.d.ts.map +1 -0
  27. package/dist/tabular/xlsx.js +190 -0
  28. package/dist/tabular/xlsx.js.map +1 -0
  29. package/dist/vision/geotech-document.d.ts +4 -0
  30. package/dist/vision/geotech-document.d.ts.map +1 -1
  31. package/dist/vision/geotech-document.js +38 -11
  32. package/dist/vision/geotech-document.js.map +1 -1
  33. package/dist/vision/ocr.d.ts +1 -0
  34. package/dist/vision/ocr.d.ts.map +1 -1
  35. package/dist/vision/ocr.js +5 -1
  36. package/dist/vision/ocr.js.map +1 -1
  37. package/dist/workspace/classifier.d.ts +7 -0
  38. package/dist/workspace/classifier.d.ts.map +1 -0
  39. package/dist/workspace/classifier.js +186 -0
  40. package/dist/workspace/classifier.js.map +1 -0
  41. package/dist/workspace/dossier.d.ts +3 -0
  42. package/dist/workspace/dossier.d.ts.map +1 -0
  43. package/dist/workspace/dossier.js +266 -0
  44. package/dist/workspace/dossier.js.map +1 -0
  45. package/dist/workspace/index.d.ts +5 -0
  46. package/dist/workspace/index.d.ts.map +1 -0
  47. package/dist/workspace/index.js +5 -0
  48. package/dist/workspace/index.js.map +1 -0
  49. package/dist/workspace/manifest.d.ts +58 -0
  50. package/dist/workspace/manifest.d.ts.map +1 -0
  51. package/dist/workspace/manifest.js +7 -0
  52. package/dist/workspace/manifest.js.map +1 -0
  53. package/dist/workspace/scanner.d.ts +3 -0
  54. package/dist/workspace/scanner.d.ts.map +1 -0
  55. package/dist/workspace/scanner.js +245 -0
  56. package/dist/workspace/scanner.js.map +1 -0
  57. package/package.json +1 -1
@@ -0,0 +1,272 @@
1
+ function normalizeHeader(value) {
2
+ return value.toLowerCase().replace(/[^a-z0-9]+/g, '');
3
+ }
4
+ function humanCell(value) {
5
+ if (value == null)
6
+ return '';
7
+ if (typeof value === 'number')
8
+ return String(value);
9
+ if (typeof value === 'boolean')
10
+ return value ? 'true' : 'false';
11
+ return value;
12
+ }
13
+ function isFiniteNumber(value) {
14
+ return typeof value === 'number' && Number.isFinite(value);
15
+ }
16
+ function inferColumnType(values) {
17
+ const present = values.filter((value) => value !== null && String(value).trim().length > 0);
18
+ if (present.length === 0)
19
+ return 'empty';
20
+ const numeric = present.filter(isFiniteNumber).length;
21
+ const boolean = present.filter((value) => typeof value === 'boolean').length;
22
+ const dates = present.filter((value) => typeof value === 'string' && !Number.isNaN(Date.parse(value))).length;
23
+ if (numeric === present.length)
24
+ return 'number';
25
+ if (boolean === present.length)
26
+ return 'boolean';
27
+ if (dates >= Math.max(2, Math.ceil(present.length * 0.8)))
28
+ return 'date';
29
+ if (numeric > 0 || boolean > 0)
30
+ return 'mixed';
31
+ return 'text';
32
+ }
33
+ function inferUnit(header) {
34
+ const explicit = header.match(/\(([^)]+)\)|\[([^\]]+)\]/);
35
+ if (explicit) {
36
+ return (explicit[1] ?? explicit[2]).trim();
37
+ }
38
+ const normalized = header
39
+ .toLowerCase()
40
+ .replace(/([a-z])([A-Z])/g, '$1_$2')
41
+ .replace(/[^a-z0-9]+/g, '_');
42
+ const units = [
43
+ 'kn_m3',
44
+ 'g_cm3',
45
+ 'kpa',
46
+ 'mpa',
47
+ 'gpa',
48
+ 'kn',
49
+ 'mn',
50
+ 'mm',
51
+ 'cm',
52
+ 'm',
53
+ 'deg',
54
+ 'degree',
55
+ 'percent',
56
+ 'pct',
57
+ 'hz',
58
+ 's',
59
+ 'sec',
60
+ 'min',
61
+ 'hr',
62
+ 'day',
63
+ 'year',
64
+ ];
65
+ return units.find((unit) => new RegExp(`(?:^|_)${unit}(?:_|$)`).test(normalized));
66
+ }
67
+ function normalizeAlias(value) {
68
+ return value.toLowerCase().replace(/[^a-z0-9]+/g, '');
69
+ }
70
+ function hasAny(normalized, aliases) {
71
+ return aliases.some((rawAlias) => {
72
+ const alias = normalizeAlias(rawAlias);
73
+ if (alias.length <= 3) {
74
+ return normalized === alias || normalized.startsWith(alias);
75
+ }
76
+ return normalized.includes(alias);
77
+ });
78
+ }
79
+ function inferRoles(header, type) {
80
+ const normalized = normalizeHeader(header);
81
+ const roles = new Set();
82
+ if (hasAny(normalized, ['depth', 'belowground', 'bgl', 'elevationdepth', 'z']))
83
+ roles.add('depth');
84
+ if (hasAny(normalized, ['time', 'elapsed', 'duration']))
85
+ roles.add('time');
86
+ if (hasAny(normalized, ['date', 'timestamp', 'datetime']))
87
+ roles.add('date');
88
+ if (hasAny(normalized, ['easting', 'eastcoord', 'xcoord', 'gridx']))
89
+ roles.add('easting');
90
+ if (hasAny(normalized, ['northing', 'northcoord', 'ycoord', 'gridy']))
91
+ roles.add('northing');
92
+ if (hasAny(normalized, ['latitude', 'lat']))
93
+ roles.add('latitude');
94
+ if (hasAny(normalized, ['longitude', 'lon', 'lng']))
95
+ roles.add('longitude');
96
+ if (hasAny(normalized, ['borehole', 'holeid', 'bhid', 'bhno', 'locationid']))
97
+ roles.add('borehole_id');
98
+ if (hasAny(normalized, ['sampleid', 'sampleno', 'sample', 'specimen']))
99
+ roles.add('sample_id');
100
+ if (hasAny(normalized, ['testid', 'testno']))
101
+ roles.add('test_id');
102
+ if (hasAny(normalized, ['sptn', 'nvalue', 'blowcount', 'blows', 'n60']))
103
+ roles.add('spt_n');
104
+ if (hasAny(normalized, ['qc', 'coneresistance', 'tipresistance']))
105
+ roles.add('cpt_qc');
106
+ if (hasAny(normalized, ['fs', 'sleevefriction', 'frictionsleeve']))
107
+ roles.add('cpt_fs');
108
+ if (hasAny(normalized, ['watercontent', 'moisturecontent', 'naturalmoisture', 'wpercent']))
109
+ roles.add('water_content');
110
+ if (hasAny(normalized, ['liquidlimit', 'll']))
111
+ roles.add('liquid_limit');
112
+ if (hasAny(normalized, ['plasticlimit', 'pl']))
113
+ roles.add('plastic_limit');
114
+ if (hasAny(normalized, ['plasticityindex', 'pi']))
115
+ roles.add('plasticity_index');
116
+ if (hasAny(normalized, ['particlesize', 'grainsize', 'sievemm', 'openingmm']))
117
+ roles.add('gradation_size');
118
+ if (hasAny(normalized, ['percentpassing', 'passingpercent', 'percentfiner', 'finerpercent']))
119
+ roles.add('percent_passing');
120
+ if (hasAny(normalized, ['unitweight', 'bulkunitweight', 'dryunitweight', 'drydensity', 'gammad']))
121
+ roles.add('unit_weight');
122
+ if (hasAny(normalized, ['settlement', 'heave', 'displacement']))
123
+ roles.add('settlement');
124
