ai-document-parser 0.1.101 → 0.1.104
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 +3 -2
- package/dist/convert/blocks.d.ts +14 -0
- package/dist/convert/blocks.d.ts.map +1 -0
- package/dist/convert/blocks.js +254 -0
- package/dist/convert/blocks.js.map +1 -0
- package/dist/convert/converter.d.ts.map +1 -1
- package/dist/convert/converter.js +84 -62
- package/dist/convert/converter.js.map +1 -1
- package/dist/convert/emit.d.ts +1 -0
- package/dist/convert/emit.d.ts.map +1 -1
- package/dist/convert/emit.js +32 -33
- package/dist/convert/emit.js.map +1 -1
- package/dist/edit/docx-ooxml.d.ts.map +1 -1
- package/dist/edit/docx-ooxml.js +5 -33
- package/dist/edit/docx-ooxml.js.map +1 -1
- package/dist/index.d.ts +1 -1
- package/dist/index.d.ts.map +1 -1
- package/dist/parse/docx-blocks.d.ts +18 -0
- package/dist/parse/docx-blocks.d.ts.map +1 -0
- package/dist/parse/docx-blocks.js +654 -0
- package/dist/parse/docx-blocks.js.map +1 -0
- package/dist/parse/html-blocks.d.ts +14 -0
- package/dist/parse/html-blocks.d.ts.map +1 -0
- package/dist/parse/html-blocks.js +244 -0
- package/dist/parse/html-blocks.js.map +1 -0
- package/dist/parse/pptx.d.ts +0 -1
- package/dist/parse/pptx.d.ts.map +1 -1
- package/dist/parse/pptx.js +95 -24
- package/dist/parse/pptx.js.map +1 -1
- package/dist/parse/router.d.ts.map +1 -1
- package/dist/parse/router.js +25 -6
- package/dist/parse/router.js.map +1 -1
- package/dist/types.d.ts +37 -0
- package/dist/types.d.ts.map +1 -1
- package/package.json +2 -2
package/README.md
CHANGED
|
@@ -21,8 +21,8 @@ import { parseDocument, convertDocument, editDocument, writeDocument } from 'ai-
|
|
|
21
21
|
const result = await parseDocument('report.docx', { output: 'text' });
|
|
22
22
|
// output: 'text' | 'structured' | 'ast' | 'chunks'
|
|
23
23
|
|
|
24
|
-
// Convert
|
|
25
|
-
const { value } = await convertDocument('
|
|
24
|
+
// Convert preserves structure where possible (headings, tables, emphasis)
|
|
25
|
+
const { value } = await convertDocument('report.docx', 'md');
|
|
26
26
|
const slides = await convertDocument('deck.pptx', 'md'); // ## Slide N …
|
|
27
27
|
|
|
28
28
|
// Surgical edit
|
|
@@ -76,6 +76,7 @@ See [docs/AGENT_GUIDE.md](docs/AGENT_GUIDE.md).
|
|
|
76
76
|
|
|
77
77
|
Notes:
|
|
78
78
|
|
|
79
|
+
- Convert → **md/html** preserves structure for DOCX (headings, tables, bold/italic/strike, ordered lists, hyperlinks, footnotes; images omitted), HTML, CSV (GFM table), PPTX slides, XLSX sheets. MD→md is passthrough. PDF / legacy `.doc` stay plain text.
|
|
79
80
|
- Convert targets `rtf` / `epub` / `pdf` are **not** supported (create PDF with `writeDocument`).
|
|
80
81
|
- **DOCX** and **PPTX** cannot convert **to csv** (use a spreadsheet source).
|
|
81
82
|
|
|
@@ -0,0 +1,14 @@
|
|
|
1
|
+
import type { ContentBlock, InlineRun } from '../types.js';
|
|
2
|
+
export type { ContentBlock, InlineRun };
|
|
3
|
+
export declare function runsToPlain(runs: InlineRun[]): string;
|
|
4
|
+
export declare function flattenBlocksToParagraphs(blocks: ContentBlock[]): string[];
|
|
5
|
+
/** Merge adjacent runs with identical formatting. */
|
|
6
|
+
export declare function mergeRuns(runs: InlineRun[]): InlineRun[];
|
|
7
|
+
export declare function runsToMarkdown(runs: InlineRun[]): string;
|
|
8
|
+
export declare function emitMarkdownTable(rows: string[][]): string;
|
|
9
|
+
export declare function emitMarkdownFromBlocks(blocks: ContentBlock[]): string;
|
|
10
|
+
export declare function emitHtmlFromBlocks(blocks: ContentBlock[]): string;
|
|
11
|
+
/** Build blocks from spreadsheet sheets (heading per sheet + table). */
|
|
12
|
+
export declare function blocksFromSpreadsheet(sheets: Record<string, (string | number | boolean | null)[][]>): ContentBlock[];
|
|
13
|
+
export declare function blocksFromCsvRows(rows: string[][]): ContentBlock[];
|
|
14
|
+
//# sourceMappingURL=blocks.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"blocks.d.ts","sourceRoot":"","sources":["../../src/convert/blocks.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,SAAS,EAAE,MAAM,aAAa,CAAC;AAG3D,YAAY,EAAE,YAAY,EAAE,SAAS,EAAE,CAAC;AAExC,wBAAgB,WAAW,CAAC,IAAI,EAAE,SAAS,EAAE,GAAG,MAAM,CAErD;AAED,wBAAgB,yBAAyB,CAAC,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,EAAE,CAkB1E;AAsBD,qDAAqD;AACrD,wBAAgB,SAAS,CAAC,IAAI,EAAE,SAAS,EAAE,GAAG,SAAS,EAAE,CAkBxD;AAED,wBAAgB,cAAc,CAAC,IAAI,EAAE,SAAS,EAAE,GAAG,MAAM,CAuBxD;AAED,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE,GAAG,MAAM,CA6B1D;AAED,wBAAgB,sBAAsB,CAAC,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,CAqCrE;AAED,wBAAgB,kBAAkB,CAAC,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,CAyDjE;AAED,wEAAwE;AACxE,wBAAgB,qBAAqB,CACnC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,CAAC,MAAM,GAAG,MAAM,GAAG,OAAO,GAAG,IAAI,CAAC,EAAE,EAAE,CAAC,GAC7D,YAAY,EAAE,CAgBhB;AAED,wBAAgB,iBAAiB,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE,GAAG,YAAY,EAAE,CAGlE"}
|
|
@@ -0,0 +1,254 @@
|
|
|
1
|
+
import { escapeHtml } from '../utils/text.js';
|
|
2
|
+
export function runsToPlain(runs) {
|
|
3
|
+
return runs.map((r) => r.text).join('');
|
|
4
|
+
}
|
|
5
|
+
export function flattenBlocksToParagraphs(blocks) {
|
|
6
|
+
const out = [];
|
|
7
|
+
for (const b of blocks) {
|
|
8
|
+
if (b.type === 'heading' || b.type === 'paragraph' || b.type === 'list_item') {
|
|
9
|
+
const t = runsToPlain(b.runs).trim();
|
|
10
|
+
// Keep edit targets free of GFM footnote definition lines.
