@cesarandreslopez/occ 0.3.1 → 0.4.1
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.
- package/README.md +130 -4
- package/dist/src/cli-validation.d.ts +14 -0
- package/dist/src/cli-validation.js +58 -0
- package/dist/src/cli-validation.js.map +1 -0
- package/dist/src/cli.js +26 -8
- package/dist/src/cli.js.map +1 -1
- package/dist/src/code/build.d.ts +7 -5
- package/dist/src/code/build.js +68 -11
- package/dist/src/code/build.js.map +1 -1
- package/dist/src/code/command.js +26 -19
- package/dist/src/code/command.js.map +1 -1
- package/dist/src/code/discover.d.ts +6 -4
- package/dist/src/code/discover.js +5 -0
- package/dist/src/code/discover.js.map +1 -1
- package/dist/src/code/languages.d.ts +28 -7
- package/dist/src/code/languages.js +9 -0
- package/dist/src/code/languages.js.map +1 -1
- package/dist/src/code/output.js +12 -3
- package/dist/src/code/output.js.map +1 -1
- package/dist/src/code/parsers.js +75 -10
- package/dist/src/code/parsers.js.map +1 -1
- package/dist/src/code/query.js +82 -22
- package/dist/src/code/query.js.map +1 -1
- package/dist/src/code/types.d.ts +1201 -145
- package/dist/src/code/types.js +150 -1
- package/dist/src/code/types.js.map +1 -1
- package/dist/src/doc/command.d.ts +2 -0
- package/dist/src/doc/command.js +54 -0
- package/dist/src/doc/command.js.map +1 -0
- package/dist/src/doc/inspect-docx.d.ts +14 -0
- package/dist/src/doc/inspect-docx.js +188 -0
- package/dist/src/doc/inspect-docx.js.map +1 -0
- package/dist/src/doc/inspect-odt.d.ts +8 -0
- package/dist/src/doc/inspect-odt.js +92 -0
- package/dist/src/doc/inspect-odt.js.map +1 -0
- package/dist/src/doc/inspect-pdf.d.ts +8 -0
- package/dist/src/doc/inspect-pdf.js +88 -0
- package/dist/src/doc/inspect-pdf.js.map +1 -0
- package/dist/src/doc/inspect.d.ts +2 -0
- package/dist/src/doc/inspect.js +70 -0
- package/dist/src/doc/inspect.js.map +1 -0
- package/dist/src/doc/output.d.ts +3 -0
- package/dist/src/doc/output.js +82 -0
- package/dist/src/doc/output.js.map +1 -0
- package/dist/src/doc/types.d.ts +172 -0
- package/dist/src/doc/types.js +62 -0
- package/dist/src/doc/types.js.map +1 -0
- package/dist/src/inspect/shared.d.ts +20 -0
- package/dist/src/inspect/shared.js +29 -0
- package/dist/src/inspect/shared.js.map +1 -0
- package/dist/src/inspect/xlsx-cells.d.ts +4 -0
- package/dist/src/inspect/xlsx-cells.js +31 -0
- package/dist/src/inspect/xlsx-cells.js.map +1 -0
- package/dist/src/markdown/convert.js +6 -4
- package/dist/src/markdown/convert.js.map +1 -1
- package/dist/src/output/tabular.d.ts +6 -4
- package/dist/src/output/tabular.js +5 -0
- package/dist/src/output/tabular.js.map +1 -1
- package/dist/src/output/tree.d.ts +21 -9
- package/dist/src/output/tree.js +10 -0
- package/dist/src/output/tree.js.map +1 -1
- package/dist/src/parsers/docx.js +5 -1
- package/dist/src/parsers/docx.js.map +1 -1
- package/dist/src/parsers/odf.js +4 -3
- package/dist/src/parsers/odf.js.map +1 -1
- package/dist/src/parsers/pdf.js +6 -1
- package/dist/src/parsers/pdf.js.map +1 -1
- package/dist/src/parsers/pptx.js +2 -1
- package/dist/src/parsers/pptx.js.map +1 -1
- package/dist/src/parsers/xlsx.js +11 -4
- package/dist/src/parsers/xlsx.js.map +1 -1
- package/dist/src/progress.d.ts +7 -5
- package/dist/src/progress.js +6 -0
- package/dist/src/progress.js.map +1 -1
- package/dist/src/scc.d.ts +35 -26
- package/dist/src/scc.js +26 -1
- package/dist/src/scc.js.map +1 -1
- package/dist/src/sheet/command.d.ts +2 -0
- package/dist/src/sheet/command.js +61 -0
- package/dist/src/sheet/command.js.map +1 -0
- package/dist/src/sheet/inspect.d.ts +3 -0
- package/dist/src/sheet/inspect.js +426 -0
- package/dist/src/sheet/inspect.js.map +1 -0
- package/dist/src/sheet/output.d.ts +3 -0
- package/dist/src/sheet/output.js +158 -0
- package/dist/src/sheet/output.js.map +1 -0
- package/dist/src/sheet/types.d.ts +472 -0
- package/dist/src/sheet/types.js +119 -0
- package/dist/src/sheet/types.js.map +1 -0
- package/dist/src/slide/command.d.ts +2 -0
- package/dist/src/slide/command.js +54 -0
- package/dist/src/slide/command.js.map +1 -0
- package/dist/src/slide/inspect-odp.d.ts +7 -0
- package/dist/src/slide/inspect-odp.js +105 -0
- package/dist/src/slide/inspect-odp.js.map +1 -0
- package/dist/src/slide/inspect-pptx.d.ts +11 -0
- package/dist/src/slide/inspect-pptx.js +157 -0
- package/dist/src/slide/inspect-pptx.js.map +1 -0
- package/dist/src/slide/inspect.d.ts +2 -0
- package/dist/src/slide/inspect.js +76 -0
- package/dist/src/slide/inspect.js.map +1 -0
- package/dist/src/slide/output.d.ts +3 -0
- package/dist/src/slide/output.js +92 -0
- package/dist/src/slide/output.js.map +1 -0
- package/dist/src/slide/types.d.ts +167 -0
- package/dist/src/slide/types.js +61 -0
- package/dist/src/slide/types.js.map +1 -0
- package/dist/src/stats.d.ts +6 -4
- package/dist/src/stats.js +5 -0
- package/dist/src/stats.js.map +1 -1
- package/dist/src/structure/types.d.ts +19 -11
- package/dist/src/structure/types.js +26 -21
- package/dist/src/structure/types.js.map +1 -1
- package/dist/src/table/command.d.ts +2 -0
- package/dist/src/table/command.js +58 -0
- package/dist/src/table/command.js.map +1 -0
- package/dist/src/table/inspect-docx.d.ts +2 -0
- package/dist/src/table/inspect-docx.js +108 -0
- package/dist/src/table/inspect-docx.js.map +1 -0
- package/dist/src/table/inspect-odp.d.ts +2 -0
- package/dist/src/table/inspect-odp.js +106 -0
- package/dist/src/table/inspect-odp.js.map +1 -0
- package/dist/src/table/inspect-odt.d.ts +2 -0
- package/dist/src/table/inspect-odt.js +100 -0
- package/dist/src/table/inspect-odt.js.map +1 -0
- package/dist/src/table/inspect-pptx.d.ts +2 -0
- package/dist/src/table/inspect-pptx.js +112 -0
- package/dist/src/table/inspect-pptx.js.map +1 -0
- package/dist/src/table/inspect-xlsx.d.ts +2 -0
- package/dist/src/table/inspect-xlsx.js +161 -0
- package/dist/src/table/inspect-xlsx.js.map +1 -0
- package/dist/src/table/inspect.d.ts +2 -0
- package/dist/src/table/inspect.js +65 -0
- package/dist/src/table/inspect.js.map +1 -0
- package/dist/src/table/output.d.ts +3 -0
- package/dist/src/table/output.js +93 -0
- package/dist/src/table/output.js.map +1 -0
- package/dist/src/table/types.d.ts +104 -0
- package/dist/src/table/types.js +40 -0
- package/dist/src/table/types.js.map +1 -0
- package/dist/src/types.d.ts +25 -20
- package/dist/src/types.js +21 -1
- package/dist/src/types.js.map +1 -1
- package/dist/src/utils.d.ts +11 -1
- package/dist/src/utils.js +2 -0
- package/dist/src/utils.js.map +1 -1
- package/dist/src/walker.d.ts +21 -12
- package/dist/src/walker.js +13 -0
- package/dist/src/walker.js.map +1 -1
- package/package.json +3 -2