+ if (hasAny(normalized, ['porepressure', 'piezometer', 'waterpressure']))
125
+ roles.add('pore_pressure');
126
+ if (hasAny(normalized, ['inclinometer', 'inclination', 'tilt']))
127
+ roles.add('inclination');
128
+ if (hasAny(normalized, ['vibration', 'acceleration', 'velocity', 'frequency', 'fft', 'psd']))
129
+ roles.add('vibration');
130
+ if (roles.size === 0 && type === 'date')
131
+ roles.add('date');
132
+ return roles.size > 0 ? [...roles] : ['unknown'];
133
+ }
134
+ function namesByRole(columns, role) {
135
+ return columns.filter((column) => column.roles.includes(role)).map((column) => column.name);
136
+ }
137
+ function inferDatasetType(columns) {
138
+ const has = (role) => columns.some((column) => column.roles.includes(role));
139
+ const hasAnyRole = (roles) => roles.some(has);
140
+ if (has('cpt_qc') && has('depth'))
141
+ return 'cpt-profile';
142
+ if (has('spt_n') && has('depth'))
143
+ return 'spt-profile';
144
+ if (hasAnyRole(['liquid_limit', 'plastic_limit', 'plasticity_index', 'gradation_size', 'percent_passing', 'water_content', 'unit_weight']))
145
+ return 'lab-test-summary';
146
+ if ((has('easting') && has('northing')) || (has('latitude') && has('longitude')))
147
+ return 'coordinate-table';
148
+ if (hasAnyRole(['settlement', 'pore_pressure', 'inclination']) && hasAnyRole(['time', 'date']))
149
+ return 'monitoring-time-series';
150
+ if (has('vibration'))
151
+ return 'signal-record';
152
+ if (has('borehole_id') && has('depth'))
153
+ return 'borehole-table';
154
+ return 'generic-table';
155
+ }
156
+ function inferBranches(datasetType, columns) {
157
+ const branches = new Set();
158
+ const has = (role) => columns.some((column) => column.roles.includes(role));
159
+ switch (datasetType) {
160
+ case 'cpt-profile':
161
+ case 'spt-profile':
162
+ case 'borehole-table':
163
+ case 'lab-test-summary':
164
+ branches.add('site-investigation');
165
+ branches.add('foundation');
166
+ break;
167
+ case 'coordinate-table':
168
+ branches.add('site-investigation');
169
+ branches.add('mapping');
170
+ break;
171
+ case 'monitoring-time-series':
172
+ branches.add('monitoring');
173
+ branches.add('settlement');
174
+ break;
175
+ case 'signal-record':
176
+ branches.add('monitoring');
177
+ branches.add('signal-processing');
178
+ break;
179
+ default:
180
+ break;
181
+ }
182
+ if (has('spt_n'))
183
+ branches.add('liquefaction');
184
+ if (has('settlement'))
185
+ branches.add('settlement');
186
+ if (has('pore_pressure'))
187
+ branches.add('seepage');
188
+ return [...branches];
189
+ }
190
+ function inferConfidence(datasetType, columns, sampledRowCount) {
191
+ if (sampledRowCount === 0 || columns.length === 0)
192
+ return 0.2;
193
+ let score = datasetType === 'generic-table' ? 0.45 : 0.68;
194
+ const recognizedColumns = columns.filter((column) => !column.roles.includes('unknown')).length;
195
+ score += Math.min(0.22, recognizedColumns / Math.max(1, columns.length) * 0.22);
196
+ if (sampledRowCount >= 5)
197
+ score += 0.08;
198
+ if (sampledRowCount >= 20)
199
+ score += 0.04;
200
+ return Math.round(Math.min(score, 0.96) * 100) / 100;
201
+ }
202
+ export function inferTabularSchema(options) {
203
+ const headers = options.rows.length > 0 ? Object.keys(options.rows[0]) : [];
204
+ const columns = headers.map((header) => {
205
+ const values = options.rows.map((row) => row[header]);
206
+ const type = inferColumnType(values);
207
+ const examples = values
208
+ .map(humanCell)
209
+ .filter((value) => value.length > 0)
210
+ .slice(0, 3);
211
+ const nonEmptyCount = values.filter((value) => value !== null && String(value).trim().length > 0).length;
212
+ const numericCount = values.filter(isFiniteNumber).length;
213
+ return {
214
+ name: header,
215
+ normalizedName: normalizeHeader(header),
216
+ type,
217
+ unit: inferUnit(header),
218
+ roles: inferRoles(header, type),
219
+ nonEmptyCount,
220
+ numericCount,
221
+ examples,
222
+ };
223
+ });
224
+ const datasetType = inferDatasetType(columns);
225
+ const branches = inferBranches(datasetType, columns);
226
+ const warnings = [...(options.warnings ?? [])];
227
+ if (datasetType === 'generic-table') {
228
+ warnings.push('No geotechnical schema pattern was detected from sampled headers.');
229
+ }
230
+ return {
231
+ sourceName: options.sourceName,
232
+ sheetName: options.sheetName,
233
+ rowCount: options.rowCount ?? options.rows.length,
234
+ sampledRowCount: options.rows.length,
235
+ columnCount: columns.length,
236
+ columns,
237
+ datasetType,
238
+ branches,
239
+ detected: {
240
+ depthColumns: namesByRole(columns, 'depth'),
241
+ timeColumns: [...new Set([...namesByRole(columns, 'time'), ...namesByRole(columns, 'date')])],
242
+ coordinateColumns: [...new Set([
243
+ ...namesByRole(columns, 'easting'),
244
+ ...namesByRole(columns, 'northing'),
245
+ ...namesByRole(columns, 'latitude'),
246
+ ...namesByRole(columns, 'longitude'),
247
+ ])],
248
+ boreholeIdColumns: namesByRole(columns, 'borehole_id'),
249
+ sampleIdColumns: namesByRole(columns, 'sample_id'),
250
+ sptColumns: namesByRole(columns, 'spt_n'),
251
+ cptColumns: [...new Set([...namesByRole(columns, 'cpt_qc'), ...namesByRole(columns, 'cpt_fs')])],
252
+ labColumns: [...new Set([
253
+ ...namesByRole(columns, 'water_content'),
254
+ ...namesByRole(columns, 'liquid_limit'),
255
+ ...namesByRole(columns, 'plastic_limit'),
256
+ ...namesByRole(columns, 'plasticity_index'),
257
+ ...namesByRole(columns, 'gradation_size'),
258
+ ...namesByRole(columns, 'percent_passing'),
259
+ ...namesByRole(columns, 'unit_weight'),
260
+ ])],
261
+ monitoringColumns: [...new Set([
262
+ ...namesByRole(columns, 'settlement'),
263
+ ...namesByRole(columns, 'pore_pressure'),
264
+ ...namesByRole(columns, 'inclination'),
265
+ ...namesByRole(columns, 'vibration'),
266
+ ])],
267
+ },
268
+ confidence: inferConfidence(datasetType, columns, options.rows.length),
269
+ warnings,
270
+ };
271
+ }
272