|
|
11
|
+
if (t && !/^\[\^[\w-]+\]:/.test(t))
|
|
12
|
+
out.push(t);
|
|
13
|
+
}
|
|
14
|
+
else if (b.type === 'table') {
|
|
15
|
+
for (const row of b.rows) {
|
|
16
|
+
const line = row.map((c) => String(c ?? '').trim()).join(' | ');
|
|
17
|
+
if (line.replace(/\s*\|\s*/g, '').trim())
|
|
18
|
+
out.push(line);
|
|
19
|
+
}
|
|
20
|
+
}
|
|
21
|
+
else if (b.type === 'code') {
|
|
22
|
+
const t = b.text.trim();
|
|
23
|
+
if (t)
|
|
24
|
+
out.push(t);
|
|
25
|
+
}
|
|
26
|
+
}
|
|
27
|
+
return out;
|
|
28
|
+
}
|
|
29
|
+
function escapeMdCell(s) {
|
|
30
|
+
return String(s ?? '')
|
|
31
|
+
.replace(/\|/g, '\\|')
|
|
32
|
+
.replace(/\n/g, ' ')
|
|
33
|
+
.trim();
|
|
34
|
+
}
|
|
35
|
+
/** Escape only characters that break inline markdown emphasis/links. */
|
|
36
|
+
function escapeMdInline(s) {
|
|
37
|
+
return s.replace(/([\\`*_\[\]])/g, '\\$1');
|
|
38
|
+
}
|
|
39
|
+
function runKey(r) {
|
|
40
|
+
return `${r.bold ? 1 : 0}|${r.italic ? 1 : 0}|${r.strike ? 1 : 0}|${r.href ?? ''}`;
|
|
41
|
+
}
|
|
42
|
+
function isFootnoteAtom(text) {
|
|
43
|
+
return /^\[\^[\w-]+\](?::[\s\S]*)?$/.test(text);
|
|
44
|
+
}
|
|
45
|
+
/** Merge adjacent runs with identical formatting. */
|
|
46
|
+
export function mergeRuns(runs) {
|
|
47
|
+
if (!runs.length)
|
|
48
|
+
return runs;
|
|
49
|
+
const out = [{ ...runs[0], text: runs[0].text }];
|
|
50
|
+
for (let i = 1; i < runs.length; i++) {
|
|
51
|
+
const prev = out[out.length - 1];
|
|
52
|
+
const cur = runs[i];
|
|
53
|
+
// Keep GFM footnote markers as standalone atoms so emit can leave brackets raw.
|
|
54
|
+
if (isFootnoteAtom(prev.text) || isFootnoteAtom(cur.text)) {
|
|
55
|
+
out.push({ ...cur, text: cur.text });
|
|
56
|
+
continue;
|
|
57
|
+
}
|
|
58
|
+
if (runKey(prev) === runKey(cur)) {
|
|
59
|
+
prev.text += cur.text;
|
|
60
|
+
}
|
|
61
|
+
else {
|
|
62
|
+
out.push({ ...cur, text: cur.text });
|
|
63
|
+
}
|
|
64
|
+
}
|
|
65
|
+
return out;
|
|
66
|
+
}
|
|
67
|
+
export function runsToMarkdown(runs) {
|
|
68
|
+
// Use ** for bold and _ for italic so adjacent runs cannot form **** / *** traps
|
|
69
|
+
// that hang markdown-it (VS Code / Cursor preview). Keep spaces outside markers.
|
|
70
|
+
return mergeRuns(runs)
|
|
71
|
+
.map((r) => {
|
|
72
|
+
if (/^\[\^[\w-]+\](?::[\s\S]*)?$/.test(r.text) && !r.href && !r.bold && !r.italic) {
|
|
73
|
+
return r.text;
|
|
74
|
+
}
|
|
75
|
+
const m = /^(?<lead>\s*)(?<core>[\s\S]*?)(?<trail>\s*)$/.exec(r.text);
|
|
76
|
+
const lead = m?.groups?.lead ?? '';
|
|
77
|
+
const core = m?.groups?.core ?? r.text;
|
|
78
|
+
const trail = m?.groups?.trail ?? '';
|
|
79
|
+
if (!core)
|
|
80
|
+
return escapeMdInline(r.text);
|
|
81
|
+
let t = escapeMdInline(core);
|
|
82
|
+
if (r.strike)
|
|
83
|
+
t = `~~${t}~~`;
|
|
84
|
+
if (r.bold && r.italic)
|
|
85
|
+
t = `**_${t}_**`;
|
|
86
|
+
else if (r.bold)
|
|
87
|
+
t = `**${t}**`;
|
|
88
|
+
else if (r.italic)
|
|
89
|
+
t = `_${t}_`;
|
|
90
|
+
if (r.href)
|
|
91
|
+
t = `[${t}](${r.href})`;
|
|
92
|
+
return `${lead}${t}${trail}`;
|
|
93
|
+
})
|
|
94
|
+
.join('');
|
|
95
|
+
}
|
|
96
|
+
export function emitMarkdownTable(rows) {
|
|
97
|
+
if (!rows.length)
|
|
98
|
+
return '';
|
|
99
|
+
const width = Math.max(...rows.map((r) => r.length), 1);
|
|
100
|
+
let norm = rows.map((r) => {
|
|
101
|
+
const cells = r.map((c) => String(c ?? '').trim());
|
|
102
|
+
while (cells.length < width)
|
|
103
|
+
cells.push('');
|
|
104
|
+
return cells.slice(0, width);
|
|
105
|
+
});
|
|
106
|
+
// Drop columns that are empty in every row (common after Word merged-cell calendars).
|
|
107
|
+
const keepCols = [];
|
|
108
|
+
for (let c = 0; c < width; c++) {
|
|
109
|
+
if (norm.some((row) => row[c]))
|
|
110
|
+
keepCols.push(c);
|
|
111
|
+
}
|
|
112
|
+
if (!keepCols.length)
|
|
113
|
+
return '';
|
|
114
|
+
norm = norm.map((row) => keepCols.map((c) => row[c]));
|
|
115
|
+
// Drop fully empty body rows (keep header even if sparse).