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"inspect-pdf.js","sourceRoot":"","sources":["../../../src/doc/inspect-pdf.ts"],"names":[],"mappings":"AACA,OAAO,GAAG,MAAM,WAAW,CAAC;AAC5B,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,gBAAgB,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AAIxE,MAAM,oBAAoB,GAAG,CAAC,CAAC,MAAM,CAAC;IACpC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE;IAChB,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE;IACpB,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC;CACpD,CAAC,CAAC;AAEH,SAAS,YAAY,CAAC,KAAc;IAClC,IAAI,CAAC,KAAK;QAAE,OAAO,SAAS,CAAC;IAC7B,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;IAC1B,mDAAmD;IACnD,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,+CAA+C,CAAC,CAAC;IACzE,IAAI,KAAK,EAAE,CAAC;QACV,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC;IACpF,CAAC;IACD,OAAO,cAAc,CAAC,GAAG,CAAC,CAAC;AAC7B,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,UAAU,CAC9B,QAAgB,EAChB,MAAc,EACd,gBAAwB;IAOxB,2BAA2B;IAC3B,MAAM,WAAW,GAAG,OAAO,CAAC,GAAG,CAAC;IAChC,OAAO,CAAC,GAAG,GAAG,CAAC,GAAG,IAAe,EAAE,EAAE;QACnC,IAAI,OAAO,IAAI,CAAC,CAAC,CAAC,KAAK,QAAQ,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,WAAW,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,QAAQ,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,wBAAwB,CAAC,CAAC,EAAE,CAAC;YACrJ,OAAO;QACT,CAAC;QACD,WAAW,CAAC,KAAK,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;IACnC,CAAC,CAAC;IAEF,IAAI,IAA0C,CAAC;IAC/C,IAAI,CAAC;QACH,IAAI,GAAG,oBAAoB,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC;IACvD,CAAC;YAAS,CAAC;QACT,OAAO,CAAC,GAAG,GAAG,WAAW,CAAC;IAC5B,CAAC;IAED,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;IAEvB,MAAM,UAAU,GAAuB;QACrC,KAAK,EAAE,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC;QACnC,OAAO,EAAE,gBAAgB,CAAC,IAAI,CAAC,OAAO,CAAC;QACvC,MAAM,EAAE,gBAAgB,CAAC,IAAI,CAAC,MAAM,CAAC;QACrC,QAAQ,EAAE,gBAAgB,CAAC,IAAI,CAAC,QAAQ,CAAC;QACzC,WAAW,EAAE,YAAY,CAAC,IAAI,CAAC,YAAY,CAAC;QAC5C,YAAY,EAAE,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC;KACzC,CAAC;IAEF,wCAAwC;IACxC,MAAM,SAAS,GAAiB;QAC9B,QAAQ,EAAE,KAAK;QACf,cAAc,EAAE,KAAK;QACrB,UAAU,EAAE,KAAK;QACjB,eAAe,EAAE,KAAK;QACtB,SAAS,EAAE,KAAK;QAChB,QAAQ,EAAE,KAAK;QACf,MAAM,EAAE,KAAK;QACb,YAAY,EAAE,KAAK;QACnB,MAAM,EAAE,KAAK;QACb,SAAS,EAAE,CAAC,CAAC,IAAI,CAAC,iBAAiB,IAAK,IAAgC,CAAC,SAAS,KAAK,IAAI;KAC5F,CAAC;IAEF,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,IAAI,EAAE,CAAC;IAC7B,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC;IAC/B,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC;IAE/E,MAAM,YAAY,GAAoB;QACpC,KAAK;QACL,KAAK,EAAE,IAAI,CAAC,QAAQ;QACpB,cAAc,EAAE,KAAK;QACrB,UAAU;QACV,UAAU,EAAE,IAAI,CAAC,MAAM;QACvB,MAAM,EAAE,CAAC;QACT,MAAM,EAAE,CAAC;KACV,CAAC;IAEF,0BAA0B;IAC1B,MAAM,cAAc,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC5E,MAAM,iBAAiB,GAAiC,EAAE,CAAC;IAC3D,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,gBAAgB,EAAE,cAAc,CAAC,MAAM,CAAC,CAAC;IAChE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/B,iBAAiB,CAAC,IAAI,CAAC;YACrB,KAAK,EAAE,CAAC;YACR,IAAI,EAAE,cAAc,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;YACnD,SAAS,EAAE,KAAK;SACjB,CAAC,CAAC;IACL,CAAC;IAED,MAAM,OAAO,GAAmB;QAC9B,SAAS,EAAE,cAAc,CAAC,MAAM,GAAG,gBAAgB;QACnD,UAAU,EAAE,iBAAiB;KAC9B,CAAC;IAEF,OAAO,EAAE,UAAU,EAAE,SAAS,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;AAC1D,CAAC"}
|
|
@@ -0,0 +1,70 @@
|
|
|
1
|
+
import path from 'node:path';
|
|
2
|
+
import { readFile } from 'node:fs/promises';
|
|
3
|
+
import { getExtension } from '../utils.js';
|
|
4
|
+
import { estimateTokens } from '../inspect/shared.js';
|
|
5
|
+
import { documentToMarkdown } from '../markdown/convert.js';
|
|
6
|
+
import { extractFromMarkdown } from '../structure/index.js';
|
|
7
|
+
import { inspectDocx } from './inspect-docx.js';
|
|
8
|
+
import { inspectPdf } from './inspect-pdf.js';
|
|
9
|
+
import { inspectOdt } from './inspect-odt.js';
|
|
10
|
+
const SUPPORTED_EXTENSIONS = new Set(['docx', 'pdf', 'odt']);
|
|
11
|
+
export async function inspectDocument(filePath, options) {
|
|
12
|
+
const resolvedPath = path.resolve(filePath);
|
|
13
|
+
const ext = getExtension(resolvedPath);
|
|
14
|
+
if (!SUPPORTED_EXTENSIONS.has(ext)) {
|
|
15
|
+
throw new Error(`Unsupported document format: .${ext || '(none)'} (expected .docx, .pdf, or .odt)`);
|
|
16
|
+
}
|
|
17
|
+
const buffer = await readFile(resolvedPath);
|
|
18
|
+
const format = ext;
|
|
19
|
+
let result;
|
|
20
|
+
switch (format) {
|
|
21
|
+
case 'docx':
|
|
22
|
+
result = await inspectDocx(resolvedPath, buffer, options.sampleParagraphs);
|
|
23
|
+
break;
|
|
24
|
+
case 'pdf':
|
|
25
|
+
result = await inspectPdf(resolvedPath, buffer, options.sampleParagraphs);
|
|
26
|
+
break;
|
|
27
|
+
case 'odt':
|
|
28
|
+
result = await inspectOdt(resolvedPath, buffer, options.sampleParagraphs);
|
|
29
|
+
break;
|
|
30
|
+
}
|
|
31
|
+
// Structure extraction
|
|
32
|
+
let structure = null;
|
|
33
|
+
if (options.includeStructure) {
|
|
34
|
+
try {
|
|
35
|
+
const markdown = await documentToMarkdown(resolvedPath);
|
|
36
|
+
if (markdown) {
|
|
37
|
+
const docStructure = extractFromMarkdown(markdown);
|
|
38
|
+
if (docStructure.totalNodes > 0) {
|
|
39
|
+
const topLevelSections = docStructure.rootNodes.map(n => n.title);
|
|
40
|
+
structure = {
|
|
41
|
+
headingCount: docStructure.totalNodes,
|
|
42
|
+
maxDepth: docStructure.maxDepth,
|
|
43
|
+
topLevelSections,
|
|
44
|
+
tree: docStructure.rootNodes,
|
|
45
|
+
};
|
|
46
|
+
}
|
|
47
|
+
}
|
|
48
|
+
}
|
|
49
|
+
catch { /* structure extraction is best-effort */ }
|
|
50
|
+
}
|
|
51
|
+
// Token estimates
|
|
52
|
+
const fullText = [
|
|
53
|
+
result.preview.paragraphs.map(p => p.text).join(' '),
|
|
54
|
+
].join(' ');
|
|
55
|
+
const fullTokenEstimate = estimateTokens(result.contentStats.characters);
|
|
56
|
+
const previewTokenEstimate = estimateTokens(result.preview.paragraphs.reduce((sum, p) => sum + p.text.length, 0));
|
|
57
|
+
return {
|
|
58
|
+
file: resolvedPath,
|
|
59
|
+
format,
|
|
60
|
+
size: buffer.length,
|
|
61
|
+
properties: result.properties,
|
|
62
|
+
riskFlags: result.riskFlags,
|
|
63
|
+
contentStats: result.contentStats,
|
|
64
|
+
structure,
|
|
65
|
+
contentPreview: result.preview,
|
|
66
|
+
fullTokenEstimate,
|
|
67
|
+
previewTokenEstimate,
|
|
68
|
+
};
|
|
69
|
+
}
|
|
70
|
+
