@geotechcli/core 0.4.22 → 0.4.23

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 (144) hide show
  1. package/dist/agents/brain.d.ts.map +1 -1
  2. package/dist/agents/brain.js +2 -1
  3. package/dist/agents/brain.js.map +1 -1
  4. package/dist/agents/data-tools.js +759 -0
  5. package/dist/agents/data-tools.js.map +1 -1
  6. package/dist/agents/swarm.d.ts.map +1 -1
  7. package/dist/agents/swarm.js +22 -2
  8. package/dist/agents/swarm.js.map +1 -1
  9. package/dist/agents/tool-runtime.d.ts +7 -0
  10. package/dist/agents/tool-runtime.d.ts.map +1 -0
  11. package/dist/agents/tool-runtime.js +9 -0
  12. package/dist/agents/tool-runtime.js.map +1 -0
  13. package/dist/config/index.d.ts +4 -4
  14. package/dist/config/index.js +1 -1
  15. package/dist/config/index.js.map +1 -1
  16. package/dist/geo/coordinates.d.ts +40 -0
  17. package/dist/geo/coordinates.d.ts.map +1 -0
  18. package/dist/geo/coordinates.js +461 -0
  19. package/dist/geo/coordinates.js.map +1 -0
  20. package/dist/geo/index.d.ts +1 -0
  21. package/dist/geo/index.d.ts.map +1 -1
  22. package/dist/geo/index.js +1 -0
  23. package/dist/geo/index.js.map +1 -1
  24. package/dist/index.d.ts +3 -2
  25. package/dist/index.d.ts.map +1 -1
  26. package/dist/index.js +3 -2
  27. package/dist/index.js.map +1 -1
  28. package/dist/ingest/ags.d.ts +3 -0
  29. package/dist/ingest/ags.d.ts.map +1 -1
  30. package/dist/ingest/ags.js +98 -9
  31. package/dist/ingest/ags.js.map +1 -1
  32. package/dist/ingest/cpt.d.ts +4 -0
  33. package/dist/ingest/cpt.d.ts.map +1 -1
  34. package/dist/ingest/cpt.js +87 -25
  35. package/dist/ingest/cpt.js.map +1 -1
  36. package/dist/ingest/document-inputs.d.ts +37 -0
  37. package/dist/ingest/document-inputs.d.ts.map +1 -0
  38. package/dist/ingest/document-inputs.js +197 -0
  39. package/dist/ingest/document-inputs.js.map +1 -0
  40. package/dist/ingest/geotech-document.d.ts +118 -0
  41. package/dist/ingest/geotech-document.d.ts.map +1 -0
  42. package/dist/ingest/geotech-document.js +1006 -0
  43. package/dist/ingest/geotech-document.js.map +1 -0
  44. package/dist/ingest/geotech-extract.d.ts +86 -0
  45. package/dist/ingest/geotech-extract.d.ts.map +1 -0
  46. package/dist/ingest/geotech-extract.js +652 -0
  47. package/dist/ingest/geotech-extract.js.map +1 -0
  48. package/dist/ingest/geotech-schemas.d.ts +248 -0
  49. package/dist/ingest/geotech-schemas.d.ts.map +1 -0
  50. package/dist/ingest/geotech-schemas.js +150 -0
  51. package/dist/ingest/geotech-schemas.js.map +1 -0
  52. package/dist/ingest/index.d.ts +8 -0
  53. package/dist/ingest/index.d.ts.map +1 -1
  54. package/dist/ingest/index.js +8 -0
  55. package/dist/ingest/index.js.map +1 -1
  56. package/dist/ingest/ingest-job-child.d.ts +2 -0
  57. package/dist/ingest/ingest-job-child.d.ts.map +1 -0
  58. package/dist/ingest/ingest-job-child.js +45 -0
  59. package/dist/ingest/ingest-job-child.js.map +1 -0
  60. package/dist/ingest/job-store.d.ts +117 -0
  61. package/dist/ingest/job-store.d.ts.map +1 -0
  62. package/dist/ingest/job-store.js +541 -0
  63. package/dist/ingest/job-store.js.map +1 -0
  64. package/dist/ingest/job-worker.d.ts +24 -0
  65. package/dist/ingest/job-worker.d.ts.map +1 -0
  66. package/dist/ingest/job-worker.js +1129 -0
  67. package/dist/ingest/job-worker.js.map +1 -0
  68. package/dist/ingest/pdf.d.ts +102 -0
  69. package/dist/ingest/pdf.d.ts.map +1 -0
  70. package/dist/ingest/pdf.js +1544 -0
  71. package/dist/ingest/pdf.js.map +1 -0
  72. package/dist/ingest/review-store.d.ts +215 -0
  73. package/dist/ingest/review-store.d.ts.map +1 -0
  74. package/dist/ingest/review-store.js +1995 -0
  75. package/dist/ingest/review-store.js.map +1 -0
  76. package/dist/llm/capabilities.d.ts +8 -0