+ //# sourceMappingURL=schema-inference.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"schema-inference.js","sourceRoot":"","sources":["../../src/tabular/schema-inference.ts"],"names":[],"mappings":"AA6EA,SAAS,eAAe,CAAC,KAAa;IACpC,OAAO,KAAK,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC;AACxD,CAAC;AAED,SAAS,SAAS,CAAC,KAAkB;IACnC,IAAI,KAAK,IAAI,IAAI;QAAE,OAAO,EAAE,CAAC;IAC7B,IAAI,OAAO,KAAK,KAAK,QAAQ;QAAE,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC;IACpD,IAAI,OAAO,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC;IAChE,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,cAAc,CAAC,KAAkB;IACxC,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;AAC7D,CAAC;AAED,SAAS,eAAe,CAAC,MAAqB;IAC5C,MAAM,OAAO,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,IAAI,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5F,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,OAAO,CAAC;IAEzC,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC,MAAM,CAAC;IACtD,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,MAAM,CAAC;IAC7E,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAE9G,IAAI,OAAO,KAAK,OAAO,CAAC,MAAM;QAAE,OAAO,QAAQ,CAAC;IAChD,IAAI,OAAO,KAAK,OAAO,CAAC,MAAM;QAAE,OAAO,SAAS,CAAC;IACjD,IAAI,KAAK,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;QAAE,OAAO,MAAM,CAAC;IACzE,IAAI,OAAO,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC;IAC/C,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,SAAS,CAAC,MAAc;IAC/B,MAAM,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,0BAA0B,CAAC,CAAC;IAC1D,IAAI,QAAQ,EAAE,CAAC;QACb,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAC7C,CAAC;IAED,MAAM,UAAU,GAAG,MAAM;SACtB,WAAW,EAAE;SACb,OAAO,CAAC,iBAAiB,EAAE,OAAO,CAAC;SACnC,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC,CAAC;IAC/B,MAAM,KAAK,GAAG;QACZ,OAAO;QACP,OAAO;QACP,KAAK;QACL,KAAK;QACL,KAAK;QACL,IAAI;QACJ,IAAI;QACJ,IAAI;QACJ,IAAI;QACJ,GAAG;QACH,KAAK;QACL,QAAQ;QACR,SAAS;QACT,KAAK;QACL,IAAI;QACJ,GAAG;QACH,KAAK;QACL,KAAK;QACL,IAAI;QACJ,KAAK;QACL,MAAM;KACP,CAAC;IAEF,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,UAAU,IAAI,SAAS,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC;AACpF,CAAC;AAED,SAAS,cAAc,CAAC,KAAa;IACnC,OAAO,KAAK,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC;AACxD,CAAC;AAED,SAAS,MAAM,CAAC,UAAkB,EAAE,OAAiB;IACnD,OAAO,OAAO,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE;QAC/B,MAAM,KAAK,GAAG,cAAc,CAAC,QAAQ,CAAC,CAAC;QACvC,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;YACtB,OAAO,UAAU,KAAK,KAAK,IAAI,UAAU,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;QAC9D,CAAC;QACD,OAAO,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IACpC,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,UAAU,CAAC,MAAc,EAAE,IAAwB;IAC1D,MAAM,UAAU,GAAG,eAAe,CAAC,MAAM,CAAC,CAAC;IAC3C,MAAM,KAAK,GAAG,IAAI,GAAG,EAAsB,CAAC;IAE5C,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,OAAO,EAAE,aAAa,EAAE,KAAK,EAAE,gBAAgB,EAAE,GAAG,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IACnG,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,MAAM,EAAE,SAAS,EAAE,UAAU,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;IAC3E,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,MAAM,EAAE,WAAW,EAAE,UAAU,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;IAC7E,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,SAAS,EAAE,WAAW,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;IAC1F,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,UAAU,EAAE,YAAY,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;IAC7F,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,UAAU,EAAE,KAAK,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;IACnE,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,WAAW,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;IAC5E,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,UAAU,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,EAAE,YAAY,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC;IACvG,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,UAAU,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;IAC/F,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;IACnE,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,MAAM,EAAE,QAAQ,EAAE,WAAW,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;IAC5F,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,gBAAgB,EAAE,eAAe,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IACvF,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,IAAI,EAAE,gBAAgB,EAAE,gBAAgB,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IACxF,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,cAAc,EAAE,iBAAiB,EAAE,iBAAiB,EAAE,UAAU,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,eAAe,CAAC,CAAC;IACvH,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,aAAa,EAAE,IAAI,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,cAAc,CAAC,CAAC;IACzE,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,cAAc,EAAE,IAAI,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,eAAe,CAAC,CAAC;IAC3E,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,iBAAiB,EAAE,IAAI,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,kBAAkB,CAAC,CAAC;IACjF,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,cAAc,EAAE,WAAW,EAAE,SAAS,EAAE,WAAW,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC;IAC3G,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,gBAAgB,EAAE,gBAAgB,EAAE,cAAc,EAAE,cAAc,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,iBAAiB,CAAC,CAAC;IAC3H,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,YAAY,EAAE,gBAAgB,EAAE,eAAe,EAAE,YAAY,EAAE,QAAQ,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC;IAC5H,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,YAAY,EAAE,OAAO,EAAE,cAAc,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;IACzF,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,cAAc,EAAE,YAAY,EAAE,eAAe,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,eAAe,CAAC,CAAC;IACpG,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,cAAc,EAAE,aAAa,EAAE,MAAM,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC;IAC1F,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC,WAAW,EAAE,cAAc,EAAE,UAAU,EAAE,WAAW,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;QAAE,KAAK,