//# sourceMappingURL=inspect.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"inspect.js","sourceRoot":"","sources":["../../../src/doc/inspect.ts"],"names":[],"mappings":"AAAA,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAC3C,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC;AACtD,OAAO,EAAE,kBAAkB,EAAE,MAAM,wBAAwB,CAAC;AAC5D,OAAO,EAAE,mBAAmB,EAAE,MAAM,uBAAuB,CAAC;AAC5D,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAG9C,MAAM,oBAAoB,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC;AAE7D,MAAM,CAAC,KAAK,UAAU,eAAe,CAAC,QAAgB,EAAE,OAA0B;IAChF,MAAM,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;IAC5C,MAAM,GAAG,GAAG,YAAY,CAAC,YAAY,CAAC,CAAC;IACvC,IAAI,CAAC,oBAAoB,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;QACnC,MAAM,IAAI,KAAK,CAAC,iCAAiC,GAAG,IAAI,QAAQ,kCAAkC,CAAC,CAAC;IACtG,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,YAAY,CAAC,CAAC;IAC5C,MAAM,MAAM,GAAG,GAA6B,CAAC;IAE7C,IAAI,MAA+C,CAAC;IACpD,QAAQ,MAAM,EAAE,CAAC;QACf,KAAK,MAAM;YACT,MAAM,GAAG,MAAM,WAAW,CAAC,YAAY,EAAE,MAAM,EAAE,OAAO,CAAC,gBAAgB,CAAC,CAAC;YAC3E,MAAM;QACR,KAAK,KAAK;YACR,MAAM,GAAG,MAAM,UAAU,CAAC,YAAY,EAAE,MAAM,EAAE,OAAO,CAAC,gBAAgB,CAAC,CAAC;YAC1E,MAAM;QACR,KAAK,KAAK;YACR,MAAM,GAAG,MAAM,UAAU,CAAC,YAAY,EAAE,MAAM,EAAE,OAAO,CAAC,gBAAgB,CAAC,CAAC;YAC1E,MAAM;IACV,CAAC;IAED,uBAAuB;IACvB,IAAI,SAAS,GAA4B,IAAI,CAAC;IAC9C,IAAI,OAAO,CAAC,gBAAgB,EAAE,CAAC;QAC7B,IAAI,CAAC;YACH,MAAM,QAAQ,GAAG,MAAM,kBAAkB,CAAC,YAAY,CAAC,CAAC;YACxD,IAAI,QAAQ,EAAE,CAAC;gBACb,MAAM,YAAY,GAAG,mBAAmB,CAAC,QAAQ,CAAC,CAAC;gBACnD,IAAI,YAAY,CAAC,UAAU,GAAG,CAAC,EAAE,CAAC;oBAChC,MAAM,gBAAgB,GAAG,YAAY,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;oBAClE,SAAS,GAAG;wBACV,YAAY,EAAE,YAAY,CAAC,UAAU;wBACrC,QAAQ,EAAE,YAAY,CAAC,QAAQ;wBAC/B,gBAAgB;wBAChB,IAAI,EAAE,YAAY,CAAC,SAAS;qBAC7B,CAAC;gBACJ,CAAC;YACH,CAAC;QACH,CAAC;QAAC,MAAM,CAAC,CAAC,yCAAyC,CAAC,CAAC;IACvD,CAAC;IAED,kBAAkB;IAClB,MAAM,QAAQ,GAAG;QACf,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC;KACrD,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACZ,MAAM,iBAAiB,GAAG,cAAc,CAAC,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,CAAC;IACzE,MAAM,oBAAoB,GAAG,cAAc,CACzC,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,CACrE,CAAC;IAEF,OAAO;QACL,IAAI,EAAE,YAAY;QAClB,MAAM;QACN,IAAI,EAAE,MAAM,CAAC,MAAM;QACnB,UAAU,EAAE,MAAM,CAAC,UAAU;QAC7B,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,YAAY,EAAE,MAAM,CAAC,YAAY;QACjC,SAAS;QACT,cAAc,EAAE,MAAM,CAAC,OAAO;QAC9B,iBAAiB;QACjB,oBAAoB;KACrB,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,82 @@
|
|
|
1
|
+
import chalk from 'chalk';
|
|
2
|
+
import { formatBytes, formatNumber } from '../utils.js';
|
|
3
|
+
import { sectionHeader, stripAnsi } from '../output/tabular.js';
|
|
4
|
+
const colors = {
|
|
5
|
+
header: (value) => chalk.bold(value),
|
|
6
|
+
key: (value) => chalk.cyan(value),
|
|
7
|
+
value: (value) => value,
|
|
8
|
+
dim: (value) => chalk.dim(value),
|
|
9
|
+
ok: (value) => chalk.green(value),
|
|
10
|
+
warn: (value) => chalk.yellow(value),
|
|
11
|
+
};
|
|
12
|
+
const noColor = Object.fromEntries(Object.keys(colors).map(key => [key, (value) => value]));
|
|
13
|
+
function palette(ci = false) {
|
|
14
|
+
return ci ? noColor : colors;
|
|
15
|
+
}
|
|
16
|
+
export function formatDocPayloadJson(payload) {
|
|
17
|
+
return JSON.stringify(payload, null, 2);
|
|
18
|
+
}
|
|
19
|
+
export function formatDocInspection(result, ci = false) {
|
|
20
|
+
const c = palette(ci);
|
|
21
|
+
const sections = [];
|
|
22
|
+
// Overview
|
|
23
|
+
const overviewLines = [
|
|
24
|
+
`${c.key('File')}: ${c.value(result.file)}`,
|
|
25
|
+
`${c.key('Format')}: ${c.value(result.format.toUpperCase())}`,
|
|
26
|
+
`${c.key('Size')}: ${c.value(formatBytes(result.size))}`,
|
|
27
|
+
];
|
|
28
|
+
const propEntries = Object.entries(result.properties).filter(([, value]) => value);
|
|
29
|
+
if (propEntries.length > 0) {
|
|
30
|
+
overviewLines.push(`${c.key('Properties')}: ${propEntries.map(([key, value]) => `${key}=${value}`).join(' | ')}`);
|
|
31
|
+
}
|
|
32
|
+
const riskFlags = Object.entries(result.riskFlags).filter(([, enabled]) => enabled).map(([key]) => key);
|
|
33
|
+
overviewLines.push(`${c.key('Risk Flags')}: ${riskFlags.length > 0 ? c.warn(riskFlags.join(', ')) : c.ok('none')}`);
|
|
34
|
+
sections.push(overviewLines.join('\n'));
|
|
35
|
+
// Content Stats
|
|
36
|
+
const pagesLabel = result.contentStats.pagesEstimated ? ' (estimated)' : '';
|
|
37
|
+
const statsLines = [
|
|
38
|
+
` ${c.key('Words')}: ${formatNumber(result.contentStats.words)} ${c.key('Pages')}: ${formatNumber(result.contentStats.pages)}${pagesLabel}`,
|
|
39
|
+
` ${c.key('Paragraphs')}: ${formatNumber(result.contentStats.paragraphs)} ${c.key('Characters')}: ${formatNumber(result.contentStats.characters)}`,
|
|
40
|
+
` ${c.key('Tables')}: ${formatNumber(result.contentStats.tables)} ${c.key('Images')}: ${formatNumber(result.contentStats.images)}`,
|
|
41
|
+
` ${c.key('Token Estimate')}: preview=${formatNumber(result.previewTokenEstimate)} | full=${formatNumber(result.fullTokenEstimate)}`,
|
|
42
|
+
];
|
|
43
|
+
const statsWidth = Math.max(...statsLines.map(line => stripAnsi(line).length), 56);
|
|
44
|
+
sections.push(`${c.header(sectionHeader('Content Stats', statsWidth, ci))}\n${statsLines.join('\n')}`);
|
|
45
|
+
// Structure
|
|
46
|
+
if (result.structure) {
|
|
47
|
+
const structLines = [];
|
|
48
|
+
for (const node of result.structure.tree) {
|
|
49
|
+
formatTreeNode(node, structLines, '');
|
|
50
|
+
}
|
|
51
|
+
const structTitle = `Structure (${result.structure.headingCount} headings, depth ${result.structure.maxDepth})`;
|
|
52
|
+
const structWidth = Math.max(...structLines.map(line => line.length), structTitle.length + 6, 56);
|
|
53
|
+
sections.push(`${c.header(sectionHeader(structTitle, structWidth, ci))}\n${structLines.join('\n')}`);
|
|
54
|
+
}
|
|
55
|
+
// Content Preview
|
|
56
|
+
if (result.contentPreview.paragraphs.length > 0) {
|
|
57
|
+
const previewLines = [];
|
|
58
|
+
for (const p of result.contentPreview.paragraphs) {
|
|
59
|
+
if (p.isHeading) {
|
|
60
|
+
previewLines.push(` ${c.key(`[H${p.headingLevel}]`)} ${p.text}`);
|
|
61
|
+
}
|
|
62
|
+
else {
|
|
63
|
+
previewLines.push(` ${p.text.length > 120 ? p.text.slice(0, 120) + '...' : p.text}`);
|
|
64
|
+
}
|
|
65
|
+
}
|
|
66
|
+
if (result.contentPreview.truncated) {
|
|
67
|
+
const remainingWords = result.contentStats.words - result.contentPreview.paragraphs.reduce((sum, p) => sum + p.text.split(/\s+/).length, 0);
|
|
68
|
+
const remainingParas = result.contentStats.paragraphs - result.contentPreview.paragraphs.length;
|
|
69
|
+
previewLines.push(c.dim(` (${remainingParas} more paragraphs, ~${formatNumber(remainingWords)} remaining words)`));
|
|
70
|
+
}
|
|
71
|
+
const previewWidth = Math.max(...previewLines.map(line => stripAnsi(line).length), 56);
|
|
72
|
+
sections.push(`${c.header(sectionHeader('Content Preview', previewWidth, ci))}\n${previewLines.join('\n')}`);
|
|
73
|
+
}
|
|
74
|
+
return sections.join('\n\n') + '\n';
|
|
75
|
+
}
|
|
76
|
+
function formatTreeNode(node, lines, indent) {
|
|
77
|
+
lines.push(` ${indent}${node.structureCode ?? ''} ${node.title}`);
|
|
78
|
+
for (const child of node.children) {
|
|
79
|
+
formatTreeNode(child, lines, indent + ' ');
|
|
80
|
+
}
|
|
81
|
+
}
|
|
82
|
+