|
|
116
|
+
norm = norm.filter((row, i) => i === 0 || row.some((c) => c));
|
|
117
|
+
const escaped = norm.map((r) => r.map(escapeMdCell));
|
|
118
|
+
const header = escaped[0];
|
|
119
|
+
const sep = header.map(() => '---');
|
|
120
|
+
const lines = [
|
|
121
|
+
`| ${header.join(' | ')} |`,
|
|
122
|
+
`| ${sep.join(' | ')} |`,
|
|
123
|
+
...escaped.slice(1).map((r) => `| ${r.join(' | ')} |`),
|
|
124
|
+
];
|
|
125
|
+
return lines.join('\n');
|
|
126
|
+
}
|
|
127
|
+
export function emitMarkdownFromBlocks(blocks) {
|
|
128
|
+
const parts = [];
|
|
129
|
+
let listBuf = [];
|
|
130
|
+
const flushList = () => {
|
|
131
|
+
if (listBuf.length) {
|
|
132
|
+
parts.push(listBuf.join('\n'));
|
|
133
|
+
listBuf = [];
|
|
134
|
+
}
|
|
135
|
+
};
|
|
136
|
+
for (const b of blocks) {
|
|
137
|
+
if (b.type === 'list_item') {
|
|
138
|
+
const indent = ' '.repeat(Math.max(0, b.level));
|
|
139
|
+
const marker = b.ordered ? '1.' : '-';
|
|
140
|
+
listBuf.push(`${indent}${marker} ${runsToMarkdown(b.runs)}`.trimEnd());
|
|
141
|
+
continue;
|
|
142
|
+
}
|
|
143
|
+
flushList();
|
|
144
|
+
if (b.type === 'heading') {
|
|
145
|
+
const hashes = '#'.repeat(b.level);
|
|
146
|
+
parts.push(`${hashes} ${runsToMarkdown(b.runs)}`.trimEnd());
|
|
147
|
+
}
|
|
148
|
+
else if (b.type === 'paragraph') {
|
|
149
|
+
const line = runsToMarkdown(b.runs).trim();
|
|
150
|
+
if (line)
|
|
151
|
+
parts.push(line);
|
|
152
|
+
}
|
|
153
|
+
else if (b.type === 'table') {
|
|
154
|
+
const md = emitMarkdownTable(b.rows);
|
|
155
|
+
if (md)
|
|
156
|
+
parts.push(md);
|
|
157
|
+
}
|
|
158
|
+
else if (b.type === 'thematic_break') {
|
|
159
|
+
parts.push('---');
|
|
160
|
+
}
|
|
161
|
+
else if (b.type === 'code') {
|
|
162
|
+
const lang = b.language ?? '';
|
|
163
|
+
parts.push(`\`\`\`${lang}\n${b.text.replace(/\n$/, '')}\n\`\`\``);
|
|
164
|
+
}
|
|
165
|
+
}
|
|
166
|
+
flushList();
|
|
167
|
+
return parts.join('\n\n').trim();
|
|
168
|
+
}
|
|
169
|
+
export function emitHtmlFromBlocks(blocks) {
|
|
170
|
+
const parts = [];
|
|
171
|
+
let listOpen = null;
|
|
172
|
+
const closeList = () => {
|
|
173
|
+
if (listOpen) {
|
|
174
|
+
parts.push(`</${listOpen}>`);
|
|
175
|
+
listOpen = null;
|
|
176
|
+
}
|
|
177
|
+
};
|
|
178
|
+
const runsToHtml = (runs) => mergeRuns(runs)
|
|
179
|
+
.map((r) => {
|
|
180
|
+
let t = escapeHtml(r.text);
|
|
181
|
+
if (r.strike)
|
|
182
|
+
t = `<s>${t}</s>`;
|
|
183
|
+
if (r.bold)
|
|
184
|
+
t = `<strong>${t}</strong>`;
|
|
185
|
+
if (r.italic)
|
|
186
|
+
t = `<em>${t}</em>`;
|
|
187
|
+
if (r.href)
|
|
188
|
+
t = `<a href="${escapeHtml(r.href)}">${t}</a>`;
|
|
189
|
+
return t;
|
|
190
|
+
})
|
|
191
|
+
.join('');
|
|
192
|
+
for (const b of blocks) {
|
|
193
|
+
if (b.type === 'list_item') {
|
|
194
|
+
const tag = b.ordered ? 'ol' : 'ul';
|
|
195
|
+
if (listOpen !== tag) {
|
|
196
|
+
closeList();
|
|
197
|
+
parts.push(`<${tag}>`);
|
|
198
|
+
listOpen = tag;
|
|
199
|
+
}
|
|
200
|
+
parts.push(`<li>${runsToHtml(b.runs)}</li>`);
|
|
201
|
+
continue;
|
|
202
|
+
}
|
|
203
|
+
closeList();
|
|
204
|
+
if (b.type === 'heading') {
|
|
205
|
+
parts.push(`<h${b.level}>${runsToHtml(b.runs)}</h${b.level}>`);
|
|
206
|
+
}
|
|
207
|
+
else if (b.type === 'paragraph') {
|
|
208
|
+
parts.push(`<p>${runsToHtml(b.runs)}</p>`);
|
|
209
|
+
}
|
|
210
|
+
else if (b.type === 'table') {
|
|
211
|
+
const body = b.rows
|
|
212
|
+
.map((row, i) => {
|
|
213
|
+
const cell = i === 0 ? 'th' : 'td';
|
|
214
|
+
return `<tr>${row
|
|
215
|
+
.map((c) => `<${cell}>${escapeHtml(String(c ?? ''))}</${cell}>`)
|
|
216
|
+
.join('')}</tr>`;
|
|
217
|
+
})
|
|
218
|
+
.join('');
|
|
219
|
+
parts.push(`<table>${body}</table>`);
|
|
220
|
+
}
|
|
221
|
+
else if (b.type === 'thematic_break') {
|
|
222
|
+
parts.push('<hr />');
|
|
223
|
+
}
|
|
224
|
+
else if (b.type === 'code') {
|
|
225
|
+
parts.push(`<pre><code>${escapeHtml(b.text)}</code></pre>`);
|
|
226
|
+
}
|
|
227
|
+
}
|
|
228
|
+
closeList();
|
|
229
|
+
return `<!DOCTYPE html>\n<html><body>\n${parts.join('\n')}\n</body></html>\n`;
|
|
230
|
+
}
|
|
231
|
+
/** Build blocks from spreadsheet sheets (heading per sheet + table). */
|
|
232
|
+
export function blocksFromSpreadsheet(sheets) {
|
|
233
|
+
const blocks = [];
|
|
234
|
+
for (const [name, grid] of Object.entries(sheets)) {
|
|
235
|
+
blocks.push({
|
|
236
|
+
type: 'heading',
|
|
237
|
+
level: 2,
|
|
238
|
+
runs: [{ text: name }],
|
|
239
|
+
});
|
|
240
|
+
if (grid.length) {
|
|
241
|
+
blocks.push({
|
|
242
|
+
type: 'table',
|
|
243
|
+
rows: grid.map((row) => row.map((c) => (c == null ? '' : String(c)))),
|
|
244
|
+
});
|
|
245
|
+
}
|
|
246
|
+
}
|
|
247
|
+
return blocks;
|
|
248
|
+
}
|
|
249
|
+
export function blocksFromCsvRows(rows) {
|
|
250
|
+
if (!rows.length)
|
|
251
|
+
return [];
|
|
252
|
+
return [{ type: 'table', rows }];
|
|
253
|
+
}
|
|
254
|
+
//# sourceMappingURL=blocks.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"blocks.js","sourceRoot":"","sources":["../../src/convert/blocks.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAI9C,MAAM,UAAU,WAAW,CAAC,IAAiB;IAC3C,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAC1C,CAAC;AAED,MAAM,UAAU,yBAAyB,CAAC,MAAsB;IAC9D,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,IAAI,CAAC,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,CAAC,IAAI,KAAK,WAAW,IAAI,CAAC,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;YAC7E,MAAM,CAAC,GAAG,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;YACrC,2DAA2D;YAC3D,IAAI,CAAC,IAAI,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC;gBAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAClD,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC9B,KAAK,MAAM,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;gBACzB,MAAM,IAAI,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBAChE,IAAI,IAAI,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE;oBAAE,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC3D,CAAC;QACH,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC7B,MAAM,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YACxB,IAAI,CAAC;gBAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACrB,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,YAAY,CAAC,CAAS;IAC7B,OAAO,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;SACnB,OAAO,CAAC,KAAK,EAAE,KAAK,CAAC;SACrB,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC;SACnB,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,wEAAwE;AACxE,SAAS,cAAc,CAAC,CAAS;IAC/B,OAAO,CAAC,CAAC,OAAO,CAAC,gBAAgB,EAAE,MAAM,CAAC,CAAC;AAC7C,CAAC;AAED,SAAS,MAAM,CAAC,CAAY;IAC1B,OAAO,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,EAAE,EAAE,CAAC;AACrF,CAAC;AAED,SAAS,cAAc,CAAC,IAAY;IAClC,OAAO,6BAA6B,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAClD,CAAC;AAED,qDAAqD;AACrD,MAAM,UAAU,SAAS,CAAC,IAAiB;IACzC,IAAI,CAAC,IAAI,CAAC,MAAM;QAAE,OAAO,IAAI,CAAC;IAC9B,MAAM,GAAG,GAAgB,CAAC,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;IAC9D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,IAAI,GAAG,GAAG,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;QACjC,MAAM,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACpB,gFAAgF;QAChF,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC;YAC1D,GAAG,CAAC,IAAI,CAAC,EAAE,GAAG,GAAG,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;YACrC,SAAS;QACX,CAAC;QACD,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;YACjC,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC,IAAI,CAAC;QACxB,CAAC;aAAM,CAAC;YACN,GAAG,CAAC,IAAI,CAAC,EAAE,GAAG,GAAG,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,IAAiB;IAC9C,iFAAiF;IACjF,iFAAiF;IACjF,OAAO,SAAS,CAAC,IAAI,CAAC;SACnB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;QACT,IAAI,6BAA6B,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC;YAClF,OAAO,CAAC,CAAC,IAAI,CAAC;QAChB,CAAC;QACD,MAAM,CAAC,GAAG,8CAA8C,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QACtE,MAAM,IAAI,GAAG,CAAC,EAAE,MAAM,EAAE,IAAI,IAAI,EAAE,CAAC;QACnC,MAAM,IAAI,GAAG,CAAC,EAAE,MAAM,EAAE,IAAI,IAAI,CAAC,CAAC,IAAI,CAAC;QACvC,MAAM,KAAK,GAAG,CAAC,EAAE,MAAM,EAAE,KAAK,IAAI,EAAE,CAAC;QACrC,IAAI,CAAC,IAAI;YAAE,OAAO,cAAc,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAEzC,IAAI,CAAC,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;QAC7B,IAAI,CAAC,CAAC,MAAM;YAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC;QAC7B,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,MAAM;YAAE,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC;aACpC,IAAI,CAAC,CAAC,IAAI;YAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC;aAC3B,IAAI,CAAC,CAAC,MAAM;YAAE,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC;QAChC,IAAI,CAAC,CAAC,IAAI;YAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,GAAG,CAAC;QACpC,OAAO,GAAG,IAAI,GAAG,CAAC,GAAG,KAAK,EAAE,CAAC;IAC/B,CAAC,CAAC;SACD,IAAI,CAAC,EAAE,CAAC,CAAC;AACd,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,IAAgB;IAChD,IAAI,CAAC,IAAI,CAAC,MAAM;QAAE,OAAO,EAAE,CAAC;IAC5B,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC;IACxD,IAAI,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;QACxB,MAAM,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QACnD,OAAO,KAAK,CAAC,MAAM,GAAG,KAAK;YAAE,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAC5C,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAC/B,CAAC,CAAC,CAAC;IAEH,sFAAsF;IACtF,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/B,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YAAE,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACnD,CAAC;IACD,IAAI,CAAC,QAAQ,CAAC,MAAM;QAAE,OAAO,EAAE,CAAC;IAChC,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAEtD,2DAA2D;IAC3D,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAE9D,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC;IACrD,MAAM,MAAM,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;IAC1B,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC;IACpC,MAAM,KAAK,GAAG;QACZ,KAAK,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI;QAC3B,KAAK,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI;QACxB,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC;KACvD,CAAC;IACF,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED,MAAM,UAAU,sBAAsB,CAAC,MAAsB;IAC3D,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,OAAO,GAAa,EAAE,CAAC;IAE3B,MAAM,SAAS,GAAG,GAAG,EAAE;QACrB,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;YACnB,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;YAC/B,OAAO,GAAG,EAAE,CAAC;QACf,CAAC;IACH,CAAC,CAAC;IAEF,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,IAAI,CAAC,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;YAC3B,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC;YACjD,MAAM,MAAM,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC;YACtC,OAAO,CAAC,IAAI,CAAC,GAAG,MAAM,GAAG,MAAM,IAAI,cAAc,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,CAAC;YACvE,SAAS;QACX,CAAC;QACD,SAAS,EAAE,CAAC;QACZ,IAAI,CAAC,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YACzB,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;YACnC,KAAK,CAAC,IAAI,CAAC,GAAG,MAAM,IAAI,cAAc,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,CAAC;QAC9D,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;YAClC,MAAM,IAAI,GAAG,cAAc,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;YAC3C,IAAI,IAAI;gBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC9B,MAAM,EAAE,GAAG,iBAAiB,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YACrC,IAAI,EAAE;gBAAE,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QACzB,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,gBAAgB,EAAE,CAAC;YACvC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACpB,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC7B,MAAM,IAAI,GAAG,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC;YAC9B,KAAK,CAAC,IAAI,CAAC,SAAS,IAAI,KAAK,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC;QACpE,CAAC;IACH,CAAC;IACD,SAAS,EAAE,CAAC;IACZ,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