  77. package/dist/llm/capabilities.d.ts.map +1 -0
  78. package/dist/llm/capabilities.js +73 -0
  79. package/dist/llm/capabilities.js.map +1 -0
  80. package/dist/llm/index.d.ts +3 -2
  81. package/dist/llm/index.d.ts.map +1 -1
  82. package/dist/llm/index.js +2 -1
  83. package/dist/llm/index.js.map +1 -1
  84. package/dist/llm/providers/anthropic.d.ts +6 -0
  85. package/dist/llm/providers/anthropic.d.ts.map +1 -1
  86. package/dist/llm/providers/anthropic.js +10 -1
  87. package/dist/llm/providers/anthropic.js.map +1 -1
  88. package/dist/llm/providers/hosted-beta.d.ts +6 -0
  89. package/dist/llm/providers/hosted-beta.d.ts.map +1 -1
  90. package/dist/llm/providers/hosted-beta.js +40 -10
  91. package/dist/llm/providers/hosted-beta.js.map +1 -1
  92. package/dist/llm/providers/huggingface.d.ts +6 -0
  93. package/dist/llm/providers/huggingface.d.ts.map +1 -1
  94. package/dist/llm/providers/huggingface.js +21 -1
  95. package/dist/llm/providers/huggingface.js.map +1 -1
  96. package/dist/llm/providers/openai-compatible.d.ts +6 -0
  97. package/dist/llm/providers/openai-compatible.d.ts.map +1 -1
  98. package/dist/llm/providers/openai-compatible.js +21 -1
  99. package/dist/llm/providers/openai-compatible.js.map +1 -1
  100. package/dist/llm/providers/zhipu.d.ts +6 -0
  101. package/dist/llm/providers/zhipu.d.ts.map +1 -1
  102. package/dist/llm/providers/zhipu.js +15 -1
  103. package/dist/llm/providers/zhipu.js.map +1 -1
  104. package/dist/llm/router.d.ts +7 -0
  105. package/dist/llm/router.d.ts.map +1 -1
  106. package/dist/llm/router.js +33 -13
  107. package/dist/llm/router.js.map +1 -1
  108. package/dist/llm/types.d.ts +22 -4
  109. package/dist/llm/types.d.ts.map +1 -1
  110. package/dist/llm/types.js.map +1 -1
  111. package/dist/meta/metadata.json +1 -1
  112. package/dist/report/html.d.ts +3 -0
  113. package/dist/report/html.d.ts.map +1 -0
  114. package/dist/report/html.js +626 -0
  115. package/dist/report/html.js.map +1 -0
  116. package/dist/report/index.d.ts +2 -0
  117. package/dist/report/index.d.ts.map +1 -1
  118. package/dist/report/index.js +2 -0
  119. package/dist/report/index.js.map +1 -1
  120. package/dist/report/ingest-dossier.d.ts +81 -0
  121. package/dist/report/ingest-dossier.d.ts.map +1 -0
  122. package/dist/report/ingest-dossier.js +324 -0
  123. package/dist/report/ingest-dossier.js.map +1 -0
  124. package/dist/storage/index.d.ts +5 -0
  125. package/dist/storage/index.d.ts.map +1 -1
  126. package/dist/storage/index.js +12 -6
  127. package/dist/storage/index.js.map +1 -1
  128. package/dist/vision/geotech-document.d.ts +46 -0
  129. package/dist/vision/geotech-document.d.ts.map +1 -0
  130. package/dist/vision/geotech-document.js +576 -0
  131. package/dist/vision/geotech-document.js.map +1 -0
  132. package/dist/vision/index.d.ts +31 -0
  133. package/dist/vision/index.d.ts.map +1 -1
  134. package/dist/vision/index.js +659 -27
  135. package/dist/vision/index.js.map +1 -1
  136. package/dist/vision/ocr.d.ts +29 -0
  137. package/dist/vision/ocr.d.ts.map +1 -0
  138. package/dist/vision/ocr.js +287 -0
  139. package/dist/vision/ocr.js.map +1 -0
  140. package/dist/vision/preprocess.d.ts +26 -0
  141. package/dist/vision/preprocess.d.ts.map +1 -0
  142. package/dist/vision/preprocess.js +194 -0
  143. package/dist/vision/preprocess.js.map +1 -0
  144. package/package.json +5 -1
@@ -0,0 +1,197 @@
1
+ import { readFileSync } from 'node:fs';
2
+ import { extname } from 'node:path';
3
+ import { PDFDocument } from 'pdf-lib';
4
+ import { extractPrimaryPdfPageImages, renderPdfPageImage, } from './pdf.js';
5
+ const MIME_TYPES = {
6
+ jpg: 'image/jpeg',
7
+ jpeg: 'image/jpeg',
8
+ png: 'image/png',
9
+ gif: 'image/gif',
10
+ webp: 'image/webp',
11
+ pdf: 'application/pdf',