//# sourceMappingURL=output.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"output.js","sourceRoot":"","sources":["../../../src/doc/output.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AACxD,OAAO,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,sBAAsB,CAAC;AAchE,MAAM,MAAM,GAAgB;IAC1B,MAAM,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;IACpC,GAAG,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC;IACjC,KAAK,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK;IACvB,GAAG,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC;IAChC,EAAE,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC;IACjC,IAAI,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC;CACrC,CAAC;AAEF,MAAM,OAAO,GAAgB,MAAM,CAAC,WAAW,CAC7C,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC,KAAa,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CACtC,CAAC;AAE5B,SAAS,OAAO,CAAC,EAAE,GAAG,KAAK;IACzB,OAAO,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC;AAC/B,CAAC;AAED,MAAM,UAAU,oBAAoB,CAAC,OAA0B;IAC7D,OAAO,IAAI,CAAC,SAAS,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;AAC1C,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,MAA0B,EAAE,EAAE,GAAG,KAAK;IACxE,MAAM,CAAC,GAAG,OAAO,CAAC,EAAE,CAAC,CAAC;IACtB,MAAM,QAAQ,GAAa,EAAE,CAAC;IAE9B,WAAW;IACX,MAAM,aAAa,GAAG;QACpB,GAAG,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE;QAC3C,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC,EAAE;QAC7D,GAAG,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE;KACzD,CAAC;IAEF,MAAM,WAAW,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC;IACnF,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC3B,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,KAAK,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,IAAI,KAAK,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACpH,CAAC;IAED,MAAM,SAAS,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC;IACxG,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,KAAK,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;IAEpH,QAAQ,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IAExC,gBAAgB;IAChB,MAAM,UAAU,GAAG,MAAM,CAAC,YAAY,CAAC,cAAc,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,EAAE,CAAC;IAC5E,MAAM,UAAU,GAAG;QACjB,KAAK,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,YAAY,CAAC,MAAM,CAAC,YAAY,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,YAAY,CAAC,MAAM,CAAC,YAAY,CAAC,KAAK,CAAC,GAAG,UAAU,EAAE;QACnJ,KAAK,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,KAAK,YAAY,CAAC,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,KAAK,YAAY,CAAC,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,EAAE;QACvJ,KAAK,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,YAAY,CAAC,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,YAAY,CAAC,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,EAAE;QAC7I,KAAK,CAAC,CAAC,GAAG,CAAC,gBAAgB,CAAC,aAAa,YAAY,CAAC,MAAM,CAAC,oBAAoB,CAAC,WAAW,YAAY,CAAC,MAAM,CAAC,iBAAiB,CAAC,EAAE;KACtI,CAAC;IAEF,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,EAAE,EAAE,CAAC,CAAC;IACnF,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,eAAe,EAAE,UAAU,EAAE,EAAE,CAAC,CAAC,KAAK,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAEvG,YAAY;IACZ,IAAI,MAAM,CAAC,SAAS,EAAE,CAAC;QACrB,MAAM,WAAW,GAAa,EAAE,CAAC;QACjC,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;YACzC,cAAc,CAAC,IAAI,EAAE,WAAW,EAAE,EAAE,CAAC,CAAC;QACxC,CAAC;QACD,MAAM,WAAW,GAAG,cAAc,MAAM,CAAC,SAAS,CAAC,YAAY,oBAAoB,MAAM,CAAC,SAAS,CAAC,QAAQ,GAAG,CAAC;QAChH,MAAM,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC;QAClG,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,WAAW,EAAE,WAAW,EAAE,EAAE,CAAC,CAAC,KAAK,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACvG,CAAC;IAED,kBAAkB;IAClB,IAAI,MAAM,CAAC,cAAc,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAChD,MAAM,YAAY,GAAa,EAAE,CAAC;QAClC,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,cAAc,CAAC,UAAU,EAAE,CAAC;YACjD,IAAI,CAAC,CAAC,SAAS,EAAE,CAAC;gBAChB,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,YAAY,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YACpE,CAAC;iBAAM,CAAC;gBACN,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YACxF,CAAC;QACH,CAAC;QACD,IAAI,MAAM,CAAC,cAAc,CAAC,SAAS,EAAE,CAAC;YACpC,MAAM,cAAc,GAAG,MAAM,CAAC,YAAY,CAAC,KAAK,GAAG,MAAM,CAAC,cAAc,CAAC,UAAU,CAAC,MAAM,CACxF,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,EAAE,CAAC,CAChD,CAAC;YACF,MAAM,cAAc,GAAG,MAAM,CAAC,YAAY,CAAC,UAAU,GAAG,MAAM,CAAC,cAAc,CAAC,UAAU,CAAC,MAAM,CAAC;YAChG,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,cAAc,sBAAsB,YAAY,CAAC,cAAc,CAAC,mBAAmB,CAAC,CAAC,CAAC;QACtH,CAAC;QACD,MAAM,YAAY,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,YAAY,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,EAAE,EAAE,CAAC,CAAC;QACvF,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,iBAAiB,EAAE,YAAY,EAAE,EAAE,CAAC,CAAC,KAAK,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAC/G,CAAC;IAED,OAAO,QAAQ,CAAC,IAAI,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;AACtC,CAAC;AAED,SAAS,cAAc,CAAC,IAAwE,EAAE,KAAe,EAAE,MAAc;IAC/H,KAAK,CAAC,IAAI,CAAC,KAAK,MAAM,GAAG,IAAI,CAAC,aAAa,IAAI,EAAE,MAAM,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC;IACrE,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;QAClC,cAAc,CAAC,KAAK,EAAE,KAAK,EAAE,MAAM,GAAG,IAAI,CAAC,CAAC;IAC9C,CAAC;AACH,CAAC"}
|
|
@@ -0,0 +1,172 @@
|
|
|
1
|
+
import { z } from 'zod';
|
|
2
|
+
export declare const DocRiskFlagsSchema: z.ZodObject<{
|
|
3
|
+
comments: z.ZodBoolean;
|
|
4
|
+
trackedChanges: z.ZodBoolean;
|
|
5
|
+
hyperlinks: z.ZodBoolean;
|
|
6
|
+
embeddedObjects: z.ZodBoolean;
|
|
7
|
+
footnotes: z.ZodBoolean;
|
|
8
|
+
endnotes: z.ZodBoolean;
|
|
9
|
+
macros: z.ZodBoolean;
|
|
10
|
+
headerFooter: z.ZodBoolean;
|
|
11
|
+
tables: z.ZodBoolean;
|
|
12
|
+
encrypted: z.ZodBoolean;
|
|
13
|
+
}, z.core.$strip>;
|
|
14
|
+
export type DocRiskFlags = z.infer<typeof DocRiskFlagsSchema>;
|
|
15
|
+
export declare const DocContentStatsSchema: z.ZodObject<{
|
|
16
|
+
words: z.ZodNumber;
|
|
17
|
+
pages: z.ZodNumber;
|
|
18
|
+
pagesEstimated: z.ZodBoolean;
|
|
19
|
+
paragraphs: z.ZodNumber;
|
|
20
|
+
characters: z.ZodNumber;
|
|
21
|
+
tables: z.ZodNumber;
|
|
22
|
+
images: z.ZodNumber;
|
|
23
|
+
}, z.core.$strip>;
|
|
24
|
+
export type DocContentStats = z.infer<typeof DocContentStatsSchema>;
|
|
25
|
+
export declare const StructureSummarySchema: z.ZodObject<{
|
|
26
|
+
headingCount: z.ZodNumber;
|
|
27
|
+
maxDepth: z.ZodNumber;
|
|
28
|
+
topLevelSections: z.ZodArray<z.ZodString>;
|
|
29
|
+
tree: z.ZodArray<z.ZodType<import("../structure/types.js").StructureNode, unknown, z.core.$ZodTypeInternals<import("../structure/types.js").StructureNode, unknown>>>;
|
|
30
|
+
}, z.core.$strip>;
|
|
31
|
+
export type StructureSummary = z.infer<typeof StructureSummarySchema>;
|
|
32
|
+
export declare const ContentPreviewParagraphSchema: z.ZodObject<{
|
|
33
|
+
index: z.ZodNumber;
|
|
34
|
+
text: z.ZodString;
|
|
35
|
+
isHeading: z.ZodBoolean;
|
|
36
|
+
headingLevel: z.ZodOptional<z.ZodNumber>;
|
|
37
|
+
}, z.core.$strip>;
|
|
38
|
+