;AACnC,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,MAAsB;IACvD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,QAAQ,GAAuB,IAAI,CAAC;IAExC,MAAM,SAAS,GAAG,GAAG,EAAE;QACrB,IAAI,QAAQ,EAAE,CAAC;YACb,KAAK,CAAC,IAAI,CAAC,KAAK,QAAQ,GAAG,CAAC,CAAC;YAC7B,QAAQ,GAAG,IAAI,CAAC;QAClB,CAAC;IACH,CAAC,CAAC;IAEF,MAAM,UAAU,GAAG,CAAC,IAAiB,EAAU,EAAE,CAC/C,SAAS,CAAC,IAAI,CAAC;SACZ,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;QACT,IAAI,CAAC,GAAG,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAC3B,IAAI,CAAC,CAAC,MAAM;YAAE,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC;QAChC,IAAI,CAAC,CAAC,IAAI;YAAE,CAAC,GAAG,WAAW,CAAC,WAAW,CAAC;QACxC,IAAI,CAAC,CAAC,MAAM;YAAE,CAAC,GAAG,OAAO,CAAC,OAAO,CAAC;QAClC,IAAI,CAAC,CAAC,IAAI;YAAE,CAAC,GAAG,YAAY,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC;QAC3D,OAAO,CAAC,CAAC;IACX,CAAC,CAAC;SACD,IAAI,CAAC,EAAE,CAAC,CAAC;IAEd,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,IAAI,CAAC,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;YAC3B,MAAM,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;YACpC,IAAI,QAAQ,KAAK,GAAG,EAAE,CAAC;gBACrB,SAAS,EAAE,CAAC;gBACZ,KAAK,CAAC,IAAI,CAAC,IAAI,GAAG,GAAG,CAAC,CAAC;gBACvB,QAAQ,GAAG,GAAG,CAAC;YACjB,CAAC;YACD,KAAK,CAAC,IAAI,CAAC,OAAO,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YAC7C,SAAS;QACX,CAAC;QACD,SAAS,EAAE,CAAC;QACZ,IAAI,CAAC,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YACzB,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,KAAK,IAAI,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC;QACjE,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;YAClC,KAAK,CAAC,IAAI,CAAC,MAAM,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAC7C,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,OAAO,EAAE,CAAC;YAC9B,MAAM,IAAI,GAAG,CAAC,CAAC,IAAI;iBAChB,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE;gBACd,MAAM,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;gBACnC,OAAO,OAAO,GAAG;qBACd,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,IAAI,IAAI,UAAU,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,IAAI,GAAG,CAAC;qBAC/D,IAAI,CAAC,EAAE,CAAC,OAAO,CAAC;YACrB,CAAC,CAAC;iBACD,IAAI,CAAC,EAAE,CAAC,CAAC;YACZ,KAAK,CAAC,IAAI,CAAC,UAAU,IAAI,UAAU,CAAC,CAAC;QACvC,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,gBAAgB,EAAE,CAAC;YACvC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QACvB,CAAC;aAAM,IAAI,CAAC,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YAC7B,KAAK,CAAC,IAAI,CAAC,cAAc,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;QAC9D,CAAC;IACH,CAAC;IACD,SAAS,EAAE,CAAC;IACZ,OAAO,kCAAkC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,oBAAoB,CAAC;AAChF,CAAC;AAED,wEAAwE;AACxE,MAAM,UAAU,qBAAqB,CACnC,MAA8D;IAE9D,MAAM,MAAM,GAAmB,EAAE,CAAC;IAClC,KAAK,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CAAC;QAClD,MAAM,CAAC,IAAI,CAAC;YACV,IAAI,EAAE,SAAS;YACf,KAAK,EAAE,CAAC;YACR,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;SACvB,CAAC,CAAC;QACH,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YAChB,MAAM,CAAC,IAAI,CAAC;gBACV,IAAI,EAAE,OAAO;gBACb,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;aACtE,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,IAAgB;IAChD,IAAI,CAAC,IAAI,CAAC,MAAM;QAAE,OAAO,EAAE,CAAC;IAC5B,OAAO,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;AACnC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"converter.d.ts","sourceRoot":"","sources":["../../src/convert/converter.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EACV,aAAa,EACb,cAAc,EACd,aAAa,EACb,aAAa,
|
|
1
|
+
{"version":3,"file":"converter.d.ts","sourceRoot":"","sources":["../../src/convert/converter.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EACV,aAAa,EACb,cAAc,EACd,aAAa,EACb,aAAa,EAMd,MAAM,aAAa,CAAC;AAkCrB;;;GAGG;AACH,wBAAsB,eAAe,CACnC,KAAK,EAAE,aAAa,EACpB,EAAE,EAAE,aAAa,GAAG,MAAM,EAC1B,OAAO,GAAE,cAAmB,GAC3B,OAAO,CAAC,aAAa,CAAC,CAiDxB"}
|
|
@@ -1,7 +1,9 @@
|
|
|
1
1
|
import { DocumentError, wrapError } from '../utils/errors.js';
|
|
2
2
|
import { parseDocument } from '../parse/router.js';
|
|
3
|
-
import { emitCsvFromSheet, emitHtmlFromParagraphs,
|
|
3
|
+
import { emitCsvFromSheet, emitHtmlFromParagraphs, emitHtmlFromBlocks, emitMarkdownFromParagraphs, emitMarkdownFromBlocks, emitMarkdownFromText, emitHtmlFromText, emitText, emitTextFromSpreadsheet, chunkText, csvEscapeCell, blocksFromSpreadsheet, blocksFromCsvRows, } from './emit.js';
|
|
4
4
|
import { pptxToText } from '../parse/pptx.js';
|
|
5
|
+
import { isPathLike } from '../utils/paths.js';
|
|
6
|
+
import { readFile } from 'node:fs/promises';
|
|
5
7
|
const FORMAT_MAP = {
|
|
6
8
|
md: 'md',
|
|
7
9
|
markdown: 'md',
|
|
@@ -31,7 +33,22 @@ export async function convertDocument(input, to, options = {}) {
|
|
|
31
33
|
if (dest === 'pdf') {
|
|
32
34
|
throw new DocumentError('UNSUPPORTED_FORMAT', 'Convert to PDF is not supported in core. Use writeDocument for plain-text PDF creation.');
|
|
33
35
|
}
|
|
36
|
+
// Markdown → md: passthrough original source text
|
|
37
|
+
if (dest === 'md' && typeof input === 'string' && isPathLike(input)) {
|
|
38
|
+
const ext = input.toLowerCase();
|
|
39
|
+
if (ext.endsWith('.md') || ext.endsWith('.markdown')) {
|
|
40
|
+
const raw = (await readFile(input, 'utf8')).replace(/\r\n/g, '\n');
|
|
41
|
+
return { value: raw, messages: [], format: 'md' };
|
|
42
|
+
}
|
|
43
|
+
}
|
|
34
44
|
const parsed = await parseDocument(input, { output: 'structured' });
|
|
45
|
+
// MD structured may carry raw for passthrough when input was buffer/string content
|
|
46
|
+
if (dest === 'md' && parsed.format === 'md') {
|
|
47
|
+
const raw = parsed.structured?.raw;
|
|
48
|
+
if (typeof raw === 'string') {
|
|
49
|
+
return { value: raw.replace(/\r\n/g, '\n'), messages: parsed.warnings ?? [], format: 'md' };
|
|
50
|
+
}
|
|
51
|
+
}
|
|
35
52
|
return emitFromParsed(parsed, dest);
|
|
36
53
|
}
|
|
37
54
|
catch (err) {
|
|
@@ -47,6 +64,33 @@ function paragraphsOf(parsed) {
|
|
|
47
64
|
return paras;
|
|
48