12
+ };
13
+ const PROVIDER_SAFE_IMAGE_MIME_TYPES = new Set([
14
+ 'image/jpeg',
15
+ 'image/png',
16
+ 'image/gif',
17
+ 'image/webp',
18
+ ]);
19
+ export function readDocumentVisionInput(filePath) {
20
+ const buffer = readFileSync(filePath);
21
+ const ext = extname(filePath).slice(1).toLowerCase();
22
+ const kind = ext === 'pdf' ? 'pdf' : MIME_TYPES[ext] ? 'image' : 'unknown';
23
+ return {
24
+ base64: buffer.toString('base64'),
25
+ mimeType: MIME_TYPES[ext] ?? 'image/png',
26
+ fileBytes: buffer.length,
27
+ filePath,
28
+ ext,
29
+ kind,
30
+ };
31
+ }
32
+ export async function countDocumentPdfPages(filePath) {
33
+ const buffer = readFileSync(filePath);
34
+ const quickCount = buffer
35
+ .toString('latin1')
36
+ .match(/\/Type\s*\/Page(?!s)\b/g)?.length ?? 0;
37
+ if (quickCount > 0) {
38
+ return quickCount;
39
+ }
40
+ const source = await PDFDocument.load(buffer, { ignoreEncryption: true });
41
+ return source.getPageCount();
42
+ }
43
+ function extFromMimeType(mimeType) {
44
+ switch (mimeType) {
45
+ case 'image/jpeg':
46
+ return 'jpg';
47
+ case 'image/png':
48
+ return 'png';
49
+ case 'image/gif':
50
+ return 'gif';
51
+ case 'image/webp':
52
+ return 'webp';
53
+ default:
54
+ return 'bin';
55
+ }
56
+ }
57
+ function imageBufferMatchesMimeType(buffer, mimeType) {
58
+ const bytes = Buffer.from(buffer);
59
+ if (bytes.length < 12) {
60
+ return false;
61
+ }
62
+ switch (mimeType) {
63
+ case 'image/png':
64
+ return bytes.subarray(0, 8).equals(Buffer.from([0x89, 0x50, 0x4E, 0x47, 0x0D, 0x0A, 0x1A, 0x0A]));
65
+ case 'image/jpeg':
66
+ return bytes[0] === 0xFF && bytes[1] === 0xD8;
67
+ case 'image/gif':
68
+ return bytes.subarray(0, 6).toString('ascii') === 'GIF87a'
69
+ || bytes.subarray(0, 6).toString('ascii') === 'GIF89a';
70
+ case 'image/webp':
71
+ return bytes.subarray(0, 4).toString('ascii') === 'RIFF'
72
+ && bytes.subarray(8, 12).toString('ascii') === 'WEBP';
73
+ default:
74
+ return false;
75
+ }
76
+ }
77
+ function isProviderSafeRasterImage(image) {
78
+ return PROVIDER_SAFE_IMAGE_MIME_TYPES.has(image.mimeType)
79
+ && imageBufferMatchesMimeType(image.data, image.mimeType);
80
+ }
81
+ function createDocumentPdfPageNormalizedArtifact(input) {
82
+ return {
83
+ kind: input.kind,
84
+ source: input.source,
85
+ mimeType: input.mimeType,
86
+ fileBytes: input.fileBytes,
87
+ textSource: input.inspectionArtifact?.textSource ?? 'none',
88
+ textQuality: input.inspectionArtifact?.textQuality ?? null,
89
+ warnings: [
90
+ ...new Set([
91
+ ...(input.inspectionArtifact?.warnings ?? []),
92
+ ...(input.warnings ?? []),
93
+ ]),
94
+ ],
95
+ };
96
+ }
97
+ export async function readDocumentPdfPageInputs(filePath, options) {
98
+ const buffer = readFileSync(filePath);
99
+ const source = await PDFDocument.load(buffer, { ignoreEncryption: true });
100
+ const totalPages = source.getPageCount();
101
+ const extractPageImages = options?.dependencies?.extractPageImages ?? extractPrimaryPdfPageImages;
102
+ const renderPageImage = options?.dependencies?.renderPageImage ?? renderPdfPageImage;
103
+ const shouldAttemptRasterExtraction = options?.preferExtractedPageImages !== false
104
+ && options?.inspection != null;
105
+ const extractedImages = !shouldAttemptRasterExtraction
106
+ ? []
107
+ : await extractPageImages(buffer);
108
+ const extractedImageByPage = new Map(extractedImages.map((image) => [image.pageNumber, image]));
109
+ const pageInputs = [];
110
+ for (let index = 0; index < totalPages; index++) {
111
+ const pageNumber = index + 1;
112
+ const inspectionPage = options?.inspection?.pages[index];
113
+ const extractedImage = extractedImageByPage.get(pageNumber);
114
+ const shouldPreferRasterImage = inspectionPage?.classification === 'image-only'