export declare const ContentPreviewSchema: z.ZodObject<{
|
|
39
|
+
truncated: z.ZodBoolean;
|
|
40
|
+
paragraphs: z.ZodArray<z.ZodObject<{
|
|
41
|
+
index: z.ZodNumber;
|
|
42
|
+
text: z.ZodString;
|
|
43
|
+
isHeading: z.ZodBoolean;
|
|
44
|
+
headingLevel: z.ZodOptional<z.ZodNumber>;
|
|
45
|
+
}, z.core.$strip>>;
|
|
46
|
+
}, z.core.$strip>;
|
|
47
|
+
export type ContentPreview = z.infer<typeof ContentPreviewSchema>;
|
|
48
|
+
export declare const DocumentInspectionSchema: z.ZodObject<{
|
|
49
|
+
file: z.ZodString;
|
|
50
|
+
format: z.ZodEnum<{
|
|
51
|
+
docx: "docx";
|
|
52
|
+
pdf: "pdf";
|
|
53
|
+
odt: "odt";
|
|
54
|
+
}>;
|
|
55
|
+
size: z.ZodNumber;
|
|
56
|
+
properties: z.ZodObject<{
|
|
57
|
+
title: z.ZodOptional<z.ZodString>;
|
|
58
|
+
subject: z.ZodOptional<z.ZodString>;
|
|
59
|
+
author: z.ZodOptional<z.ZodString>;
|
|
60
|
+
company: z.ZodOptional<z.ZodString>;
|
|
61
|
+
createdDate: z.ZodOptional<z.ZodString>;
|
|
62
|
+
modifiedDate: z.ZodOptional<z.ZodString>;
|
|
63
|
+
lastModifiedBy: z.ZodOptional<z.ZodString>;
|
|
64
|
+
keywords: z.ZodOptional<z.ZodString>;
|
|
65
|
+
}, z.core.$strip>;
|
|
66
|
+
riskFlags: z.ZodObject<{
|
|
67
|
+
comments: z.ZodBoolean;
|
|
68
|
+
trackedChanges: z.ZodBoolean;
|
|
69
|
+
hyperlinks: z.ZodBoolean;
|
|
70
|
+
embeddedObjects: z.ZodBoolean;
|
|
71
|
+
footnotes: z.ZodBoolean;
|
|
72
|
+
endnotes: z.ZodBoolean;
|
|
73
|
+
macros: z.ZodBoolean;
|
|
74
|
+
headerFooter: z.ZodBoolean;
|
|
75
|
+
tables: z.ZodBoolean;
|
|
76
|
+
encrypted: z.ZodBoolean;
|
|
77
|
+
}, z.core.$strip>;
|
|
78
|
+
contentStats: z.ZodObject<{
|
|
79
|
+
words: z.ZodNumber;
|
|
80
|
+
pages: z.ZodNumber;
|
|
81
|
+
pagesEstimated: z.ZodBoolean;
|
|
82
|
+
paragraphs: z.ZodNumber;
|
|
83
|
+
characters: z.ZodNumber;
|
|
84
|
+
tables: z.ZodNumber;
|
|
85
|
+
images: z.ZodNumber;
|
|
86
|
+
}, z.core.$strip>;
|
|
87
|
+
structure: z.ZodNullable<z.ZodObject<{
|
|
88
|
+
headingCount: z.ZodNumber;
|
|
89
|
+
maxDepth: z.ZodNumber;
|
|
90
|
+
topLevelSections: z.ZodArray<z.ZodString>;
|
|
91
|
+
tree: z.ZodArray<z.ZodType<import("../structure/types.js").StructureNode, unknown, z.core.$ZodTypeInternals<import("../structure/types.js").StructureNode, unknown>>>;
|
|
92
|
+
}, z.core.$strip>>;
|
|
93
|
+
contentPreview: z.ZodObject<{
|
|
94
|
+
truncated: z.ZodBoolean;
|
|
95
|
+
paragraphs: z.ZodArray<z.ZodObject<{
|
|
96
|
+
index: z.ZodNumber;
|
|
97
|
+
text: z.ZodString;
|
|
98
|
+
isHeading: z.ZodBoolean;
|
|
99
|
+
headingLevel: z.ZodOptional<z.ZodNumber>;
|
|
100
|
+
}, z.core.$strip>>;
|
|
101
|
+
}, z.core.$strip>;
|
|
102
|
+
fullTokenEstimate: z.ZodNumber;
|
|
103
|
+
previewTokenEstimate: z.ZodNumber;
|
|
104
|
+
}, z.core.$strip>;
|
|
105
|
+
export type DocumentInspection = z.infer<typeof DocumentInspectionSchema>;
|
|
106
|
+
export declare const DocInspectPayloadSchema: z.ZodObject<{
|
|
107
|
+
file: z.ZodString;
|
|
108
|
+
query: z.ZodRecord<z.ZodString, z.ZodUnknown>;
|
|
109
|
+
results: z.ZodObject<{
|
|
110
|
+
file: z.ZodString;
|
|
111
|
+
format: z.ZodEnum<{
|
|
112
|
+
docx: "docx";
|
|
113
|
+
pdf: "pdf";
|
|
114
|
+
odt: "odt";
|
|
115
|
+
}>;
|
|
116
|
+
size: z.ZodNumber;
|
|
117
|
+
properties: z.ZodObject<{
|
|
118
|
+
title: z.ZodOptional<z.ZodString>;
|
|
119
|
+
subject: z.ZodOptional<z.ZodString>;
|
|
120
|
+
author: z.ZodOptional<z.ZodString>;
|
|
121
|
+
company: z.ZodOptional<z.ZodString>;
|
|
122
|
+
createdDate: z.ZodOptional<z.ZodString>;
|
|
123
|
+
modifiedDate: z.ZodOptional<z.ZodString>;
|
|
124
|
+
lastModifiedBy: z.ZodOptional<z.ZodString>;
|
|
125
|
+
keywords: z.ZodOptional<z.ZodString>;
|
|
126
|
+
}, z.core.$strip>;
|
|
127
|
+
riskFlags: z.ZodObject<{
|
|
128
|
+
comments: z.ZodBoolean;
|
|
129
|
+
trackedChanges: z.ZodBoolean;
|
|
130
|
+
hyperlinks: z.ZodBoolean;
|
|
131
|
+
embeddedObjects: z.ZodBoolean;
|
|
132
|
+
footnotes: z.ZodBoolean;
|
|
133
|
+
endnotes: z.ZodBoolean;
|
|
134
|
+
macros: z.ZodBoolean;
|
|
135
|
+
headerFooter: z.ZodBoolean;
|
|
136
|
+
tables: z.ZodBoolean;
|
|
137
|
+
encrypted: z.ZodBoolean;
|
|
138
|
+
}, z.core.$strip>;
|
|
139
|
+
contentStats: z.ZodObject<{
|
|
140
|
+
words: z.ZodNumber;
|
|
141
|
+
pages: z.ZodNumber;
|
|
142
|
+
pagesEstimated: z.ZodBoolean;
|
|
143
|
+
paragraphs: z.ZodNumber;
|
|
144
|
+
characters: z.ZodNumber;
|
|
145
|
+
tables: z.ZodNumber;
|
|
146
|
+
images: z.ZodNumber;
|
|
147
|
+
}, z.core.$strip>;
|
|
148
|
+
structure: z.ZodNullable<z.ZodObject<{
|
|
149
|
+
headingCount: z.ZodNumber;
|
|
150
|
+
maxDepth: z.ZodNumber;
|
|
151
|
+
topLevelSections: z.ZodArray<z.ZodString>;
|
|
152
|
+
tree: z.ZodArray<z.ZodType<import("../structure/types.js").StructureNode, unknown, z.core.$ZodTypeInternals<import("../structure/types.js").StructureNode, unknown>>>;
|
|
153
|
+
}, z.core.$strip>>;
|
|
154
|
+
contentPreview: z.ZodObject<{
|
|
155
|
+
truncated: z.ZodBoolean;
|
|
156
|
+
paragraphs: z.ZodArray<z.ZodObject<{
|
|
157
|
+
index: z.ZodNumber;
|
|
158
|
+
text: z.ZodString;
|
|
159
|
+
isHeading: z.ZodBoolean;
|
|
160
|
+
headingLevel: z.ZodOptional<z.ZodNumber>;
|
|
161
|
+
}, z.core.$strip>>;
|
|
162
|
+
}, z.core.$strip>;
|
|
163
|
+
fullTokenEstimate: z.ZodNumber;
|
|
164
|
+
previewTokenEstimate: z.ZodNumber;
|
|
165
|
+
}, z.core.$strip>;
|
|
166
|
+
}, z.core.$strip>;
|
|
167
|
+
export type DocInspectPayload = z.infer<typeof DocInspectPayloadSchema>;
|
|
168
|
+
export declare const InspectDocOptionsSchema: z.ZodObject<{
|
|
169
|
+
sampleParagraphs: z.ZodNumber;
|
|
170
|
+
includeStructure: z.ZodBoolean;
|
|
171
|
+
}, z.core.$strip>;
|
|
172
|
+
export type InspectDocOptions = z.infer<typeof InspectDocOptionsSchema>;
|
|
@@ -0,0 +1,62 @@
|
|
|
1
|
+
import { z } from 'zod';
|
|
2
|
+
import { DocumentPropertiesSchema } from '../inspect/shared.js';
|
|
3
|
+
import { StructureNodeSchema } from '../structure/types.js';
|
|
4
|
+
export const DocRiskFlagsSchema = z.object({
|
|
5
|
+
comments: z.boolean(),
|
|
6
|
+
trackedChanges: z.boolean(),
|
|
7
|
+
hyperlinks: z.boolean(),
|
|
8
|
+
embeddedObjects: z.boolean(),
|
|
9
|
+
footnotes: z.boolean(),
|
|
10
|
+
endnotes: z.boolean(),
|
|
11
|
+
macros: z.boolean(),
|
|
12
|
+
headerFooter: z.boolean(),
|
|
13
|
+
tables: z.boolean(),
|
|
14
|
+
encrypted: z.boolean(),
|
|
15
|
+
});
|
|
16
|
+
export const DocContentStatsSchema = z.object({
|
|
17
|
+
words: z.number(),
|
|
18
|
+
pages: z.number(),
|
|
19
|
+
pagesEstimated: z.boolean(),
|
|
20
|
+
paragraphs: z.number(),
|
|
21
|
+
characters: z.number(),
|
|
22
|
+
tables: z.number(),
|
|
23
|
+
images: z.number(),
|
|
24
|
+
});
|
|
25
|
+
export const StructureSummarySchema = z.object({
|
|
26
|
+
headingCount: z.number(),
|
|
27
|
+
maxDepth: z.number(),
|
|
28
|
+
topLevelSections: z.array(z.string()),
|
|
29
|
+
tree: z.array(StructureNodeSchema),
|
|
30
|
+
});
|
|
31
|
+
