@xberg-io/xberg-wasm 1.0.10 → 1.0.12

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (8) hide show
  1. package/README.md +76 -44
  2. package/package.json +1 -1
  3. package/pkg/web/README.md +76 -44
  4. package/pkg/web/package.json +1 -1
  5. package/pkg/web/xberg_wasm.d.ts +81 -81
  6. package/pkg/web/xberg_wasm.js +1 -1
  7. package/pkg/web/xberg_wasm_bg.wasm +0 -0
  8. package/pkg/web/xberg_wasm_bg.wasm.d.ts +81 -81
package/README.md CHANGED
@@ -100,6 +100,7 @@ pnpm add @xberg-io/xberg-wasm
100
100
  ```
101
101
 
102
102
  ### System Requirements
103
+
103
104
  - Modern browser with WebAssembly support, or Deno 1.0+, or Cloudflare Workers
104
105
  - Optional: [Tesseract WASM](https://github.com/naptha/tesseract.js) for OCR functionality
105
106
 
@@ -261,53 +262,84 @@ extractDocuments(fileBytes, mimes)
261
262
 
262
263
  ## Features
263
264
 
264
- ### Supported File Formats (101 formats · 115 file extensions) 101 formats across 115 file extensions in 8 major
265
- categories with intelligent format detection and comprehensive metadata extraction. #### Office Documents | Category |
266
- Formats | Capabilities | |----------|---------|--------------| | **Word Processing** | `.docx`, `.docm`, `.doc`,
267
- `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles
268
- | | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers`
269
- | Sheet data, formulas, cell metadata, charts | | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`,
270
- `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata | | **PDF** | `.pdf` | Text, tables, images,
271
- metadata, OCR support | | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources | | **Database** |
272
- `.dbf` | Table data extraction, field type support | | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text
273
- extraction | #### Images (OCR-Enabled) | Category | Formats | Features | |----------|---------|----------| | **Raster**
274
- | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions,
275
- color space | | **Advanced** | `.jp2`, `.jpx`, `.jpm`, `.mj2`, `.jbig2`, `.jb2`, `.pnm`, `.pbm`, `.pgm`, `.ppm` | OCR
276
- via hayro-jpeg2000 (pure Rust decoder), JBIG2 support, table detection, format-specific metadata |
265
+ ### Supported File Formats (101 formats · 115 file extensions)
266
+
267
+ 101 formats across 115 file extensions in 8 major categories with intelligent format detection and comprehensive metadata extraction.
268
+
269
+ #### Office Documents
270
+
271
+ | Category | Formats | Capabilities |
272
+ |----------|---------|--------------|
273
+ | **Word Processing** | `.docx`, `.docm`, `.doc`, `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles |
274
+ | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers` | Sheet data, formulas, cell metadata, charts |
275
+ | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`, `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata |
276
+ | **PDF** | `.pdf` | Text, tables, images, metadata, OCR support |
277
+ | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources |
278
+ | **Database** | `.dbf` | Table data extraction, field type support |
279
+ | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text extraction |
280
+
281
+ #### Images (OCR-Enabled)
282
+
283
+ | Category | Formats | Features |
284
+ |----------|---------|----------|
285
+ | **Raster** | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions, color space |
286
+ | **Advanced** | `.jp2`, `.jpx`, `.jpm`, `.mj2`, `.jbig2`, `.jb2`, `.pnm`, `.pbm`, `.pgm`, `.ppm` | OCR via hayro-jpeg2000 (pure Rust decoder), JBIG2 support, table detection, format-specific metadata |
277
287
  | **Vector** | `.svg` | DOM parsing, embedded text, graphics metadata |
278
288
 
279
- #### Web & Data | Category | Formats | Features | |----------|---------|----------| | **Markup** | `.html`, `.htm`,
280
- `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction | | **Structured Data** |
281
- `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation | | **Text &
282
- Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX,
283
- reStructuredText, Org Mode | #### Email & Archives | Category | Formats | Features | |----------|---------|----------| |
284
- **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading | | **Archives** | `.zip`,
285
- `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
286
- #### Academic & Scientific | Category | Formats | Features | |----------|---------|----------| | **Citations** | `.bib`,
287
- `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
288
- | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS
289
- | | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML
290
- outlines | | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats | #### Code Intelligence
291