115
+ || inspectionPage?.classification === 'text-unreadable'
116
+ || inspectionPage?.normalizedArtifact?.textSource === 'native-text-low-quality';
117
+ if (shouldPreferRasterImage) {
118
+ const rasterWarnings = [];
119
+ if (extractedImage) {
120
+ if (isProviderSafeRasterImage(extractedImage)) {
121
+ pageInputs.push({
122
+ base64: Buffer.from(extractedImage.data).toString('base64'),
123
+ mimeType: extractedImage.mimeType,
124
+ fileBytes: extractedImage.byteLength,
125
+ filePath,
126
+ ext: extFromMimeType(extractedImage.mimeType),
127
+ kind: 'image',
128
+ pageNumber,
129
+ totalPages,
130
+ sourceKind: 'raster-image',
131
+ normalizedArtifact: createDocumentPdfPageNormalizedArtifact({
132
+ kind: 'image',
133
+ source: 'xobject-raster',
134
+ mimeType: extractedImage.mimeType,
135
+ fileBytes: extractedImage.byteLength,
136
+ inspectionArtifact: inspectionPage?.normalizedArtifact,
137
+ warnings: extractedImage.warnings,
138
+ }),
139
+ });
140
+ continue;
141
+ }
142
+ rasterWarnings.push(`Extracted raster image for page ${pageNumber} was not a provider-safe ${extractedImage.mimeType} payload, so the full page was rerendered to PNG.`);
143
+ }
144
+ const renderedPage = await renderPageImage(buffer, pageNumber);
145
+ if (renderedPage) {
146
+ pageInputs.push({
147
+ base64: Buffer.from(renderedPage.data).toString('base64'),
148
+ mimeType: renderedPage.mimeType,
149
+ fileBytes: renderedPage.byteLength,
150
+ filePath,
151
+ ext: extFromMimeType(renderedPage.mimeType),
152
+ kind: 'image',
153
+ pageNumber,
154
+ totalPages,
155
+ sourceKind: 'raster-image',
156
+ normalizedArtifact: createDocumentPdfPageNormalizedArtifact({
157
+ kind: 'image',
158
+ source: 'full-page-raster',
159
+ mimeType: renderedPage.mimeType,
160
+ fileBytes: renderedPage.byteLength,
161
+ inspectionArtifact: inspectionPage?.normalizedArtifact,
162
+ warnings: [...rasterWarnings, ...renderedPage.warnings],
163
+ }),
164
+ });
165
+ continue;
166
+ }
167
+ }
168
+ const pageDoc = await PDFDocument.create();
169
+ const [copiedPage] = await pageDoc.copyPages(source, [index]);
170
+ pageDoc.addPage(copiedPage);
171
+ const pageBytes = await pageDoc.save();
172
+ const pageBuffer = Buffer.from(pageBytes);
173
+ pageInputs.push({
174
+ base64: pageBuffer.toString('base64'),
175
+ mimeType: 'application/pdf',
176
+ fileBytes: pageBuffer.length,
177
+ filePath,
178
+ ext: 'pdf',
179
+ kind: 'pdf',
180
+ pageNumber,
181
+ totalPages,
182
+ sourceKind: 'pdf-page',
183
+ normalizedArtifact: createDocumentPdfPageNormalizedArtifact({
184
+ kind: 'pdf',
185
+ source: 'native-pdf-page',
186
+ mimeType: 'application/pdf',
187
+ fileBytes: pageBuffer.length,
188
+ inspectionArtifact: inspectionPage?.normalizedArtifact,
189
+ warnings: shouldPreferRasterImage && extractedImage && !isProviderSafeRasterImage(extractedImage)
190
+ ? [`Extracted raster image for page ${pageNumber} could not be normalized into a provider-safe image payload, so the job fell back to a native PDF page.`]
191
+ : [],
192
+ }),
193
+ });
194
+ }
195
+ return pageInputs;
196
+ }
197
+ //# sourceMappingURL=document-inputs.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"document-inputs.js","sourceRoot":"","sources":["../../src/ingest/document-inputs.