export const ContentPreviewParagraphSchema = z.object({
|
|
32
|
+
index: z.number(),
|
|
33
|
+
text: z.string(),
|
|
34
|
+
isHeading: z.boolean(),
|
|
35
|
+
headingLevel: z.number().optional(),
|
|
36
|
+
});
|
|
37
|
+
export const ContentPreviewSchema = z.object({
|
|
38
|
+
truncated: z.boolean(),
|
|
39
|
+
paragraphs: z.array(ContentPreviewParagraphSchema),
|
|
40
|
+
});
|
|
41
|
+
export const DocumentInspectionSchema = z.object({
|
|
42
|
+
file: z.string(),
|
|
43
|
+
format: z.enum(['docx', 'pdf', 'odt']),
|
|
44
|
+
size: z.number(),
|
|
45
|
+
properties: DocumentPropertiesSchema,
|
|
46
|
+
riskFlags: DocRiskFlagsSchema,
|
|
47
|
+
contentStats: DocContentStatsSchema,
|
|
48
|
+
structure: StructureSummarySchema.nullable(),
|
|
49
|
+
contentPreview: ContentPreviewSchema,
|
|
50
|
+
fullTokenEstimate: z.number(),
|
|
51
|
+
previewTokenEstimate: z.number(),
|
|
52
|
+
});
|
|
53
|
+
export const DocInspectPayloadSchema = z.object({
|
|
54
|
+
file: z.string(),
|
|
55
|
+
query: z.record(z.string(), z.unknown()),
|
|
56
|
+
results: DocumentInspectionSchema,
|
|
57
|
+
});
|
|
58
|
+
export const InspectDocOptionsSchema = z.object({
|
|
59
|
+
sampleParagraphs: z.number(),
|
|
60
|
+
includeStructure: z.boolean(),
|
|
61
|
+
});
|
|
62
|
+
//# sourceMappingURL=types.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"types.js","sourceRoot":"","sources":["../../../src/doc/types.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,EAAE,wBAAwB,EAAE,MAAM,sBAAsB,CAAC;AAChE,OAAO,EAAE,mBAAmB,EAAE,MAAM,uBAAuB,CAAC;AAE5D,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,CAAC,MAAM,CAAC;IACzC,QAAQ,EAAE,CAAC,CAAC,OAAO,EAAE;IACrB,cAAc,EAAE,CAAC,CAAC,OAAO,EAAE;IAC3B,UAAU,EAAE,CAAC,CAAC,OAAO,EAAE;IACvB,eAAe,EAAE,CAAC,CAAC,OAAO,EAAE;IAC5B,SAAS,EAAE,CAAC,CAAC,OAAO,EAAE;IACtB,QAAQ,EAAE,CAAC,CAAC,OAAO,EAAE;IACrB,MAAM,EAAE,CAAC,CAAC,OAAO,EAAE;IACnB,YAAY,EAAE,CAAC,CAAC,OAAO,EAAE;IACzB,MAAM,EAAE,CAAC,CAAC,OAAO,EAAE;IACnB,SAAS,EAAE,CAAC,CAAC,OAAO,EAAE;CACvB,CAAC,CAAC;AAGH,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC5C,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE;IACjB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE;IACjB,cAAc,EAAE,CAAC,CAAC,OAAO,EAAE;IAC3B,UAAU,EAAE,CAAC,CAAC,MAAM,EAAE;IACtB,UAAU,EAAE,CAAC,CAAC,MAAM,EAAE;IACtB,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE;IAClB,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE;CACnB,CAAC,CAAC;AAGH,MAAM,CAAC,MAAM,sBAAsB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC7C,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE;IACxB,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE;IACpB,gBAAgB,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC;IACrC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,mBAAmB,CAAC;CACnC,CAAC,CAAC;AAGH,MAAM,CAAC,MAAM,6BAA6B,GAAG,CAAC,CAAC,MAAM,CAAC;IACpD,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE;IACjB,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE;IAChB,SAAS,EAAE,CAAC,CAAC,OAAO,EAAE;IACtB,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;CACpC,CAAC,CAAC;AAEH,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC3C,SAAS,EAAE,CAAC,CAAC,OAAO,EAAE;IACtB,UAAU,EAAE,CAAC,CAAC,KAAK,CAAC,6BAA6B,CAAC;CACnD,CAAC,CAAC;AAGH,MAAM,CAAC,MAAM,wBAAwB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC/C,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE;IAChB,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;IACtC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE;IAChB,UAAU,EAAE,wBAAwB;IACpC,SAAS,EAAE,kBAAkB;IAC7B,YAAY,EAAE,qBAAqB;IACnC,SAAS,EAAE,sBAAsB,CAAC,QAAQ,EAAE;IAC5C,cAAc,EAAE,oBAAoB;IACpC,iBAAiB,EAAE,CAAC,CAAC,MAAM,EAAE;IAC7B,oBAAoB,EAAE,CAAC,CAAC,MAAM,EAAE;CACjC,CAAC,CAAC;AAGH,MAAM,CAAC,MAAM,uBAAuB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC9C,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE;IAChB,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;IACxC,OAAO,EAAE,wBAAwB;CAClC,CAAC,CAAC;AAGH,MAAM,CAAC,MAAM,uBAAuB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC9C,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE;IAC5B,gBAAgB,EAAE,CAAC,CAAC,OAAO,EAAE;CAC9B,CAAC,CAAC"}
|
|
@@ -0,0 +1,20 @@
|
|
|
1
|
+
import { z } from 'zod';
|
|
2
|
+
export declare const DocumentPropertiesSchema: z.ZodObject<{
|
|
3
|
+
title: z.ZodOptional<z.ZodString>;
|
|
4
|
+
subject: z.ZodOptional<z.ZodString>;
|
|
5
|
+
author: z.ZodOptional<z.ZodString>;
|
|
6
|
+
company: z.ZodOptional<z.ZodString>;
|
|
7
|
+
createdDate: z.ZodOptional<z.ZodString>;
|
|
8
|
+
modifiedDate: z.ZodOptional<z.ZodString>;
|
|
9
|
+
lastModifiedBy: z.ZodOptional<z.ZodString>;
|
|
10
|
+
keywords: z.ZodOptional<z.ZodString>;
|
|
11
|
+
}, z.core.$strip>;
|
|
12
|
+
export type DocumentProperties = z.infer<typeof DocumentPropertiesSchema>;
|
|
13
|
+
export declare function estimateTokens(input: string | number): number;
|
|
14
|
+
export declare function asOptionalString(value: unknown): string | undefined;
|
|
15
|
+
export declare function formatDateLike(value: unknown): string | undefined;
|
|
16
|
+
export declare function createInspectPayload<T>(file: string, query: Record<string, unknown>, results: T): {
|
|
17
|
+
file: string;
|
|
18
|
+
query: Record<string, unknown>;
|
|
19
|
+
results: T;
|
|
20
|
+
};
|
|
@@ -0,0 +1,29 @@
|
|
|
1
|
+
import { z } from 'zod';
|
|
2
|
+
export const DocumentPropertiesSchema = z.object({
|
|
3
|
+
title: z.string().optional(),
|
|
4
|
+
subject: z.string().optional(),
|
|
5
|
+
author: z.string().optional(),
|
|
6
|
+
company: z.string().optional(),
|
|
7
|
+
createdDate: z.string().optional(),
|
|
8
|
+
modifiedDate: z.string().optional(),
|
|
9
|
+
lastModifiedBy: z.string().optional(),
|
|
10
|
+
keywords: z.string().optional(),
|
|
11
|
+
});
|
|
12
|
+
export function estimateTokens(input) {
|
|
13
|
+
const chars = typeof input === 'number' ? input : input.length;
|
|
14
|
+
return Math.max(0, Math.ceil(chars / 4));
|
|
15
|
+
}
|
|
16
|
+
export function asOptionalString(value) {
|
|
17
|
+
return typeof value === 'string' && value.trim() ? value : undefined;
|
|
18
|
+
}
|
|
19
|
+
export function formatDateLike(value) {
|
|
20
|
+
if (value instanceof Date)
|
|
21
|
+
return value.toISOString();
|
|
22
|
+
if (typeof value === 'string' && value.trim())
|
|
23
|
+
return value;
|
|
24
|
+
return undefined;
|
|
25
|
+
}
|
|
26
|
+
export function createInspectPayload(file, query, results) {
|
|
27
|
+
return { file, query, results };
|
|
28
|
+
}
|
|
29
|
+