65
|
return [];
|
|
49
66
|
}
|
|
67
|
+
function blocksOf(parsed) {
|
|
68
|
+
const blocks = parsed.structured
|
|
69
|
+
?.blocks;
|
|
70
|
+
return Array.isArray(blocks) ? blocks : undefined;
|
|
71
|
+
}
|
|
72
|
+
function emitRich(blocks, dest, messages, plainFallback) {
|
|
73
|
+
if (dest === 'md') {
|
|
74
|
+
return { value: emitMarkdownFromBlocks(blocks), messages, format: 'md' };
|
|
75
|
+
}
|
|
76
|
+
if (dest === 'html') {
|
|
77
|
+
return { value: emitHtmlFromBlocks(blocks), messages, format: 'html' };
|
|
78
|
+
}
|
|
79
|
+
if (dest === 'text') {
|
|
80
|
+
return { value: emitText(plainFallback), messages, format: 'text' };
|
|
81
|
+
}
|
|
82
|
+
if (dest === 'chunks') {
|
|
83
|
+
return { value: chunkText(plainFallback), messages, format: 'chunks' };
|
|
84
|
+
}
|
|
85
|
+
if (dest === 'csv') {
|
|
86
|
+
const csv = plainFallback
|
|
87
|
+
.split('\n')
|
|
88
|
+
.map((line) => csvEscapeCell(line))
|
|
89
|
+
.join('\n');
|
|
90
|
+
return { value: csv, messages, format: 'csv' };
|
|
91
|
+
}
|
|
92
|
+
throw new DocumentError('UNSUPPORTED_FORMAT', `Unsupported convert target: ${dest}`);
|
|
93
|
+
}
|
|
50
94
|
function emitFromParsed(parsed, dest) {
|
|
51
95
|
const messages = parsed.warnings ?? [];
|
|
52
96
|
const format = parsed.format;
|
|
@@ -55,34 +99,9 @@ function emitFromParsed(parsed, dest) {
|
|
|
55
99
|
if (dest === 'csv') {
|
|
56
100
|
return { value: emitCsvFromSheet(structured), messages, format: 'csv' };
|
|
57
101
|
}
|
|
58
|
-
|
|
59
|
-
|
|
60
|
-
|
|
61
|
-
messages,
|
|
62
|
-
format: 'md',
|
|
63
|
-
};
|
|
64
|
-
}
|
|
65
|
-
if (dest === 'text') {
|
|
66
|
-
return {
|
|
67
|
-
value: emitTextFromSpreadsheet(structured),
|
|
68
|
-
messages,
|
|
69
|
-
format: 'text',
|
|
70
|
-
};
|
|
71
|
-
}
|
|
72
|
-
if (dest === 'html') {
|
|
73
|
-
return {
|
|
74
|
-
value: emitHtmlFromText(emitTextFromSpreadsheet(structured)),
|
|
75
|
-
messages,
|
|
76
|
-
format: 'html',
|
|
77
|
-
};
|
|
78
|
-
}
|
|
79
|
-
if (dest === 'chunks') {
|
|
80
|
-
return {
|
|
81
|
-
value: chunkText(emitTextFromSpreadsheet(structured)),
|
|
82
|
-
messages,
|
|
83
|
-
format: 'chunks',
|
|
84
|
-
};
|
|
85
|
-
}
|
|
102
|
+
const blocks = blocksFromSpreadsheet(structured.sheets);
|
|
103
|
+
const plain = emitTextFromSpreadsheet(structured);
|
|
104
|
+
return emitRich(blocks, dest, messages, plain);
|
|
86
105
|
}
|
|
87
106
|
if (format === 'csv') {
|
|
88
107
|
const rows = parsed.structured?.rows ?? [];
|
|
@@ -90,56 +109,59 @@ function emitFromParsed(parsed, dest) {
|
|
|
90
109
|
if (dest === 'csv') {
|
|
91
110
|
return { value: csvText, messages, format: 'csv' };
|
|
92
111
|
}
|
|
112
|
+
const blocks = blocksFromCsvRows(rows);
|
|
113
|
+
if (dest === 'md' || dest === 'html') {
|
|
114
|
+
return emitRich(blocks, dest, messages, csvText);
|
|
115
|
+
}
|
|
93
116
|
return emitFromPlain(csvText, dest, messages);
|
|
94
117
|
}
|
|
95
|
-
if (format === 'docx'
|
|
118
|
+
if (format === 'docx') {
|
|
96
119
|
if (dest === 'csv') {
|
|
97
120
|
throw new DocumentError('UNSUPPORTED_FORMAT', 'Cannot convert DOCX/DOC to CSV; use a spreadsheet source.');
|
|
98
121
|
}
|
|
99
|
-
const
|
|
100
|
-
|
|
101
|
-
|
|
122
|
+
const structured = parsed.structured;
|
|
123
|
+
const blocks = structured.blocks ?? [];
|
|
124
|
+
const plain = (structured.paragraphs ?? []).join('\n\n');
|
|
125
|
+
if (blocks.length && (dest === 'md' || dest === 'html')) {
|
|
126
|
+
return emitRich(blocks, dest, messages, plain);
|
|
127
|
+
}
|
|
128
|
+
return emitFromParagraphs(structured.paragraphs ?? paragraphsOf(parsed), dest, messages);
|
|
129
|
+
}
|
|
130
|
+
if (format === 'doc') {
|
|
131
|
+
if (dest === 'csv') {
|
|
132
|
+
throw new DocumentError('UNSUPPORTED_FORMAT', 'Cannot convert DOCX/DOC to CSV; use a spreadsheet source.');
|
|
133
|
+
}
|
|
134
|
+
return emitFromParagraphs(paragraphsOf(parsed), dest, messages);
|
|
102
135
|
}
|
|
103
136
|
if (format === 'pptx') {
|
|
104
137
|
if (dest === 'csv') {
|
|
105
138
|
throw new DocumentError('UNSUPPORTED_FORMAT', 'Cannot convert PPTX to CSV; use md, html, text, or chunks.');
|
|
106
139
|
}
|
|
107
|
-
const
|
|
108
|
-
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
messages,
|
|
112
|
-
format: 'md',
|
|
113
|
-
};
|
|
114
|
-
}
|
|
115
|
-
if (dest === 'html') {
|
|
116
|
-
return {
|
|
117
|
-
value: emitHtmlFromPptx(slides),
|
|
118
|
-
messages,
|
|
119
|
-
format: 'html',
|
|
120
|
-
};
|
|
121
|
-
}
|
|
122
|
-
if (dest === 'text') {
|
|
123
|
-
return {
|
|
124
|
-
value: emitText(pptxToText({ format: 'pptx', slides })),
|
|
125
|
-
messages,
|
|
126
|
-
format: 'text',
|
|
127
|
-
};
|
|
140
|
+
const structured = parsed.structured;
|
|
141
|
+
const plain = pptxToText(structured);
|
|
142
|
+
if (structured.blocks?.length && (dest === 'md' || dest === 'html')) {
|
|
143
|
+
return emitRich(structured.blocks, dest, messages, plain);
|
|
128
144
|
}
|
|
129
|
-
|
|
130
|
-
|
|
131
|
-
|
|
132
|
-
|
|
133
|
-
|
|
134
|
-
|
|
145
|
+
return emitFromParagraphs(structured.slides.flatMap((s, i) => [`Slide ${i + 1}`, ...s]), dest, messages);
|
|
146
|
+
}
|
|
147
|
+
if (format === 'html' || format === 'md') {
|
|
148
|
+
const blocks = blocksOf(parsed);
|
|
149
|
+
const plain = paragraphsOf(parsed).join('\n\n') || parsed.text;
|
|
150
|
+
if (blocks?.length && (dest === 'md' || dest === 'html')) {
|
|
151
|
+
// md→md already handled via raw passthrough above when possible
|
|
152
|
+
if (format === 'md' && dest === 'md') {
|
|
153
|
+
const raw = parsed.structured?.raw;
|
|
154
|
+
if (raw)
|
|
155
|
+
return { value: raw.replace(/\r\n/g, '\n'), messages, format: 'md' };
|
|
156
|
+
}
|
|
157
|
+
return emitRich(blocks, dest, messages, plain);
|
|
135
158
|
}
|
|
136
159
|
}
|
|
137
|
-
// md / html / pdf / other paragraph-shaped structured
|
|
138
160
|
const paragraphs = paragraphsOf(parsed);
|
|
139
161
|
if (paragraphs.length > 0) {
|
|
140
162
|