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;IAErH,IAAI,KAAK,CAAC,IAAI,KAAK,CAAC,IAAI,IAAI,KAAK,MAAM;QAAE,KAAK,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;IAC3D,OAAO,KAAK,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;AACnD,CAAC;AAED,SAAS,WAAW,CAAC,OAA0B,EAAE,IAAwB;IACvE,OAAO,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;AAC9F,CAAC;AAED,SAAS,gBAAgB,CAAC,OAA0B;IAClD,MAAM,GAAG,GAAG,CAAC,IAAwB,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;IAChG,MAAM,UAAU,GAAG,CAAC,KAA2B,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAEpE,IAAI,GAAG,CAAC,QAAQ,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC;QAAE,OAAO,aAAa,CAAC;IACxD,IAAI,GAAG,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC;QAAE,OAAO,aAAa,CAAC;IACvD,IAAI,UAAU,CAAC,CAAC,cAAc,EAAE,eAAe,EAAE,kBAAkB,EAAE,gBAAgB,EAAE,iBAAiB,EAAE,eAAe,EAAE,aAAa,CAAC,CAAC;QAAE,OAAO,kBAAkB,CAAC;IACtK,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,GAAG,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,GAAG,CAAC,WAAW,CAAC,CAAC;QAAE,OAAO,kBAAkB,CAAC;IAC5G,IAAI,UAAU,CAAC,CAAC,YAAY,EAAE,eAAe,EAAE,aAAa,CAAC,CAAC,IAAI,UAAU,CAAC,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAAE,OAAO,wBAAwB,CAAC;IAChI,IAAI,GAAG,CAAC,WAAW,CAAC;QAAE,OAAO,eAAe,CAAC;IAC7C,IAAI,GAAG,CAAC,aAAa,CAAC,IAAI,GAAG,CAAC,OAAO,CAAC;QAAE,OAAO,gBAAgB,CAAC;IAChE,OAAO,eAAe,CAAC;AACzB,CAAC;AAED,SAAS,aAAa,CAAC,WAAgC,EAAE,OAA0B;IACjF,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;IACnC,MAAM,GAAG,GAAG,CAAC,IAAwB,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;IAEhG,QAAQ,WAAW,EAAE,CAAC;QACpB,KAAK,aAAa,CAAC;QACnB,KAAK,aAAa,CAAC;QACnB,KAAK,gBAAgB,CAAC;QACtB,KAAK,kBAAkB;YACrB,QAAQ,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAC;YACnC,QAAQ,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;YAC3B,MAAM;QACR,KAAK,kBAAkB;YACrB,QAAQ,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAC;YACnC,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;YACxB,MAAM;QACR,KAAK,wBAAwB;YAC3B,QAAQ,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;YAC3B,QAAQ,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;YAC3B,MAAM;QACR,KAAK,eAAe;YAClB,QAAQ,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;YAC3B,QAAQ,CAAC,GAAG,CAAC,mBAAmB,CAAC,CAAC;YAClC,MAAM;QACR;YACE,MAAM;IACV,CAAC;IAED,IAAI,GAAG,CAAC,OAAO,CAAC;QAAE,QAAQ,CAAC,GAAG,CAAC,cAAc,CAAC,CAAC;IAC/C,IAAI,GAAG,CAAC,YAAY,CAAC;QAAE,QAAQ,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;IAClD,IAAI,GAAG,CAAC,eAAe,CAAC;QAAE,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;IAClD,OAAO,CAAC,GAAG,QAAQ,CAAC,CAAC;AACvB,CAAC;AAED,SAAS,eAAe,CAAC,WAAgC,EAAE,OAA0B,EAAE,eAAuB;IAC5G,IAAI,eAAe,KAAK,CAAC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,GAAG,CAAC;IAC9D,IAAI,KAAK,GAAG,WAAW,KAAK,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;IAC1D,MAAM,iBAAiB,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC;IAC/F,KAAK,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,iBAAiB,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC;IAChF,IAAI,eAAe,IAAI,CAAC;QAAE,KAAK,IAAI,IAAI,CAAC;IACxC,IAAI,eAAe,IAAI,EAAE;QAAE,KAAK,IAAI,IAAI,CAAC;IACzC,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,GAAG,CAAC,GAAG,GAAG,CAAC;AACvD,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,OAMlC;IACC,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAC5E,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE;QACrC,MAAM,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC;QACtD,MAAM,IAAI,GAAG,eAAe,CAAC,MAAM,CAAC,CAAC;QACrC,MAAM,QAAQ,GAAG,MAAM;aACpB,GAAG,CAAC,SAAS,CAAC;aACd,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC;aACnC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;QACf,MAAM,aAAa,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,IAAI,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC;QACzG,MAAM,YAAY,GAAG,MAAM,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC,MAAM,CAAC;QAE1D,OAAO;YACL,IAAI,EAAE,MAAM;YACZ,cAAc,EAAE,eAAe,CAAC,MAAM,CAAC;YACvC,IAAI;YACJ,IAAI,EAAE,SAAS,CAAC,MAAM,CAAC;YACvB,KAAK,EAAE,UAAU,CAAC,MAAM,EAAE,IAAI,CAAC;YAC/B,aAAa;YACb,YAAY;YACZ,QAAQ;SACT,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,MAAM,WAAW,GAAG,gBAAgB,CAAC,OAAO,CAAC,CAAC;IAC9C,MAAM,QAAQ,GAAG,aAAa,CAAC,WAAW,EAAE,OAAO,CAAC,CAAC;IACrD,MAAM,QAAQ,GAAG,CAAC,GAAG,CAAC,OAAO,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,CAAC;IAE/C,IAAI,WAAW,KAAK,eAAe,EAAE,CAAC;QACpC,QAAQ,CAAC,IAAI,CAAC,mEAAmE,CAAC,CAAC;IACrF,CAAC;IAED,OAAO;QACL,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,SAAS,EAAE,OAAO,CAAC,SAAS;QAC5B,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,OAAO,CAAC,IAAI,CAAC,MAAM;QACjD,eAAe,EAAE,OAAO,CAAC,IAAI,CAAC,MAAM;QACpC,WAAW,EAAE,OAAO,CAAC,MAAM;QAC3B,OAAO;QACP,WAAW;QACX,QAAQ;QACR,QAAQ,EAAE;YACR,YAAY,EAAE,WAAW,CAAC,OAAO,EAAE,OAAO,CAAC;YAC3C,WAAW,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,WAAW,CAAC,OAAO,EAAE,MAAM,CAAC,EAAE,GAAG,WAAW,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC,CAAC;YAC7F,iBAAiB,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC;oBAC7B,GAAG,WAAW,CAAC,OAAO,EAAE,SAAS,CAAC;oBAClC,GAAG,WAAW,CAAC,OAAO,EAAE,UAAU,CAAC;oBACnC,GAAG,WAAW,CAAC,OAAO,EAAE,UAAU,CAAC;oBACnC,GAAG,WAAW,CAAC,OAAO,EAAE,WAAW,CAAC;iBACrC,CAAC,CAAC;YACH,iBAAiB,EAAE,WAAW,CAAC,OAAO,EAAE,aAAa,CAAC;YACtD,eAAe,EAAE,WAAW,CAAC,OAAO,EAAE,WAAW,CAAC;YAClD,UAAU,EAAE,WAAW,CAAC,OAAO,EAAE,OAAO,CAAC;YACzC,UAAU,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,WAAW,CAAC,OAAO,EAAE,QAAQ,CAAC,EAAE,GAAG,WAAW,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC,CAAC;YAChG,UAAU,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC;oBACtB,GAAG,WAAW,CAAC,OAAO,EAAE,eAAe,CAAC;oBACxC,GAAG,WAAW,CAAC,OAAO,EAAE,cAAc,CAAC;oBACvC,GAAG,WAAW,CAAC,OAAO,EAAE,eAAe,CAAC;oBACxC,GAAG,WAAW,CAAC,OAAO,EAAE,kBAAkB,CAAC;oBAC3C,GAAG,WAAW,CAAC,OAAO,EAAE,gBAAgB,CAAC;oBACzC,GAAG,WAAW,CAAC,OAAO,EAAE,iBAAiB,CAAC;oBAC1C,GAAG,WAAW,CAAC,OAAO,EAAE,aAAa,CAAC;iBACvC,CAAC,CAAC;YACH,iBAAiB,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC;oBAC7B,GAAG,WAAW,CAAC,OAAO,EAAE,YAAY,CAAC;oBACrC,GAAG,WAAW,CAAC,OAAO,EAAE,eAAe,CAAC;oBACxC,GAAG,WAAW,CAAC,OAAO,EAAE,aAAa,CAAC;oBACtC,GAAG,WAAW,CAAC,OAAO,EAAE,WAAW,CAAC;iBACrC,CAAC,CAAC;SACJ;QACD,UAAU,EAAE,eAAe,CAAC,WAAW,EAAE,OAAO,EAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC;QACtE,QAAQ;KACT,CAAC;AACJ,CAAC"}