//# sourceMappingURL=shared.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"shared.js","sourceRoot":"","sources":["../../../src/inspect/shared.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,MAAM,CAAC,MAAM,wBAAwB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC/C,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IAC5B,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IAC9B,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IAC7B,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IAC9B,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IAClC,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IACnC,cAAc,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IACrC,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;CAChC,CAAC,CAAC;AAGH,MAAM,UAAU,cAAc,CAAC,KAAsB;IACnD,MAAM,KAAK,GAAG,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC;IAC/D,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC;AAC3C,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,KAAc;IAC7C,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC;AACvE,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,KAAc;IAC3C,IAAI,KAAK,YAAY,IAAI;QAAE,OAAO,KAAK,CAAC,WAAW,EAAE,CAAC;IACtD,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,EAAE;QAAE,OAAO,KAAK,CAAC;IAC5D,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,MAAM,UAAU,oBAAoB,CAAI,IAAY,EAAE,KAA8B,EAAE,OAAU;IAC9F,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC;AAClC,CAAC"}
|
|
@@ -0,0 +1,4 @@
|
|
|
1
|
+
import type { CellObject, WorkSheet } from 'xlsx';
|
|
2
|
+
export declare function getCell(sheet: WorkSheet, row: number, col: number): CellObject | undefined;
|
|
3
|
+
export declare function renderCell(cell: CellObject | undefined): string;
|
|
4
|
+
export declare function isNonEmptyCell(cell: CellObject): boolean;
|
|
@@ -0,0 +1,31 @@
|
|
|
1
|
+
import XLSX from 'xlsx';
|
|
2
|
+
export function getCell(sheet, row, col) {
|
|
3
|
+
return sheet[XLSX.utils.encode_cell({ r: row, c: col })];
|
|
4
|
+
}
|
|
5
|
+
export function renderCell(cell) {
|
|
6
|
+
if (!cell)
|
|
7
|
+
return '';
|
|
8
|
+
if (typeof cell.w === 'string' && cell.w.trim())
|
|
9
|
+
return cell.w.trim();
|
|
10
|
+
if (cell.v instanceof Date)
|
|
11
|
+
return cell.v.toISOString();
|
|
12
|
+
if (cell.t === 'b')
|
|
13
|
+
return String(Boolean(cell.v));
|
|
14
|
+
if (cell.t === 'e')
|
|
15
|
+
return cell.w?.trim() || '#ERROR';
|
|
16
|
+
if (cell.v == null)
|
|
17
|
+
return cell.f?.trim() || '';
|
|
18
|
+
return String(cell.v).trim();
|
|
19
|
+
}
|
|
20
|
+
export function isNonEmptyCell(cell) {
|
|
21
|
+
if (cell.f)
|
|
22
|
+
return true;
|
|
23
|
+
if (cell.t === 'z')
|
|
24
|
+
return false;
|
|
25
|
+
if (cell.c?.length)
|
|
26
|
+
return true;
|
|
27
|
+
if (cell.l?.Target)
|
|
28
|
+
return true;
|
|
29
|
+
return renderCell(cell).length > 0;
|
|
30
|
+
}
|
|
31
|
+
//# sourceMappingURL=xlsx-cells.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"xlsx-cells.js","sourceRoot":"","sources":["../../../src/inspect/xlsx-cells.ts"],"names":[],"mappings":"AAAA,OAAO,IAAI,MAAM,MAAM,CAAC;AAGxB,MAAM,UAAU,OAAO,CAAC,KAAgB,EAAE,GAAW,EAAE,GAAW;IAChE,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,CAA2B,CAAC;AACrF,CAAC;AAED,MAAM,UAAU,UAAU,CAAC,IAA4B;IACrD,IAAI,CAAC,IAAI;QAAE,OAAO,EAAE,CAAC;IACrB,IAAI,OAAO,IAAI,CAAC,CAAC,KAAK,QAAQ,IAAI,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE;QAAE,OAAO,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IACtE,IAAI,IAAI,CAAC,CAAC,YAAY,IAAI;QAAE,OAAO,IAAI,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;IACxD,IAAI,IAAI,CAAC,CAAC,KAAK,GAAG;QAAE,OAAO,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IACnD,IAAI,IAAI,CAAC,CAAC,KAAK,GAAG;QAAE,OAAO,IAAI,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,QAAQ,CAAC;IACtD,IAAI,IAAI,CAAC,CAAC,IAAI,IAAI;QAAE,OAAO,IAAI,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IAChD,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;AAC/B,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,IAAgB;IAC7C,IAAI,IAAI,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IACxB,IAAI,IAAI,CAAC,CAAC,KAAK,GAAG;QAAE,OAAO,KAAK,CAAC;IACjC,IAAI,IAAI,CAAC,CAAC,EAAE,MAAM;QAAE,OAAO,IAAI,CAAC;IAChC,IAAI,IAAI,CAAC,CAAC,EAAE,MAAM;QAAE,OAAO,IAAI,CAAC;IAChC,OAAO,UAAU,CAAC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;AACrC,CAAC"}
|
|
@@ -4,8 +4,10 @@ import pdf from 'pdf-parse';
|
|
|
4
4
|
import JSZip from 'jszip';
|
|
5
5
|
import officeparser from 'officeparser';
|
|
6
6
|
import TurndownService from 'turndown';
|
|
7
|
+
import { z } from 'zod';
|
|
7
8
|
import { getExtension } from '../utils.js';
|
|
8
9
|
const turndown = new TurndownService({ headingStyle: 'atx' });
|
|
10
|
+
const PdfTextContentSchema = z.object({ items: z.array(z.object({ str: z.string() }).passthrough()) });
|
|
9
11
|
/** Convert a DOCX file to markdown via mammoth → HTML → turndown */
|
|
10
12
|
async function docxToMarkdown(filePath) {
|
|
11
13
|
const result = await mammoth.convertToHtml({ path: filePath });
|
|
@@ -29,7 +31,7 @@ async function pdfToMarkdown(filePath) {
|
|
|
29
31
|
try {
|
|
30
32
|
data = await pdf(buffer, {
|
|
31
33
|
pagerender: async (pageData) => {
|
|
32
|
-
const textContent = await pageData.getTextContent();
|
|
34
|
+
const textContent = PdfTextContentSchema.parse(await pageData.getTextContent());
|
|
33
35
|
const strings = textContent.items.map(item => item.str);
|
|
34
36
|
return `[Page ${pageData.pageIndex + 1}]\n${strings.join(' ')}`;
|
|
35
37
|
},
|
|
@@ -53,7 +55,7 @@ async function pptxToMarkdown(filePath) {
|
|
|
53
55
|
return numA - numB;
|
|
54
56
|
});
|
|
55
57
|
// Get full text via officeparser
|
|
56
|
-
const fullText = await officeparser.parseOffice(buffer);
|
|
58
|
+
const fullText = z.string().parse(await officeparser.parseOffice(buffer));
|
|
57
59
|
const slideCount = slideFiles.length;
|
|
58
60
|
if (slideCount <= 1) {
|
|
59
61
|
return `# Slide 1\n\n${fullText}`;
|
|
@@ -72,7 +74,7 @@ async function pptxToMarkdown(filePath) {
|
|
|
72
74
|
}
|
|
73
75
|
/** Convert an ODT file to markdown (best-effort heading detection) */
|
|
74
76
|
async function odtToMarkdown(filePath) {
|
|
75
|
-
const text = await officeparser.parseOffice(filePath);
|
|
77
|
+
const text = z.string().parse(await officeparser.parseOffice(filePath));
|
|
76
78
|
return text;
|
|
77
79
|
}
|
|
78
80
|
/** Convert an ODP file to markdown with slide headers */
|
|
@@ -83,7 +85,7 @@ async function odpToMarkdown(filePath) {
|
|
|
83
85
|
if (!contentXml)
|
|
84
86
|
return '';
|
|
85
87
|
const slides = (contentXml.match(/<draw:page /g) || []).length;
|
|
86
|
-
const text = await officeparser.parseOffice(buffer);
|
|
88
|
+
const text = z.string().parse(await officeparser.parseOffice(buffer));
|
|
87
89
|
if (slides <= 1) {
|
|
88
90
|
return `# Slide 1\n\n${text}`;
|
|
89
91
|
}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"convert.js","sourceRoot":"","sources":["../../../src/markdown/convert.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,OAAO,MAAM,SAAS,CAAC;AAC9B,OAAO,GAAG,MAAM,WAAW,CAAC;AAC5B,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,YAAY,MAAM,cAAc,CAAC;AACxC,OAAO,eAAe,MAAM,UAAU,CAAC;AACvC,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAE3C,MAAM,QAAQ,GAAG,IAAI,eAAe,CAAC,EAAE,YAAY,EAAE,KAAK,EAAE,CAAC,CAAC;