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AACtC,OAAO,EACL,2BAA2B,EAC3B,kBAAkB,GAKnB,MAAM,UAAU,CAAC;AAElB,MAAM,UAAU,GAA2B;IACzC,GAAG,EAAE,YAAY;IACjB,IAAI,EAAE,YAAY;IAClB,GAAG,EAAE,WAAW;IAChB,GAAG,EAAE,WAAW;IAChB,IAAI,EAAE,YAAY;IAClB,GAAG,EAAE,iBAAiB;CACvB,CAAC;AAuCF,MAAM,8BAA8B,GAAG,IAAI,GAAG,CAAC;IAC7C,YAAY;IACZ,WAAW;IACX,WAAW;IACX,YAAY;CACb,CAAC,CAAC;AAEH,MAAM,UAAU,uBAAuB,CAAC,QAAgB;IACtD,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;IACrD,MAAM,IAAI,GAAsB,GAAG,KAAK,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS,CAAC;IAE9F,OAAO;QACL,MAAM,EAAE,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC;QACjC,QAAQ,EAAE,UAAU,CAAC,GAAG,CAAC,IAAI,WAAW;QACxC,SAAS,EAAE,MAAM,CAAC,MAAM;QACxB,QAAQ;QACR,GAAG;QACH,IAAI;KACL,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,qBAAqB,CAAC,QAAgB;IAC1D,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,UAAU,GAAG,MAAM;SACtB,QAAQ,CAAC,QAAQ,CAAC;SAClB,KAAK,CAAC,yBAAyB,CAAC,EAAE,MAAM,IAAI,CAAC,CAAC;IACjD,IAAI,UAAU,GAAG,CAAC,EAAE,CAAC;QACnB,OAAO,UAAU,CAAC;IACpB,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,WAAW,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,gBAAgB,EAAE,IAAI,EAAE,CAAC,CAAC;IAC1E,OAAO,MAAM,CAAC,YAAY,EAAE,CAAC;AAC/B,CAAC;AAED,SAAS,eAAe,CAAC,QAAgB;IACvC,QAAQ,QAAQ,EAAE,CAAC;QACjB,KAAK,YAAY;YACf,OAAO,KAAK,CAAC;QACf,KAAK,WAAW;YACd,OAAO,KAAK,CAAC;QACf,KAAK,WAAW;YACd,OAAO,KAAK,CAAC;QACf,KAAK,YAAY;YACf,OAAO,MAAM,CAAC;QAChB;YACE,OAAO,KAAK,CAAC;IACjB,CAAC;AACH,CAAC;AAED,SAAS,0BAA0B,CAAC,MAAkB,EAAE,QAAgB;IACtE,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAClC,IAAI,KAAK,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;QACtB,OAAO,KAAK,CAAC;IACf,CAAC;IAED,QAAQ,QAAQ,EAAE,CAAC;QACjB,KAAK,WAAW;YACd,OAAO,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC;QACpG,KAAK,YAAY;YACf,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,IAAI,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC;QAChD,KAAK,WAAW;YACd,OAAO,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,QAAQ;mBACrD,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,QAAQ,CAAC;QAC3D,KAAK,YAAY;YACf,OAAO,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,MAAM;mBACnD,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,MAAM,CAAC;QAC1D;YACE,OAAO,KAAK,CAAC;IACjB,CAAC;AACH,CAAC;AAED,SAAS,yBAAyB,CAAC,KAAyB;IAC1D,OAAO,8BAA8B,CAAC,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC;WACpD,0BAA0B,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;AAC9D,CAAC;AAED,SAAS,uCAAuC,CAAC,KAOhD;IACC,OAAO;QACL,IAAI,EAAE,KAAK,CAAC,IAAI;QAChB,MAAM,EAAE,KAAK,CAAC,MAAM;QACpB,QAAQ,EAAE,KAAK,CAAC,QAAQ;QACxB,SAAS,EAAE,KAAK,CAAC,SAAS;QAC1B,UAAU,EAAE,KAAK,CAAC,kBAAkB,EAAE,UAAU,IAAI,MAAM;QAC1D,WAAW,EAAE,KAAK,CAAC,kBAAkB,EAAE,WAAW,IAAI,IAAI;QAC1D,QAAQ,EAAE;YACR,GAAG,IAAI,GAAG,CAAC;gBACT,GAAG,CAAC,KAAK,CAAC,kBAAkB,EAAE,QAAQ,IAAI,EAAE,CAAC;gBAC7C,GAAG,CAAC,KAAK,CAAC,QAAQ,IAAI,EAAE,CAAC;aAC1B,CAAC;SACH;KACF,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,yBAAyB,CAC7C,QAAgB,EAChB,OAA0C;IAE1C,MAAM,MAAM,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACtC,MAAM,MAAM,GAAG,MAAM,WAAW,CAAC,IAAI,CAAC,MAAM,EAAE,EAAE,gBAAgB,EAAE,IAAI,EAAE,CAAC,CAAC;IAC1E,MAAM,UAAU,GAAG,MAAM,CAAC,YAAY,EAAE,CAAC;IACzC,MAAM,iBAAiB,GAAG,OAAO,EAAE,YAAY,EAAE,iBAAiB,IAAI,2BAA2B,CAAC;IAClG,MAAM,eAAe,GAAG,OAAO,EAAE,YAAY,EAAE,eAAe,IAAI,kBAAkB,CAAC;IACrF,MAAM,6BAA6B,GACjC,OAAO,EAAE,yBAAyB,KAAK,KAAK;WACzC,OAAO,EAAE,UAAU,IAAI,IAAI,CAAC;IACjC,MAAM,eAAe,GACnB,CAAC,6BAA6B;QAC5B,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,MAAM,iBAAiB,CAAC,MAAM,CAAC,CAAC;IACtC,MAAM,oBAAoB,GAAG,IAAI,GAAG,CAClC,eAAe,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,UAAU,EAAE,KAAK,CAAU,CAAC,CACnE,CAAC;IACF,MAAM,UAAU,GAA2B,EAAE,CAAC;IAE9C,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,UAAU,EAAE,KAAK,EAAE,EAAE,CAAC;QAChD,MAAM,UAAU,GAAG,KAAK,GAAG,CAAC,CAAC;QAC7B,MAAM,cAAc,GAAG,OAAO,EAAE,UAAU,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QACzD,MAAM,cAAc,GAAG,oBAAoB,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;QAC5D,MAAM,uBAAuB,GAC3B,cAAc,EAAE,cAAc,KAAK,YAAY;eAC5C,cAAc,EAAE,cAAc,KAAK,iBAAiB;eACpD,cAAc,EAAE,kBAAkB,EAAE,UAAU,KAAK,yBAAyB,CAAC;QAElF,IAAI,uBAAuB,EAAE,CAAC;YAC5B,MAAM,cAAc,GAAa,EAAE,CAAC;YACpC,IAAI,cAAc,EAAE,CAAC;gBACnB,IAAI,yBAAyB,CAAC,cAAc,CAAC,EAAE,CAAC;oBAC9C,UAAU,CAAC,IAAI,CAAC;wBACd,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC;wBAC3D,QAAQ,EAAE,cAAc,CAAC,QAAQ;wBACjC,SAAS,EAAE,cAAc,CAAC,UAAU;wBACpC,QAAQ;wBACR,GAAG,EAAE,eAAe,CAAC,cAAc,CAAC,QAAQ,CAAC;wBAC7C,IAAI,EAAE,OAAO;wBACb,UAAU;wBACV,UAAU;wBACV,UAAU,EAAE,cAAc;wBAC1B,kBAAkB,EAAE,uCAAuC,CAAC;4BAC1D,IAAI,EAAE,OAAO;4BACb,MAAM,EAAE,gBAAgB;4BACxB,QAAQ,EAAE,cAAc,CAAC,QAAQ;4BACjC,SAAS,EAAE,cAAc,CAAC,UAAU;4BACpC,kBAAkB,EAAE,cAAc,EAAE,kBAAkB;4BACtD,QAAQ,EAAE,cAAc,CAAC,QAAQ;yBAClC,CAAC;qBACH,CAAC,CAAC;oBACH,SAAS;gBACX,CAAC;gBAED,cAAc,CAAC,IAAI,CACjB,mCAAmC,UAAU,4BAA4B,cAAc,CAAC,QAAQ,mDAAmD,CACpJ,CAAC;YACJ,CAAC;YAED,MAAM,YAAY,GAAG,MAAM,eAAe,CAAC,MAAM,EAAE,UAAU,CAAC,CAAC;YAC/D,IAAI,YAAY,EAAE,CAAC;gBACjB,UAAU,CAAC,IAAI,CAAC;oBACd,MAAM,EAAE,MAAM,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC;oBACzD,QAAQ,EAAE,YAAY,CAAC,QAAQ;oBAC/B,SAAS,EAAE,YAAY,CAAC,UAAU;oBAClC,QAAQ;oBACR,GAAG,EAAE,eAAe,CAAC,YAAY,CAAC,QAAQ,CAAC;oBAC3C,IAAI,EAAE,OAAO;oBACb,UAAU;oBACV,UAAU;oBACV,UAAU,EAAE,cAAc;oBAC1B,kBAAkB,EAAE,uCAAuC,CAAC;wBAC1D,IAAI,EAAE,OAAO;wBACb,MAAM,EAAE,kBAAkB;wBAC1B,QAAQ,EAAE,YAAY,CAAC,QAAQ;wBAC/B,SAAS,EAAE,YAAY,CAAC,UAAU;wBAClC,kBAAkB,EAAE,cAAc,EAAE,kBAAkB;wBACtD,QAAQ,EAAE,CAAC,GAAG,cAAc,EAAE,GAAG,YAAY,CAAC,QAAQ,CAAC;qBACxD,CAAC;iBACH,CAAC,CAAC;gBACH,SAAS;YACX,CAAC;QACH,CAAC;QAED,MAAM,OAAO,GAAG,MAAM,WAAW,CAAC,MAAM,EAAE,CAAC;QAC3C,MAAM,CAAC,UAAU,CAAC,GAAG,MAAM,OAAO,CAAC,SAAS,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC;QAC9D,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC;QAC5B,MAAM,SAAS,GAAG,MAAM,OAAO,CAAC,IAAI,EAAE,CAAC;QACvC,MAAM,UAAU,GAAG,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAE1C,UAAU,CAAC,IAAI,CAAC;YACd,MAAM,EAAE,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC;YACrC,QAAQ,EAAE,iBAAiB;YAC3B,SAAS,EAAE,UAAU,CAAC,MAAM;YAC5B,QAAQ;YACR,GAAG,EAAE,KAAK;YACV,IAAI,EAAE,KAAK;YACX,UAAU;YACV,UAAU;YACV,UAAU,EAAE,UAAU;YACtB,kBAAkB,EAAE,uCAAuC,CAAC;gBAC1D,IAAI,EAAE,KAAK;gBACX,MAAM,EAAE,iBAAiB;gBACzB,QAAQ,EAAE,iBAAiB;gBAC3B,SAAS,EAAE,UAAU,CAAC,MAAM;gBAC5B,kBAAkB,EAAE,cAAc,EAAE,kBAAkB;gBACtD,QAAQ,EACN,uBAAuB,IAAI,cAAc,IAAI,CAAC,yBAAyB,CAAC,cAAc,CAAC;oBACrF,CAAC,CAAC,CAAC,mCAAmC,UAAU,yGAAyG,CAAC;oBAC1J,CAAC,CAAC,EAAE;aACT,CAAC;SACH,CAAC,CAAC;IACL,CAAC;IAED,OAAO,UAAU,CAAC;AACpB,CAAC"}
@@ -0,0 +1,118 @@
1
+ import type { LLMConfig } from '../llm/types.js';
2
+ import { extractGeotechDocumentFactsFromText, interpretGeotechDocumentPage, type GeotechDocumentClassification, type GeotechDocumentInsight, type GeotechMaterialObservation, type GeotechParameterObservation, type ParseStatus } from '../vision/geotech-document.js';
3
+ import { transcribeDocumentImageText } from '../vision/index.js';
4
+ import { type DocumentTextHintSource } from '../vision/ocr.js';
5
+ import type { PdfDocumentInspection, PdfPageClassification } from './pdf.js';
6
+ export interface GeotechDocumentVisionInput {
7
+ base64: string;
8
+ mimeType: string;
9
+ fileBytes?: number;
10
+ }
11
+ export interface GeotechDocumentPageInput extends GeotechDocumentVisionInput {
12
+ pageNumber: number;
13
+ totalPages: number;
14
+ sourceKind?: 'pdf-page' | 'raster-image';
15
+ filePath?: string;
16
+ }
17
+ export interface GeotechDocumentSource {
18
+ filePath?: string;
19
+ fileName?: string;
20
+ inputKind: 'image' | 'pdf';
21
+ }
22
+ export interface GeotechDocumentPageAudit {
23
+ pageNumber: number;
24
+ classification: PdfPageClassification | null;
25
+ textHintSource: DocumentTextHintSource;
26
+ parseStatus: ParseStatus;
27
+ confidence: number;
28
+ materialCount: number;
29
+ classificationCount: number;
30
+ parameterCount: number;
31
+ warnings: string[];
32
+ }
33
+ export type GeotechDocumentFindingSeverity = 'advisory' | 'review' | 'blocking';
34
+ export type GeotechDocumentFindingScope = 'document' | 'page' | 'material';