@@ -0,0 +1,15 @@
1
+ import type { TabularRow } from './csv.js';
2
+ export interface ParsedWorkbookSheet {
3
+ name: string;
4
+ rows: TabularRow[];
5
+ totalRowsSampled: number;
6
+ warnings: string[];
7
+ }
8
+ export interface ParsedWorkbook {
9
+ sheets: ParsedWorkbookSheet[];
10
+ warnings: string[];
11
+ }
12
+ export declare function parseXlsxFile(filePath: string, options?: {
13
+ maxRows?: number;
14
+ }): Promise<ParsedWorkbook>;
15
+ //# sourceMappingURL=xlsx.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"xlsx.d.ts","sourceRoot":"","sources":["../../src/tabular/xlsx.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAe,UAAU,EAAE,MAAM,UAAU,CAAC;AAExD,MAAM,WAAW,mBAAmB;IAClC,IAAI,EAAE,MAAM,CAAC;IACb,IAAI,EAAE,UAAU,EAAE,CAAC;IACnB,gBAAgB,EAAE,MAAM,CAAC;IACzB,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AAED,MAAM,WAAW,cAAc;IAC7B,MAAM,EAAE,mBAAmB,EAAE,CAAC;IAC9B,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AA0LD,wBAAsB,aAAa,CAAC,QAAQ,EAAE,MAAM,EAAE,OAAO,GAAE;IAAE,OAAO,CAAC,EAAE,MAAM,CAAA;CAAO,GAAG,OAAO,CAAC,cAAc,CAAC,CAoCjH"}
@@ -0,0 +1,190 @@
1
+ import { readFile } from 'node:fs/promises';
2
+ import { strFromU8, unzipSync } from 'fflate';
3
+ const DEFAULT_MAX_ROWS = 200;
4
+ function decodeXml(value) {
5
+ return value
6
+ .replace(/&lt;/g, '<')
7
+ .replace(/&gt;/g, '>')
8
+ .replace(/&quot;/g, '"')
9
+ .replace(/&apos;/g, "'")
10
+ .replace(/&amp;/g, '&');
11
+ }
12
+ function stripTags(value) {
13
+ return decodeXml(value.replace(/<[^>]+>/g, ''));
14
+ }
15
+ function getZipText(entries, name) {
16
+ const normalized = name.replace(/^\/+/, '');
17
+ const entry = entries[normalized];
18
+ return entry ? strFromU8(entry) : undefined;
19
+ }
20
+ function parseAttributes(tag) {
21
+ const attrs = {};
22
+ const pattern = /([A-Za-z_:][A-Za-z0-9_:.-]*)="([^"]*)"/g;
23
+ let match;
24
+ while ((match = pattern.exec(tag)) !== null) {
25
+ attrs[match[1]] = decodeXml(match[2]);
26
+ }
27
+ return attrs;
28
+ }
29
+ function parseSharedStrings(xml) {
30
+ if (!xml)
31
+ return [];
32
+ const strings = [];
33
+ const siPattern = /<si\b[^>]*>([\s\S]*?)<\/si>/g;
34
+ let match;
35
+ while ((match = siPattern.exec(xml)) !== null) {
36
+ const segment = match[1];
37
+ const texts = [...segment.matchAll(/<t\b[^>]*>([\s\S]*?)<\/t>/g)].map((textMatch) => decodeXml(textMatch[1]));
38
+ strings.push(texts.length > 0 ? texts.join('') : stripTags(segment));
39
+ }
40
+ return strings;
41
+ }
42
+ function parseWorkbookSheets(workbookXml) {
43
+ if (!workbookXml)
44
+ return [];
45
+ const sheets = [];
46
+ const sheetPattern = /<sheet\b([^>]+?)\/?>/g;
47
+ let match;
48
+ while ((match = sheetPattern.exec(workbookXml)) !== null) {
49
+ const attrs = parseAttributes(match[1]);
50
+ const relId = attrs['r:id'];
51
+ if (attrs.name && relId) {
52
+ sheets.push({ name: attrs.name, relId });
53
+ }
54
+ }
55
+ return sheets;
56
+ }
57
+ function parseWorkbookRelationships(relsXml) {
58
+ const rels = new Map();
59
+ if (!relsXml)
60
+ return rels;
61
+ const relPattern = /<Relationship\b([^>]+?)\/?>/g;
62
+ let match;
63
+ while ((match = relPattern.exec(relsXml)) !== null) {
64
+ const attrs = parseAttributes(match[1]);
65
+ if (attrs.Id && attrs.Target) {
66
+ const target = attrs.Target.startsWith('/xl/') ? attrs.Target.slice(1) : `xl/${attrs.Target.replace(/^\/+/, '')}`;
67
+ rels.set(attrs.Id, target.replace(/\\/g, '/'));
68
+ }
69
+ }
70
+ return rels;
71
+ }
72
+ function columnIndexFromRef(cellRef) {
73
+ const letters = (cellRef.match(/^[A-Z]+/i)?.[0] ?? '').toUpperCase();
74
+ let index = 0;
75
+ for (const letter of letters) {
76
+ index = index * 26 + (letter.charCodeAt(0) - 64);
77
+ }
78
+ return Math.max(0, index - 1);
79
+ }
80
+ function coerceCell(value) {
81
+ const trimmed = value.trim();
82
+ if (!trimmed)
83
+ return null;
84
+ if (/^(true|false)$/i.test(trimmed))
85
+ return /^true$/i.test(trimmed);
86
+ const parsed = Number(trimmed.replace(/,/g, ''));
87
+ if (Number.isFinite(parsed) && /^[+-]?(?:\d+\.?\d*|\.\d+)(?:e[+-]?\d+)?$/i.test(trimmed.replace(/,/g, ''))) {
88
+ return parsed;
89
+ }
90
+ return trimmed;
91
+ }
92
+ function getCellValue(cellXml, sharedStrings) {
93
+ const openTag = cellXml.match(/^<c\b([^>]*)>/)?.[1] ?? '';
94
+ const attrs = parseAttributes(openTag);
95
+ const rawValue = cellXml.match(/<v\b[^>]*>([\s\S]*?)<\/v>/)?.[1];
96
+ if (attrs.t === 's' && rawValue != null) {
97
+ return sharedStrings[Number(rawValue)] ?? null;
98
+ }
99
+ if (attrs.t === 'inlineStr') {
100
+ const inlineText = [...cellXml.matchAll(/<t\b[^>]*>([\s\S]*?)<\/t>/g)]
101
+ .map((match) => decodeXml(match[1]))
102
+ .join('');
103
+ return inlineText.trim() ? inlineText : null;
104
+ }
105
+ if (attrs.t === 'str' && rawValue != null) {
106
+ return decodeXml(rawValue);
107
+ }
108
+ if (attrs.t === 'b' && rawValue != null) {
109
+ return rawValue.trim() === '1';
110
+ }
111
+ return rawValue == null ? null : coerceCell(decodeXml(rawValue));
112
+ }
113
+ function uniquifyHeaders(headers) {
114
+ const counts = new Map();
115
+ return headers.map((header, index) => {
116
+ const base = header.trim() || `Column ${index + 1}`;
117
+ const count = counts.get(base) ?? 0;
118
+ counts.set(base, count + 1);
119
+ return count === 0 ? base : `${base} ${count + 1}`;
120
+ });
121
+ }
122
+ function parseWorksheet(name, xml, sharedStrings, maxRows) {
123
+ const warnings = [];
124
+ const matrix = [];
125
+ const rowPattern = /<row\b[^>]*>([\s\S]*?)<\/row>/g;
126
+ let rowMatch;
127
+ while ((rowMatch = rowPattern.exec(xml)) !== null) {
128
+ if (matrix.length > maxRows + 1) {
129
+ warnings.push(`Sampled first ${maxRows} data rows from sheet "${name}".`);
130
+ break;
131
+ }
132
+ const rowCells = [];
133
+ const cellPattern = /<c\b[^>]*>[\s\S]*?<\/c>/g;
134
+ let cellMatch;
135
+ while ((cellMatch = cellPattern.exec(rowMatch[1])) !== null) {
136
+ const ref = cellMatch[0].match(/^<c\b[^>]*\br="([^"]+)"/)?.[1];
137