|
|
1
|
+
{"version":3,"file":"convert.js","sourceRoot":"","sources":["../../../src/markdown/convert.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,OAAO,MAAM,SAAS,CAAC;AAC9B,OAAO,GAAG,MAAM,WAAW,CAAC;AAC5B,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,YAAY,MAAM,cAAc,CAAC;AACxC,OAAO,eAAe,MAAM,UAAU,CAAC;AACvC,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAE3C,MAAM,QAAQ,GAAG,IAAI,eAAe,CAAC,EAAE,YAAY,EAAE,KAAK,EAAE,CAAC,CAAC;AAC9D,MAAM,oBAAoB,GAAG,CAAC,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC,EAAE,CAAC,CAAC;AAEvG,oEAAoE;AACpE,KAAK,UAAU,cAAc,CAAC,QAAgB;IAC5C,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,aAAa,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;IAC/D,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,IAAI,EAAE,CAAC;IAChC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE;QAAE,OAAO,EAAE,CAAC;IAC5B,OAAO,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;AACjC,CAAC;AAED,sDAAsD;AACtD,KAAK,UAAU,aAAa,CAAC,QAAgB;IAC3C,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;IAExC,2BAA2B;IAC3B,MAAM,WAAW,GAAG,OAAO,CAAC,GAAG,CAAC;IAChC,OAAO,CAAC,GAAG,GAAG,CAAC,GAAG,IAAe,EAAE,EAAE;QACnC,IAAI,OAAO,IAAI,CAAC,CAAC,CAAC,KAAK,QAAQ,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,WAAW,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,QAAQ,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,wBAAwB,CAAC,CAAC,EAAE,CAAC;YACrJ,OAAO;QACT,CAAC;QACD,WAAW,CAAC,KAAK,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;IACnC,CAAC,CAAC;IAEF,IAAI,IAAwC,CAAC;IAC7C,IAAI,CAAC;QACH,IAAI,GAAG,MAAM,GAAG,CAAC,MAAM,EAAE;YACvB,UAAU,EAAE,KAAK,EAAE,QAAuE,EAAE,EAAE;gBAC5F,MAAM,WAAW,GAAG,oBAAoB,CAAC,KAAK,CAAC,MAAM,QAAQ,CAAC,cAAc,EAAE,CAAC,CAAC;gBAChF,MAAM,OAAO,GAAG,WAAW,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gBACxD,OAAO,SAAS,QAAQ,CAAC,SAAS,GAAG,CAAC,MAAM,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YAClE,CAAC;SACF,CAAC,CAAC;IACL,CAAC;YAAS,CAAC;QACT,OAAO,CAAC,GAAG,GAAG,WAAW,CAAC;IAC5B,CAAC;IAED,OAAO,IAAI,CAAC,IAAI,CAAC;AACnB,CAAC;AAED,oDAAoD;AACpD,KAAK,UAAU,cAAc,CAAC,QAAgB;IAC5C,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACxC,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;IAE1C,uCAAuC;IACvC,MAAM,UAAU,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;SACtC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,8BAA8B,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;SACzD,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE;QACb,MAAM,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAC;QAC7D,MAAM,IAAI,GAAG,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,GAAG,EAAE,EAAE,CAAC,CAAC;QAC7D,OAAO,IAAI,GAAG,IAAI,CAAC;IACrB,CAAC,CAAC,CAAC;IAEL,iCAAiC;IACjC,MAAM,QAAQ,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,YAAY,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC;IAC1E,MAAM,UAAU,GAAG,UAAU,CAAC,MAAM,CAAC;IAErC,IAAI,UAAU,IAAI,CAAC,EAAE,CAAC;QACpB,OAAO,gBAAgB,QAAQ,EAAE,CAAC;IACpC,CAAC;IAED,oCAAoC;IACpC,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACnC,MAAM,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,UAAU,CAAC,CAAC,CAAC;IACxE,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,EAAE,CAAC,EAAE,EAAE,CAAC;QACpC,MAAM,KAAK,GAAG,CAAC,GAAG,aAAa,CAAC;QAChC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,aAAa,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;QAC5D,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5D,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,SAAS,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC5B,CAAC;AAED,sEAAsE;AACtE,KAAK,UAAU,aAAa,CAAC,QAAgB;IAC3C,MAAM,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,YAAY,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC;IACxE,OAAO,IAAI,CAAC;AACd,CAAC;AAED,yDAAyD;AACzD,KAAK,UAAU,aAAa,CAAC,QAAgB;IAC3C,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACxC,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;IAC1C,MAAM,UAAU,GAAG,MAAM,GAAG,CAAC,IAAI,CAAC,aAAa,CAAC,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IAChE,IAAI,CAAC,UAAU;QAAE,OAAO,EAAE,CAAC;IAE3B,MAAM,MAAM,GAAG,CAAC,UAAU,CAAC,KAAK,CAAC,cAAc,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;IAC/D,MAAM,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,KAAK,CAAC,MAAM,YAAY,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC;IAEtE,IAAI,MAAM,IAAI,CAAC,EAAE,CAAC;QAChB,OAAO,gBAAgB,IAAI,EAAE,CAAC;IAChC,CAAC;IAED,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC/B,MAAM,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC;IACpE,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAChC,MAAM,KAAK,GAAG,CAAC,GAAG,aAAa,CAAC;QAChC,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,aAAa,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;QAC5D,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5D,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,OAAO,SAAS,EAAE,CAAC,CAAC;IACjD,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAC5B,CAAC;AAED,wFAAwF;AACxF,MAAM,CAAC,KAAK,UAAU,kBAAkB,CAAC,QAAgB;IACvD,MAAM,GAAG,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IAEnC,QAAQ,GAAG,EAAE,CAAC;QACZ,KAAK,MAAM,CAAC,CAAC,OAAO,cAAc,CAAC,QAAQ,CAAC,CAAC;QAC7C,KAAK,KAAK,CAAC,CAAC,OAAO,aAAa,CAAC,QAAQ,CAAC,CAAC;QAC3C,KAAK,MAAM,CAAC,CAAC,OAAO,cAAc,CAAC,QAAQ,CAAC,CAAC;QAC7C,KAAK,KAAK,CAAC,CAAC,OAAO,aAAa,CAAC,QAAQ,CAAC,CAAC;QAC3C,KAAK,KAAK,CAAC,CAAC,OAAO,aAAa,CAAC,QAAQ,CAAC,CAAC;QAC3C,KAAK,MAAM,CAAC;QACZ,KAAK,KAAK;YACR,OAAO,IAAI,CAAC;QACd;YACE,OAAO,IAAI,CAAC;IAChB,CAAC;AACH,CAAC"}
|
|
@@ -1,9 +1,11 @@
|
|
|
1
|
+
import { z } from 'zod';
|
|
1
2
|
import type { AggregateResult } from '../stats.js';
|
|
2
3
|
import type { SccLanguage } from '../scc.js';
|
|
3
|
-
export
|
|
4
|
-
ci
|
|
5
|
-
byFile
|
|
6
|
-
}
|
|
4
|
+
export declare const TableOptionsSchema: z.ZodObject<{
|
|
5
|
+
ci: z.ZodOptional<z.ZodBoolean>;
|
|
6
|
+
byFile: z.ZodOptional<z.ZodBoolean>;
|
|
7
|
+
}, z.core.$strip>;
|
|
8
|
+
export type TableOptions = z.infer<typeof TableOptionsSchema>;
|
|
7
9
|
export declare function formatDocumentTable(stats: AggregateResult, options?: TableOptions): string;
|
|
8
10
|
export declare function formatSccTable(sccData: SccLanguage[], options?: TableOptions): string;
|
|
9
11
|
export declare function formatSummaryLine(stats: AggregateResult, sccData: SccLanguage[] | null, elapsed: number, options?: TableOptions): string;
|
|
@@ -1,6 +1,11 @@
|
|
|
1
1
|
import Table from 'cli-table3';
|
|
2
2
|
import chalk from 'chalk';
|
|
3
|
+
import { z } from 'zod';
|
|
3
4
|
import { formatBytes, formatNumber } from '../utils.js';
|
|
5
|
+
export const TableOptionsSchema = z.object({
|
|
6
|
+
ci: z.boolean().optional(),
|
|
7
|
+
byFile: z.boolean().optional(),
|
|
8
|
+
});
|
|
4
9
|
export function formatDocumentTable(stats, options = {}) {
|
|
5
10
|
const { ci = false } = options;
|
|
6
11
|
const c = ci ? noColor : colorize;
|