35
+ export interface GeotechDocumentFinding {
36
+ code: string;
37
+ severity: GeotechDocumentFindingSeverity;
38
+ scope: GeotechDocumentFindingScope;
39
+ message: string;
40
+ pageNumber?: number;
41
+ materialDescription?: string;
42
+ }
43
+ export interface GeotechDocumentInspectionSummary {
44
+ pageClassificationCounts: Partial<Record<PdfPageClassification, number>>;
45
+ imageHeavyPageCount: number;
46
+ nativeTextPageCount: number;
47
+ degradedPageCount: number;
48
+ ocrRecoveredPageCount: number;
49
+ }
50
+ export interface GeotechDocumentIngestResult {
51
+ kind: 'geotech-ingest-result';
52
+ schemaVersion: 1;
53
+ documentType: 'geotech-document';
54
+ generatedAt: string;
55
+ source: GeotechDocumentSource & {
56
+ totalPages: number;
57
+ successfulPages: number;
58
+ failedPages: number;
59
+ };
60
+ inspection: PdfDocumentInspection | null;
61
+ inspectionSummary: GeotechDocumentInspectionSummary | null;
62
+ documentClass: string | null;
63
+ title: string | null;
64
+ summary: string | null;
65
+ materials: GeotechMaterialObservation[];
66
+ classifications: GeotechDocumentClassification[];
67
+ parameters: GeotechParameterObservation[];
68
+ risks: string[];
69
+ recommendations: string[];
70
+ contentChunks?: GeotechDocumentContentChunk[];
71
+ pageAudits: GeotechDocumentPageAudit[];
72
+ pageFailures: string[];
73
+ warnings: string[];
74
+ reviewFindings: GeotechDocumentFinding[];
75
+ reviewReasons: string[];
76
+ parseStatus: ParseStatus;
77
+ confidence: number;
78
+ reviewRequired: boolean;
79
+ canAutoProceed: boolean;
80
+ }
81
+ export interface GeotechDocumentContentChunk {
82
+ chunkId: string;
83
+ pageRange: [number, number];
84
+ headingAncestry: string[];
85
+ scope: 'page' | 'table' | 'figure' | 'section';
86
+ sectionType?: GeotechDocumentSectionType;
87
+ significance?: number;
88
+ text: string;
89
+ sourcePages: number[];
90
+ }
91
+ export type GeotechDocumentSectionType = 'administrative' | 'summary' | 'ground-model' | 'laboratory' | 'classification' | 'groundwater' | 'recommendation' | 'visual-appendix' | 'general';
92
+ export interface IngestGeotechDocumentOptions {
93
+ config: LLMConfig;
94
+ source: GeotechDocumentSource;
95
+ inspection?: PdfDocumentInspection | null;
96
+ image?: GeotechDocumentVisionInput;
97
+ pages?: GeotechDocumentPageInput[];
98
+ interpretPage?: typeof interpretGeotechDocumentPage;
99
+ extractTextFacts?: typeof extractGeotechDocumentFactsFromText;
100
+ transcribePageImageText?: typeof transcribeDocumentImageText;
101
+ pageConcurrency?: number;
102
+ now?: () => Date;
103
+ }
104
+ export declare function inferPreflightLowYieldPageRole(input: {
105
+ inspectionPage: PdfDocumentInspection['pages'][number] | undefined;
106
+ previousInspectionPage?: PdfDocumentInspection['pages'][number] | undefined;
107
+ pageNumber?: number | null;
108
+ totalPages?: number | null;
109
+ sourceKind?: GeotechDocumentPageInput['sourceKind'];
110
+ }): GeotechDocumentSectionType | null;
111
+ export declare function buildPreflightLowYieldInsight(input: {
112
+ role: GeotechDocumentSectionType;
113
+ inspectionPage: PdfDocumentInspection['pages'][number];
114
+ pageNumber: number;
115
+ totalPages: number;
116
+ }): GeotechDocumentInsight;
117
+ export declare function ingestGeotechDocument(options: IngestGeotechDocumentOptions): Promise<GeotechDocumentIngestResult>;
118