+ const columnIndex = ref ? columnIndexFromRef(ref) : rowCells.length;
138
+ rowCells[columnIndex] = getCellValue(cellMatch[0], sharedStrings);
139
+ }
140
+ if (rowCells.some((cell) => cell != null && String(cell).trim().length > 0)) {
141
+ matrix.push(rowCells);
142
+ }
143
+ }
144
+ if (matrix.length === 0) {
145
+ return { name, rows: [], totalRowsSampled: 0, warnings: [...warnings, `Sheet "${name}" has no readable rows.`] };
146
+ }
147
+ const headers = uniquifyHeaders(matrix[0].map((cell, index) => String(cell ?? `Column ${index + 1}`).trim()));
148
+ const rows = matrix.slice(1, maxRows + 1).map((cells) => {
149
+ const row = {};
150
+ headers.forEach((header, index) => {
151
+ row[header] = cells[index] ?? null;
152
+ });
153
+ return row;
154
+ }).filter((row) => Object.values(row).some((cell) => cell != null && String(cell).trim().length > 0));
155
+ return { name, rows, totalRowsSampled: rows.length, warnings };
156
+ }
157
+ export async function parseXlsxFile(filePath, options = {}) {
158
+ const warnings = [];
159
+ const bytes = await readFile(filePath);
160
+ const entries = unzipSync(new Uint8Array(bytes));
161
+ const workbookXml = getZipText(entries, 'xl/workbook.xml');
162
+ const relsXml = getZipText(entries, 'xl/_rels/workbook.xml.rels');
163
+ const sharedStrings = parseSharedStrings(getZipText(entries, 'xl/sharedStrings.xml'));
164
+ const sheetRefs = parseWorkbookSheets(workbookXml);
165
+ const rels = parseWorkbookRelationships(relsXml);
166
+ const maxRows = Math.max(1, options.maxRows ?? DEFAULT_MAX_ROWS);
167
+ if (sheetRefs.length === 0) {
168
+ warnings.push('Workbook sheet metadata was not found.');
169
+ }
170
+ const sheetPaths = sheetRefs.length > 0
171
+ ? sheetRefs.map((sheet, index) => ({
172
+ name: sheet.name,
173
+ path: rels.get(sheet.relId) ?? `xl/worksheets/sheet${index + 1}.xml`,
174
+ }))
175
+ : Object.keys(entries)
176
+ .filter((entry) => /^xl\/worksheets\/sheet\d+\.xml$/i.test(entry))
177
+ .sort()
178
+ .map((path, index) => ({ name: `Sheet ${index + 1}`, path }));
179
+ const sheets = [];
180
+ for (const sheet of sheetPaths) {
181
+ const xml = getZipText(entries, sheet.path);
182
+ if (!xml) {
183
+ warnings.push(`Worksheet XML missing for "${sheet.name}".`);
184
+ continue;
185
+ }
186
+ sheets.push(parseWorksheet(sheet.name, xml, sharedStrings, maxRows));
187
+ }
188
+ return { sheets, warnings };
189
+ }
190
+ //# sourceMappingURL=xlsx.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"xlsx.js","sourceRoot":"","sources":["../../src/tabular/xlsx.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,QAAQ,CAAC;AAoB9C,MAAM,gBAAgB,GAAG,GAAG,CAAC;AAE7B,SAAS,SAAS,CAAC,KAAa;IAC9B,OAAO,KAAK;SACT,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;SACrB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;AAC5B,CAAC;AAED,SAAS,SAAS,CAAC,KAAa;IAC9B,OAAO,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,CAAC;AAClD,CAAC;AAED,SAAS,UAAU,CAAC,OAAmC,EAAE,IAAY;IACnE,MAAM,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IAC5C,MAAM,KAAK,GAAG,OAAO,CAAC,UAAU,CAAC,CAAC;IAClC,OAAO,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AAC9C,CAAC;AAED,SAAS,eAAe,CAAC,GAAW;IAClC,MAAM,KAAK,GAA2B,EAAE,CAAC;IACzC,MAAM,OAAO,GAAG,yCAAyC,CAAC;IAC1D,IAAI,KAA6B,CAAC;IAClC,OAAO,CAAC,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAC5C,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IACxC,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,kBAAkB,CAAC,GAAY;IACtC,IAAI,CAAC,GAAG;QAAE,OAAO,EAAE,CAAC;IACpB,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,SAAS,GAAG,8BAA8B,CAAC;IACjD,IAAI,KAA6B,CAAC;IAElC,OAAO,CAAC,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAC9C,MAAM,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACzB,MAAM,KAAK,GAAG,CAAC,GAAG,OAAO,CAAC,QAAQ,CAAC,4BAA4B,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,SAAS,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAC9G,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC;IACvE,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,mBAAmB,CAAC,WAAoB;IAC/C,IAAI,CAAC,WAAW;QAAE,OAAO,EAAE,CAAC;IAC5B,MAAM,MAAM,GAAe,EAAE,CAAC;IAC9B,MAAM,YAAY,GAAG,uBAAuB,CAAC;IAC7C,IAAI,KAA6B,CAAC;IAElC,OAAO,CAAC,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACzD,MAAM,KAAK,GAAG,eAAe,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACxC,MAAM,KAAK,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC;QAC5B,IAAI,KAAK,CAAC,IAAI,IAAI,KAAK,EAAE,CAAC;YACxB,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC;QAC3C,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,0BAA0B,CAAC,OAAgB;IAClD,MAAM,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAC;IACvC,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAC1B,MAAM,UAAU,GAAG,8BAA8B,CAAC;IAClD,IAAI,KAA6B,CAAC;IAElC,OAAO,CAAC,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QACnD,MAAM,KAAK,GAAG,eAAe,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACxC,IAAI,KAAK,CAAC,EAAE,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;YAC7B,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,KAAK,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,EAAE,CAAC;YAClH,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,EAAE,MAAM,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC;QACjD,CAAC;IACH,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,kBAAkB,CAAC,OAAe;IACzC,MAAM,OAAO,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;IACrE,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,KAAK,GAAG,KAAK,GAAG,EAAE,GAAG,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;IACnD,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;AAChC,CAAC;AAED,SAAS,UAAU,CAAC,KAAa;IAC/B,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAC7B,IAAI,CAAC,OAAO;QAAE,OAAO,IAAI,CAAC;IAC1B,IAAI,iBAAiB,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACpE,MAAM,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC;IACjD,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,2CAA2C,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC;QAC3G,OAAO,MAAM,CAAC;IAChB,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,YAAY,CAAC,OAAe,EAAE,aAAuB;IAC5D,MAAM,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,eAAe,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAC1D,MAAM,KAAK,GAAG,eAAe,CAAC,OAAO,CAAC,CAAC;IACvC,MAAM,QAAQ,GAAG,OAAO,CAAC,KAAK,CAAC,2BAA2B,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAEjE,IAAI,KAAK,CAAC,CAAC,KAAK,GAAG,IAAI,QAAQ,IAAI,IAAI,EAAE,CAAC;QACxC,OAAO,aAAa,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,IAAI,IAAI,CAAC;IACjD,CAAC;IAED,IAAI,KAAK,CAAC,CAAC,KAAK,WAAW,EAAE,CAAC;QAC5B,MAAM,UAAU,GAAG,CAAC,GAAG,OAAO,CAAC,QAAQ,CAAC,4BAA4B,CAAC,CAAC;aACnE,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;aACnC,IAAI,CAAC,EAAE,CAAC,CAAC;QACZ,OAAO,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC;IAC/C,CAAC;IAED,IAAI,KAAK,CAAC,CAAC,KAAK,KAAK,IAAI,QAAQ,IAAI,IAAI,EAAE,CAAC;QAC1C,OAAO,SAAS,CAAC,QAAQ,CAAC,CAAC;IAC7B,CAAC;IAED,IAAI,KAAK,CAAC,CAAC,KAAK,GAAG,IAAI,QAAQ,IAAI,IAAI,EAAE,CAAC;QACxC,OAAO,QAAQ,CAAC,IAAI,EAAE,KAAK,GAAG,CAAC;IACjC,CAAC;IAED,OAAO,QAAQ,IAAI,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,UAAU,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC,CAAC;AACnE,CAAC;AAED,SAAS,eAAe,CAAC,OAAiB;IACxC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE;QACnC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,EAAE,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,CAAC;QACpD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACpC,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;QAC5B,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;IACrD,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,cAAc,CAAC,IAAY,EAAE,GAAW,EAAE,aAAuB,EAAE,OAAe;IACzF,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,MAAM,GAAoB,EAAE,CAAC;IACnC,MAAM,UAAU,GAAG,gCAAgC,CAAC;IACpD,IAAI,QAAgC,CAAC;IAErC,OAAO,CAAC,QAAQ,GAAG,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;QAClD,IAAI,MAAM,CAAC,MAAM,GAAG,OAAO,GAAG,CAAC,EAAE,CAAC;YAChC,QAAQ,CAAC,IAAI,CAAC,iBAAiB,OAAO,0BAA0B,IAAI,IAAI,CAAC,CAAC;YAC1E,MAAM;QACR,CAAC;QAED,MAAM,QAAQ,GAAkB,EAAE,CAAC;QACnC,MAAM,WAAW,GAAG,0BAA0B,CAAC;QAC/C,IAAI,SAAiC,CAAC;QACtC,OAAO,CAAC,SAAS,GAAG,WAAW,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YAC5D,MAAM,GAAG,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,yBAAyB,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;YAC/D,MAAM,WAAW,GAAG,GAAG,CAAC,CAAC,CAAC,kBAAkB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC;YACpE,QAAQ,CAAC,WAAW,CAAC,GAAG,YAAY,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,aAAa,CAAC,CAAC;QACpE,CAAC;QAED,IAAI,QAAQ,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,IAAI,IAAI,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,EAAE,CAAC;YAC5E,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QACxB,CAAC;IACH,CAAC;IAED,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,gBAAgB,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,GAAG,QAAQ,EAAE,UAAU,IAAI,yBAAyB,CAAC,EAAE,CAAC;IACnH,CAAC;IAED,MAAM,OAAO,GAAG,eAAe,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,IAAI,UAAU,KAAK,GAAG,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAC9G,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE;QACtD,MAAM,GAAG,GAAe,EAAE,CAAC;QAC3B,OAAO,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE;YAChC,GAAG,CAAC,MAAM,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC;QACrC,CAAC,CAAC,CAAC;QACH,OAAO,GAAG,CAAC;IACb,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,IAAI,IAAI,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC;IAEtG,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,gBAAgB,EAAE,IAAI,CAAC,MAAM,EAAE,QAAQ,EAAE,CAAC;AACjE,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,aAAa,CAAC,QAAgB,EAAE,UAAgC,EAAE;IACtF,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACvC,MAAM,OAAO,GAAG,SAAS,CAAC,IAAI,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC;IACjD,MAAM,WAAW,GAAG,UAAU,CAAC,OAAO,EAAE,iBAAiB,CAAC,CAAC;IAC3D,MAAM,OAAO,GAAG,UAAU,CAAC,OAAO,EAAE,4BAA4B,CAAC,CAAC;IAClE,MAAM,aAAa,GAAG,kBAAkB,CAAC,UAAU,CAAC,OAAO,EAAE,sBAAsB,CAAC,CAAC,CAAC;IACtF,MAAM,SAAS,GAAG,mBAAmB,CAAC,WAAW,CAAC,CAAC;IACnD,MAAM,IAAI,GAAG,0BAA0B,CAAC,OAAO,CAAC,CAAC;IACjD,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,OAAO,IAAI,gBAAgB,CAAC,CAAC;IAEjE,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,QAAQ,CAAC,IAAI,CAAC,wCAAwC,CAAC,CAAC;IAC1D,CAAC;IAED,MAAM,UAAU,GAAG,SAAS,CAAC,MAAM,GAAG,CAAC;QACrC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;YAC/B,IAAI,EAAE,KAAK,CAAC,IAAI;YAChB,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,sBAAsB,KAAK,GAAG,CAAC,MAAM;SACrE,CAAC,CAAC;QACL,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC;aACjB,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,kCAAkC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;aACjE,IAAI,EAAE;aACN,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,SAAS,KAAK,GAAG,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IAEpE,MAAM,MAAM,GAA0B,EAAE,CAAC;IACzC,KAAK,MAAM,KAAK,IAAI,UAAU,EAAE,CAAC;QAC/B,MAAM,GAAG,GAAG,UAAU,CAAC,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;QAC5C,IAAI,CAAC,GAAG,EAAE,CAAC;YACT,QAAQ,CAAC,IAAI,CAAC,8BAA8B,KAAK,CAAC,IAAI,IAAI,CAAC,CAAC;YAC5D,SAAS;QACX,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,cAAc,CAAC,KAAK,CAAC,IAAI,EAAE,GAAG,EAAE,aAAa,EAAE,OAAO,CAAC,CAAC,CAAC;IACvE,CAAC;IAED,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAC;AAC9B,CAAC"}
@@ -42,6 +42,10 @@ export interface GeotechDocumentInsight extends ParseSafety {
42
42
  rawLLMText: string;
43
43
  latencyMs: number;
44
44
  }
45
+ export declare function extractGeotechDocumentDeterministicFactsFromText(pageText: string, context?: GeotechDocumentContext, options?: {
46
+ warning?: string;
47
+ forcePartial?: boolean;
48
+ }): GeotechDocumentInsight | null;
45
49
  export declare function extractGeotechDocumentFactsFromText(pageText: string, config: LLMConfig, context?: GeotechDocumentContext): Promise<GeotechDocumentInsight>;
46
50
  export declare function interpretGeotechDocumentPage(imageBase64: string, mimeType: string, config: LLMConfig, context?: GeotechDocumentContext): Promise<GeotechDocumentInsight>;
47
51