return emitFromParagraphs(paragraphs, dest, messages);
|
|
141
163
|
}
|
|
142
|
-
return emitFromPlain('', dest, messages);
|
|
164
|
+
return emitFromPlain(parsed.text || '', dest, messages);
|
|
143
165
|
}
|
|
144
166
|
function emitFromParagraphs(paragraphs, dest, messages) {
|
|
145
167
|
if (dest === 'md') {
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"converter.js","sourceRoot":"","sources":["../../src/convert/converter.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;
|
|
1
|
+
{"version":3,"file":"converter.js","sourceRoot":"","sources":["../../src/convert/converter.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAY9D,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAC;AACnD,OAAO,EACL,gBAAgB,EAChB,sBAAsB,EACtB,kBAAkB,EAClB,0BAA0B,EAC1B,sBAAsB,EACtB,oBAAoB,EACpB,gBAAgB,EAChB,QAAQ,EACR,uBAAuB,EACvB,SAAS,EACT,aAAa,EACb,qBAAqB,EACrB,iBAAiB,GAClB,MAAM,WAAW,CAAC;AACnB,OAAO,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC9C,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AAC/C,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAE5C,MAAM,UAAU,GAA2C;IACzD,EAAE,EAAE,IAAI;IACR,QAAQ,EAAE,IAAI;IACd,IAAI,EAAE,MAAM;IACZ,GAAG,EAAE,KAAK;IACV,IAAI,EAAE,MAAM;IACZ,GAAG,EAAE,MAAM;IACX,MAAM,EAAE,QAAQ;IAChB,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,IAAI,EAAE,MAAM;CACb,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,eAAe,CACnC,KAAoB,EACpB,EAA0B,EAC1B,UAA0B,EAAE;IAE5B,KAAK,OAAO,CAAC;IACb,MAAM,IAAI,GAAG,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IAClD,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,MAAM,IAAI,aAAa,CACrB,oBAAoB,EACpB,+BAA+B,EAAE,EAAE,CACpC,CAAC;IACJ,CAAC;IAED,IAAI,CAAC;QACH,IAAI,IAAI,KAAK,KAAK,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;YACtC,MAAM,IAAI,aAAa,CACrB,oBAAoB,EACpB,mBAAmB,IAAI,yDAAyD,CACjF,CAAC;QACJ,CAAC;QAED,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,MAAM,IAAI,aAAa,CACrB,oBAAoB,EACpB,yFAAyF,CAC1F,CAAC;QACJ,CAAC;QAED,kDAAkD;QAClD,IAAI,IAAI,KAAK,IAAI,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YACpE,MAAM,GAAG,GAAG,KAAK,CAAC,WAAW,EAAE,CAAC;YAChC,IAAI,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,QAAQ,CAAC,WAAW,CAAC,EAAE,CAAC;gBACrD,MAAM,GAAG,GAAG,CAAC,MAAM,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC,OAAO,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;gBACnE,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,QAAQ,EAAE,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC;YACpD,CAAC;QACH,CAAC;QAED,MAAM,MAAM,GAAG,MAAM,aAAa,CAAC,KAAK,EAAE,EAAE,MAAM,EAAE,YAAY,EAAE,CAAC,CAAC;QAEpE,mFAAmF;QACnF,IAAI,IAAI,KAAK,IAAI,IAAI,MAAM,CAAC,MAAM,KAAK,IAAI,EAAE,CAAC;YAC5C,MAAM,GAAG,GAAI,MAAM,CAAC,UAA2C,EAAE,GAAG,CAAC;YACrE,IAAI,OAAO,GAAG,KAAK,QAAQ,EAAE,CAAC;gBAC5B,OAAO,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC;YAC9F,CAAC;QACH,CAAC;QAED,OAAO,cAAc,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IACtC,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,IAAI,GAAG,YAAY,aAAa;YAAE,MAAM,GAAG,CAAC;QAC5C,MAAM,SAAS,CAAC,gBAAgB,EAAE,wBAAwB,EAAE,EAAE,EAAE,GAAG,CAAC,CAAC;IACvE,CAAC;AACH,CAAC;AAED,SAAS,YAAY,CAAC,MAAsB;IAC1C,MAAM,KAAK,GAAI,MAAM,CAAC,UAAoD;QACxE,EAAE,UAAU,CAAC;IACf,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC;IACvC,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,SAAS,QAAQ,CAAC,MAAsB;IACtC,MAAM,MAAM,GAAI,MAAM,CAAC,UAAsD;QAC3E,EAAE,MAAM,CAAC;IACX,OAAO,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC;AACpD,CAAC;AAED,SAAS,QAAQ,CACf,MAAsB,EACtB,IAA4B,EAC5B,QAAmC,EACnC,aAAqB;IAErB,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;QAClB,OAAO,EAAE,KAAK,EAAE,sBAAsB,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC;IAC3E,CAAC;IACD,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;QACpB,OAAO,EAAE,KAAK,EAAE,kBAAkB,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;IACzE,CAAC;IACD,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;QACpB,OAAO,EAAE,KAAK,EAAE,QAAQ,CAAC,aAAa,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;IACtE,CAAC;IACD,IAAI,IAAI,KAAK,QAAQ,EAAE,CAAC;QACtB,OAAO,EAAE,KAAK,EAAE,SAAS,CAAC,aAAa,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAC;IACzE,CAAC;IACD,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;QACnB,MAAM,GAAG,GAAG,aAAa;aACtB,KAAK,CAAC,IAAI,CAAC;aACX,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC;aAClC,IAAI,CAAC,IAAI,CAAC,CAAC;QACd,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;IACjD,CAAC;IACD,MAAM,IAAI,aAAa,CAAC,oBAAoB,EAAE,+BAA+B,IAAI,EAAE,CAAC,CAAC;AACvF,CAAC;AAED,SAAS,cAAc,CACrB,MAAsB,EACtB,IAA4B;IAE5B,MAAM,QAAQ,GAAG,MAAM,CAAC,QAAQ,IAAI,EAAE,CAAC;IACvC,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;IAE7B,IAAI,MAAM,KAAK,MAAM,IAAI,MAAM,KAAK,KAAK,EAAE,CAAC;QAC1C,MAAM,UAAU,GAAG,MAAM,CAAC,UAAmC,CAAC;QAC9D,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,OAAO,EAAE,KAAK,EAAE,gBAAgB,CAAC,UAAU,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;QAC1E,CAAC;QACD,MAAM,MAAM,GAAG,qBAAqB,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC;QACxD,MAAM,KAAK,GAAG,uBAAuB,CAAC,UAAU,CAAC,CAAC;QAClD,OAAO,QAAQ,CAAC,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,CAAC,CAAC;IACjD,CAAC;IAED,IAAI,MAAM,KAAK,KAAK,EAAE,CAAC;QACrB,MAAM,IAAI,GACP,MAAM,CAAC,UAAgD,EAAE,IAAI,IAAI,EAAE,CAAC;QACvE,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC/E,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;QACrD,CAAC;QACD,MAAM,MAAM,GAAG,iBAAiB,CAAC,IAAI,CAAC,CAAC;QACvC,IAAI,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;YACrC,OAAO,QAAQ,CAAC,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;QACnD,CAAC;QACD,OAAO,aAAa,CAAC,OAAO,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;IAChD,CAAC;IAED,IAAI,MAAM,KAAK,MAAM,EAAE,CAAC;QACtB,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,MAAM,IAAI,aAAa,CACrB,oBAAoB,EACpB,2DAA2D,CAC5D,CAAC;QACJ,CAAC;QACD,MAAM,UAAU,GAAG,MAAM,CAAC,UAA4B,CAAC;QACvD,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,IAAI,EAAE,CAAC;QACvC,MAAM,KAAK,GAAG,CAAC,UAAU,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACzD,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,MAAM,CAAC,EAAE,CAAC;YACxD,OAAO,QAAQ,CAAC,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,CAAC,CAAC;QACjD,CAAC;QACD,OAAO,kBAAkB,CAAC,UAAU,CAAC,UAAU,IAAI,YAAY,CAAC,MAAM,CAAC,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;IAC3F,CAAC;IAED,IAAI,MAAM,KAAK,KAAK,EAAE,CAAC;QACrB,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,MAAM,IAAI,aAAa,CACrB,oBAAoB,EACpB,2DAA2D,CAC5D,CAAC;QACJ,CAAC;QACD,OAAO,kBAAkB,CAAC,YAAY,CAAC,MAAM,CAAC,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;IAClE,CAAC;IAED,IAAI,MAAM,KAAK,MAAM,EAAE,CAAC;QACtB,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,MAAM,IAAI,aAAa,CACrB,oBAAoB,EACpB,4DAA4D,CAC7D,CAAC;QACJ,CAAC;QACD,MAAM,UAAU,GAAG,MAAM,CAAC,UAA4B,CAAC;QACvD,MAAM,KAAK,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;QACrC,IAAI,UAAU,CAAC,MAAM,EAAE,MAAM,IAAI,CAAC,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,MAAM,CAAC,EAAE,CAAC;YACpE,OAAO,QAAQ,CAAC,UAAU,CAAC,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,CAAC,CAAC;QAC5D,CAAC;QACD,OAAO,kBAAkB,CACvB,UAAU,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,EAC7D,IAAI,EACJ,QAAQ,CACT,CAAC;IACJ,CAAC;IAED,IAAI,MAAM,KAAK,MAAM,IAAI,MAAM,KAAK,IAAI,EAAE,CAAC;QACzC,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC;QAChC,MAAM,KAAK,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC;QAC/D,IAAI,MAAM,EAAE,MAAM,IAAI,CAAC,IAAI,KAAK,IAAI,IAAI,IAAI,KAAK,MAAM,CAAC,EAAE,CAAC;YACzD,gEAAgE;YAChE,IAAI,MAAM,KAAK,IAAI,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;gBACrC,MAAM,GAAG,GAAI,MAAM,CAAC,UAA+B,EAAE,GAAG,CAAC;gBACzD,IAAI,GAAG;oBAAE,OAAO,EAAE,KAAK,EAAE,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC;YAChF,CAAC;YACD,OAAO,QAAQ,CAAC,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,CAAC,CAAC;QACjD,CAAC;IACH,CAAC;IAED,MAAM,UAAU,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC;IACxC,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC1B,OAAO,kBAAkB,CAAC,UAAU,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;IACxD,CAAC;IAED,OAAO,aAAa,CAAC,MAAM,CAAC,IAAI,IAAI,EAAE,EAAE,IAAI,EAAE,QAAQ,CAAC,CAAC;AAC1D,CAAC;AAED,SAAS,kBAAkB,CACzB,UAAoB,EACpB,IAA4B,EAC5B,QAAmC;IAEnC,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;QAClB,OAAO;YACL,KAAK,EAAE,0BAA0B,CAAC,UAAU,CAAC;YAC7C,QAAQ;YACR,MAAM,EAAE,IAAI;SACb,CAAC;IACJ,CAAC;IACD,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;QACpB,OAAO;YACL,KAAK,EAAE,sBAAsB,CAAC,UAAU,CAAC;YACzC,QAAQ;YACR,MAAM,EAAE,MAAM;SACf,CAAC;IACJ,CAAC;IACD,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;QACpB,OAAO;YACL,KAAK,EAAE,QAAQ,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACtC,QAAQ;YACR,MAAM,EAAE,MAAM;SACf,CAAC;IACJ,CAAC;IACD,IAAI,IAAI,KAAK,QAAQ,EAAE,CAAC;QACtB,OAAO;YACL,KAAK,EAAE,SAAS,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACzC,QAAQ;YACR,MAAM,EAAE,QAAQ;SACjB,CAAC;IACJ,CAAC;IACD,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;QACnB,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACrE,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;IACjD,CAAC;IACD,MAAM,IAAI,aAAa,CAAC,oBAAoB,EAAE,+BAA+B,IAAI,EAAE,CAAC,CAAC;AACvF,CAAC;AAED,SAAS,aAAa,CACpB,IAAY,EACZ,IAA4B,EAC5B,QAAmC;IAEnC,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;QACpB,OAAO,EAAE,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;IAC7D,CAAC;IACD,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;QAClB,OAAO,EAAE,KAAK,EAAE,oBAAoB,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC;IACvE,CAAC;IACD,IAAI,IAAI,KAAK,MAAM,EAAE,CAAC;QACpB,OAAO,EAAE,KAAK,EAAE,gBAAgB,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;IACrE,CAAC;IACD,IAAI,IAAI,KAAK,QAAQ,EAAE,CAAC;QACtB,OAAO,EAAE,KAAK,EAAE,SAAS,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAC;IAChE,CAAC;IACD,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;QACnB,MAAM,GAAG,GAAG,IAAI;aACb,KAAK,CAAC,IAAI,CAAC;aACX,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC;aAClC,IAAI,CAAC,IAAI,CAAC,CAAC;QACd,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;IACjD,CAAC;IACD,MAAM,IAAI,aAAa,CAAC,oBAAoB,EAAE,+BAA+B,IAAI,EAAE,CAAC,CAAC;AACvF,CAAC"}
|
package/dist/convert/emit.d.ts
CHANGED
|
@@ -1,5 +1,6 @@
|
|
|
1
1
|
import type { SpreadsheetStructured } from '../types.js';
|
|
2
2
|
export { csvEscapeCell, escapeHtml } from '../utils/text.js';
|
|
3
|
+
export { emitMarkdownFromBlocks, emitHtmlFromBlocks, emitMarkdownTable, blocksFromSpreadsheet, blocksFromCsvRows, flattenBlocksToParagraphs, } from './blocks.js';
|
|
3
4
|
/** Emit markdown / html / text / csv from already-extracted content. */
|
|
4
5
|
export declare function emitText(text: string): string;
|
|
5
6
|
export declare function emitMarkdownFromParagraphs(paragraphs: string[]): string;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"emit.d.ts","sourceRoot":"","sources":["../../src/convert/emit.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,
|
|
1
|
+
{"version":3,"file":"emit.d.ts","sourceRoot":"","sources":["../../src/convert/emit.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAgB,qBAAqB,EAAE,MAAM,aAAa,CAAC;AASvE,OAAO,EAAE,aAAa,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC7D,OAAO,EACL,sBAAsB,EACtB,kBAAkB,EAClB,iBAAiB,EACjB,qBAAqB,EACrB,iBAAiB,EACjB,yBAAyB,GAC1B,MAAM,aAAa,CAAC;AAErB,wEAAwE;AACxE,wBAAgB,QAAQ,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAE7C;AAED,wBAAgB,0BAA0B,CAAC,UAAU,EAAE,MAAM,EAAE,GAAG,MAAM,CAEvE;AAED,wBAAgB,sBAAsB,CAAC,UAAU,EAAE,MAAM,EAAE,GAAG,MAAM,CAOnE;AAED,wBAAgB,oBAAoB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CAEzD;AAED,wBAAgB,gBAAgB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CASrD;AAED,wBAAgB,gBAAgB,CAC9B,UAAU,EAAE,qBAAqB,EACjC,SAAS,CAAC,EAAE,MAAM,GACjB,MAAM,CAMR;AAED,wBAAgB,2BAA2B,CAAC,UAAU,EAAE,qBAAqB,GAAG,MAAM,CAErF;AAED,wBAAgB,oBAAoB,CAAC,MAAM,EAAE,MAAM,EAAE,EAAE,GAAG,MAAM,CAiB/D;AAED,wBAAgB,gBAAgB,CAAC,MAAM,EAAE,MAAM,EAAE,EAAE,GAAG,MAAM,CAa3D;AAED,wBAAgB,uBAAuB,CAAC,UAAU,EAAE,qBAAqB,GAAG,MAAM,CAEjF;AAED,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,SAAO,GAAG,MAAM,EAAE,CAWjE"}
|