+ //# sourceMappingURL=geotech-document.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"geotech-document.d.ts","sourceRoot":"","sources":["../../src/ingest/geotech-document.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAEjD,OAAO,EACL,mCAAmC,EACnC,4BAA4B,EAC5B,KAAK,6BAA6B,EAElC,KAAK,sBAAsB,EAC3B,KAAK,0BAA0B,EAC/B,KAAK,2BAA2B,EAChC,KAAK,WAAW,EACjB,MAAM,+BAA+B,CAAC;AACvC,OAAO,EAAE,2BAA2B,EAAE,MAAM,oBAAoB,CAAC;AACjE,OAAO,EAA2B,KAAK,sBAAsB,EAAE,MAAM,kBAAkB,CAAC;AACxF,OAAO,KAAK,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,UAAU,CAAC;AAE7E,MAAM,WAAW,0BAA0B;IACzC,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,wBAAyB,SAAQ,0BAA0B;IAC1E,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,CAAC,EAAE,UAAU,GAAG,cAAc,CAAC;IACzC,QAAQ,CAAC,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,qBAAqB;IACpC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,OAAO,GAAG,KAAK,CAAC;CAC5B;AAED,MAAM,WAAW,wBAAwB;IACvC,UAAU,EAAE,MAAM,CAAC;IACnB,cAAc,EAAE,qBAAqB,GAAG,IAAI,CAAC;IAC7C,cAAc,EAAE,sBAAsB,CAAC;IACvC,WAAW,EAAE,WAAW,CAAC;IACzB,UAAU,EAAE,MAAM,CAAC;IACnB,aAAa,EAAE,MAAM,CAAC;IACtB,mBAAmB,EAAE,MAAM,CAAC;IAC5B,cAAc,EAAE,MAAM,CAAC;IACvB,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AAED,MAAM,MAAM,8BAA8B,GAAG,UAAU,GAAG,QAAQ,GAAG,UAAU,CAAC;AAChF,MAAM,MAAM,2BAA2B,GAAG,UAAU,GAAG,MAAM,GAAG,UAAU,CAAC;AAE3E,MAAM,WAAW,sBAAsB;IACrC,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,8BAA8B,CAAC;IACzC,KAAK,EAAE,2BAA2B,CAAC;IACnC,OAAO,EAAE,MAAM,CAAC;IAChB,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,mBAAmB,CAAC,EAAE,MAAM,CAAC;CAC9B;AAED,MAAM,WAAW,gCAAgC;IAC/C,wBAAwB,EAAE,OAAO,CAAC,MAAM,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC,CAAC;IACzE,mBAAmB,EAAE,MAAM,CAAC;IAC5B,mBAAmB,EAAE,MAAM,CAAC;IAC5B,iBAAiB,EAAE,MAAM,CAAC;IAC1B,qBAAqB,EAAE,MAAM,CAAC;CAC/B;AAED,MAAM,WAAW,2BAA2B;IAC1C,IAAI,EAAE,uBAAuB,CAAC;IAC9B,aAAa,EAAE,CAAC,CAAC;IACjB,YAAY,EAAE,kBAAkB,CAAC;IACjC,WAAW,EAAE,MAAM,CAAC;IACpB,MAAM,EAAE,qBAAqB,GAAG;QAC9B,UAAU,EAAE,MAAM,CAAC;QACnB,eAAe,EAAE,MAAM,CAAC;QACxB,WAAW,EAAE,MAAM,CAAC;KACrB,CAAC;IACF,UAAU,EAAE,qBAAqB,GAAG,IAAI,CAAC;IACzC,iBAAiB,EAAE,gCAAgC,GAAG,IAAI,CAAC;IAC3D,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,aAAa,CAAC,EAAE,2BAA2B,EAAE,CAAC;IAC9C,UAAU,EAAE,wBAAwB,EAAE,CAAC;IACvC,YAAY,EAAE,MAAM,EAAE,CAAC;IACvB,QAAQ,EAAE,MAAM,EAAE,CAAC;IACnB,cAAc,EAAE,sBAAsB,EAAE,CAAC;IACzC,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,WAAW,EAAE,WAAW,CAAC;IACzB,UAAU,EAAE,MAAM,CAAC;IACnB,cAAc,EAAE,OAAO,CAAC;IACxB,cAAc,EAAE,OAAO,CAAC;CACzB;AAED,MAAM,WAAW,2BAA2B;IAC1C,OAAO,EAAE,MAAM,CAAC;IAChB,SAAS,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC5B,eAAe,EAAE,MAAM,EAAE,CAAC;IAC1B,KAAK,EAAE,MAAM,GAAG,OAAO,GAAG,QAAQ,GAAG,SAAS,CAAC;IAC/C,WAAW,CAAC,EAAE,0BAA0B,CAAC;IACzC,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,IAAI,EAAE,MAAM,CAAC;IACb,WAAW,EAAE,MAAM,EAAE,CAAC;CACvB;AAED,MAAM,MAAM,0BAA0B,GAClC,gBAAgB,GAChB,SAAS,GACT,cAAc,GACd,YAAY,GACZ,gBAAgB,GAChB,aAAa,GACb,gBAAgB,GAChB,iBAAiB,GACjB,SAAS,CAAC;AAUd,MAAM,WAAW,4BAA4B;IAC3C,MAAM,EAAE,SAAS,CAAC;IAClB,MAAM,EAAE,qBAAqB,CAAC;IAC9B,UAAU,CAAC,EAAE,qBAAqB,GAAG,IAAI,CAAC;IAC1C,KAAK,CAAC,EAAE,0BAA0B,CAAC;IACnC,KAAK,CAAC,EAAE,wBAAwB,EAAE,CAAC;IACnC,aAAa,CAAC,EAAE,OAAO,4BAA4B,CAAC;IACpD,gBAAgB,CAAC,EAAE,OAAO,mCAAmC,CAAC;IAC9D,uBAAuB,CAAC,EAAE,OAAO,2BAA2B,CAAC;IAC7D,eAAe,CAAC,EAAE,MAAM,CAAC;IACzB,GAAG,CAAC,EAAE,MAAM,IAAI,CAAC;CAClB;AAgxBD,wBAAgB,8BAA8B,CAAC,KAAK,EAAE;IACpD,cAAc,EAAE,qBAAqB,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,GAAG,SAAS,CAAC;IACnE,sBAAsB,CAAC,EAAE,qBAAqB,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,GAAG,SAAS,CAAC;IAC5E,UAAU,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC3B,UAAU,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC3B,UAAU,CAAC,EAAE,wBAAwB,CAAC,YAAY,CAAC,CAAC;CACrD,GAAG,0BAA0B,GAAG,IAAI,CAyDpC;AAED,wBAAgB,6BAA6B,CAAC,KAAK,EAAE;IACnD,IAAI,EAAE,0BAA0B,CAAC;IACjC,cAAc,EAAE,qBAAqB,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC;IACvD,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;CACpB,GAAG,sBAAsB,CA4CzB;AA+GD,wBAAsB,qBAAqB,CACzC,OAAO,EAAE,4BAA4B,GACpC,OAAO,CAAC,2BAA2B,CAAC,CAgStC"}