  //# sourceMappingURL=geotech-document.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"geotech-document.d.ts","sourceRoot":"","sources":["../../src/vision/geotech-document.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAEjD,OAAO,EAML,KAAK,WAAW,EAEjB,MAAM,YAAY,CAAC;AAGpB,YAAY,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AA4I3D,QAAA,MAAM,cAAc,oEAAqE,CAAC;AAE1F,MAAM,WAAW,0BAA0B;IACzC,IAAI,EAAE,OAAO,cAAc,CAAC,MAAM,CAAC,CAAC;IACpC,WAAW,EAAE,MAAM,CAAC;IACpB,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,SAAS,EAAE,MAAM,GAAG,IAAI,CAAC;CAC1B;AAED,MAAM,WAAW,6BAA6B;IAC5C,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB;AAED,MAAM,WAAW,2BAA2B;IAC1C,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,IAAI,EAAE,MAAM,GAAG,IAAI,CAAC;IACpB,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB;AAED,MAAM,WAAW,sBAAsB;IACrC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,qBAAqB,CAAC,EAAE,OAAO,CAAC;CACjC;AAED,MAAM,WAAW,sBAAuB,SAAQ,WAAW;IACzD,aAAa,EAAE,MAAM,GAAG,IAAI,CAAC;IAC7B,KAAK,EAAE,MAAM,GAAG,IAAI,CAAC;IACrB,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,SAAS,EAAE,0BAA0B,EAAE,CAAC;IACxC,eAAe,EAAE,6BAA6B,EAAE,CAAC;IACjD,UAAU,EAAE,2BAA2B,EAAE,CAAC;IAC1C,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,eAAe,EAAE,MAAM,EAAE,CAAC;IAC1B,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;CACnB;AA6eD,wBAAsB,mCAAmC,CACvD,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,SAAS,EACjB,OAAO,GAAE,sBAA2B,GACnC,OAAO,CAAC,sBAAsB,CAAC,CA+IjC;AAED,wBAAsB,4BAA4B,CAChD,WAAW,EAAE,MAAM,EACnB,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,SAAS,EACjB,OAAO,GAAE,sBAA2B,GACnC,OAAO,CAAC,sBAAsB,CAAC,CA4BjC"}
1
+ {"version":3,"file":"geotech-document.d.ts","sourceRoot":"","sources":["../../src/vision/geotech-document.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAEjD,OAAO,EAML,KAAK,WAAW,EAEjB,MAAM,YAAY,CAAC;AAGpB,YAAY,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AA4I3D,QAAA,MAAM,cAAc,oEAAqE,CAAC;AAE1F,MAAM,WAAW,0BAA0B;IACzC,IAAI,EAAE,OAAO,cAAc,CAAC,MAAM,CAAC,CAAC;IACpC,WAAW,EAAE,MAAM,CAAC;IACpB,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,SAAS,EAAE,MAAM,GAAG,IAAI,CAAC;CAC1B;AAED,MAAM,WAAW,6BAA6B;IAC5C,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB;AAED,MAAM,WAAW,2BAA2B;IAC1C,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,IAAI,EAAE,MAAM,GAAG,IAAI,CAAC;IACpB,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC;IACxB,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;CACxB;AAED,MAAM,WAAW,sBAAsB;IACrC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,qBAAqB,CAAC,EAAE,OAAO,CAAC;CACjC;AAED,MAAM,WAAW,sBAAuB,SAAQ,WAAW;IACzD,aAAa,EAAE,MAAM,GAAG,IAAI,CAAC;IAC7B,KAAK,EAAE,MAAM,GAAG,IAAI,CAAC;IACrB,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,SAAS,EAAE,0BAA0B,EAAE,CAAC;IACxC,eAAe,EAAE,6BAA6B,EAAE,CAAC;IACjD,UAAU,EAAE,2BAA2B,EAAE,CAAC;IAC1C,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,eAAe,EAAE,MAAM,EAAE,CAAC;IAC1B,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;CACnB;AAigBD,wBAAgB,gDAAgD,CAC9D,QAAQ,EAAE,MAAM,EAChB,OAAO,GAAE,sBAA2B,EACpC,OAAO,GAAE;IACP,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,YAAY,CAAC,EAAE,OAAO,CAAC;CACnB,GACL,sBAAsB,GAAG,IAAI,CAyB/B;AAED,wBAAsB,mCAAmC,CACvD,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,SAAS,EACjB,OAAO,GAAE,sBAA2B,GACnC,OAAO,CAAC,sBAAsB,CAAC,CAoIjC;AAED,wBAAsB,4BAA4B,CAChD,WAAW,EAAE,MAAM,EACnB,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,SAAS,EACjB,OAAO,GAAE,sBAA2B,GACnC,OAAO,CAAC,sBAAsB,CAAC,CA4BjC"}
@@ -487,6 +487,43 @@ function shouldShortCircuitToDeterministicFallback(insight) {
487
487
  }
488
488
  return false;
489
489
  }
490
+ function buildDeterministicGeotechDocumentInsight(normalizedPageText, context) {
491
+ const fallback = extractGeotechDocumentFallback(normalizedPageText);
492
+ if (!fallback.value) {
493
+ return null;
494
+ }
495
+ return buildGeotechDocumentInsightFromValue({
496
+ mergedValue: fallback.value,
497
+ baseStatus: fallback.baseStatus,
498
+ warnings: fallback.warnings,
499
+ normalizedPageText,
500
+ context,
501
+ rawLLMText: '',
502
+ latencyMs: 0,
503
+ });
504
+ }
505
+ export function extractGeotechDocumentDeterministicFactsFromText(pageText, context = {}, options = {}) {
506
+ const normalizedPageText = pageText.replace(/\s+/g, ' ').trim();
507
+ if (!normalizedPageText) {
508
+ return null;
509
+ }
510
+ const insight = buildDeterministicGeotechDocumentInsight(normalizedPageText, context);
511
+ if (!insight) {
512
+ return null;
513
+ }
514
+ const parseStatus = options.forcePartial && insight.parseStatus !== 'failed'
515
+ ? 'partial'
516
+ : insight.parseStatus;
517
+ return {
518
+ ...insight,
519
+ parseStatus,
520
+ confidence: options.forcePartial ? Math.min(insight.confidence, 68) : insight.confidence,
521
+ canAutoProceed: options.forcePartial ? false : insight.canAutoProceed,
522
+ warnings: options.warning
523
+ ? combineWarnings(insight.warnings, [options.warning])
524
+ : insight.warnings,
525
+ };
526
+ }
490
527
  export async function extractGeotechDocumentFactsFromText(pageText, config, context = {}) {
491
528
  const normalizedPageText = pageText.replace(/\s+/g, ' ').trim();
492
529
  if (!normalizedPageText) {
@@ -514,17 +551,7 @@ export async function extractGeotechDocumentFactsFromText(pageText, config, cont
514
551
  ].filter((value) => Boolean(value));
515
552
  const sharedContext = contextParts.join('\n');
516
553
  const fallback = extractGeotechDocumentFallback(normalizedPageText);
517
- const deterministicFallbackInsight = fallback.value
518
- ? buildGeotechDocumentInsightFromValue({
519
- mergedValue: fallback.value,
520
- baseStatus: fallback.baseStatus,
521
- warnings: fallback.warnings,
522
- normalizedPageText,
523
- context,
524
- rawLLMText: '',
525
- latencyMs: 0,
526
- })
527
- : null;
554
+ const deterministicFallbackInsight = buildDeterministicGeotechDocumentInsight(normalizedPageText, context);
528
555
  if (deterministicFallbackInsight && shouldShortCircuitToDeterministicFallback(deterministicFallbackInsight)) {
529
556
  const deterministicConfidence = Math.max(deterministicFallbackInsight.confidence, 72);
530
557
  return {