- (371 Languages) | Feature | Description | |---------|-------------| | **Structure Extraction** | Functions, classes,
292
- methods, structs, interfaces, enums | | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
293
- | **Symbol Extraction** | Variables, constants, type aliases, properties | | **Docstring Parsing** | Google, NumPy,
294
- Sphinx, JSDoc, RustDoc, and 10+ formats | | **Diagnostics** | Parse errors with line/column positions | | **Syntax-Aware
295
- Chunking** | Split code by semantic boundaries, not arbitrary byte offsets | Powered by
296
- [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) —
297
- [documentation](https://docs.tree-sitter-language-pack.xberg.io). **[Complete Format
298
- Reference](https://docs.xberg.io/reference/formats/)** ### Key Capabilities - **Text Extraction** - Extract all text
299
- content with position and formatting information - **Metadata Extraction** - Retrieve document properties, creation
300
- date, author, etc. - **Table Extraction** - Parse tables with structure and cell content preservation - **Image
301
- Extraction** - Extract embedded images and render page previews
289
+ #### Web & Data
290
+
291
+ | Category | Formats | Features |
292
+ |----------|---------|----------|
293
+ | **Markup** | `.html`, `.htm`, `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction |
294
+ | **Structured Data** | `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation |
295
+ | **Text & Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX, reStructuredText, Org Mode |
296
+
297
+ #### Email & Archives
298
+
299
+ | Category | Formats | Features |
300
+ |----------|---------|----------|
301
+ | **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading |
302
+ | **Archives** | `.zip`, `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
303
+
304
+ #### Academic & Scientific
305
+
306
+ | Category | Formats | Features |
307
+ |----------|---------|----------|
308
+ | **Citations** | `.bib`, `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
309
+ | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS |
310
+ | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML outlines |
311
+ | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats |
312
+
313
+ #### Code Intelligence (371 Languages)
314
+
315
+ | Feature | Description |
316
+ |---------|-------------|
317
+ | **Structure Extraction** | Functions, classes, methods, structs, interfaces, enums |
318
+ | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
319
+ | **Symbol Extraction** | Variables, constants, type aliases, properties |
320
+ | **Docstring Parsing** | Google, NumPy, Sphinx, JSDoc, RustDoc, and 10+ formats |
321
+ | **Diagnostics** | Parse errors with line/column positions |
322
+ | **Syntax-Aware Chunking** | Split code by semantic boundaries, not arbitrary byte offsets |
323
+
324
+ Powered by [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) — [documentation](https://docs.tree-sitter-language-pack.xberg.io).
325
+
326
+ **[Complete Format Reference](https://docs.xberg.io/reference/formats/)**
327
+
328
+ ### Key Capabilities
329
+
330
+ - **Text Extraction** - Extract all text content with position and formatting information
331
+ - **Metadata Extraction** - Retrieve document properties, creation date, author, etc.
332
+ - **Table Extraction** - Parse tables with structure and cell content preservation
333
+ - **Image Extraction** - Extract embedded images and render page previews
302
334
  - **OCR Support** - Integrate multiple OCR backends for scanned documents
303
- - **Async/Await** - Non-blocking document processing with concurrent operations
304
- - **Plugin System** - Extensible post-processing for custom text transformation
305
- - **Batch Processing** - Efficiently process multiple documents in parallel - **Memory Efficient** - Stream large files
306
- without loading entirely into memory - **Language Detection** - Detect and support multiple languages in documents
307
- - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [371 programming
308
- languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
309
- - **Configuration** - Fine-grained control over extraction behavior - **Six Output Formats** - Plain text, Markdown,
310
- Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
335
+ - **Async/Await** - Non-blocking document processing with concurrent operations
336
+ - **Plugin System** - Extensible post-processing for custom text transformation
337
+ - **Batch Processing** - Efficiently process multiple documents in parallel
338
+ - **Memory Efficient** - Stream large files without loading entirely into memory
339
+ - **Language Detection** - Detect and support multiple languages in documents
340
+ - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [371 programming languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
341
+ - **Configuration** - Fine-grained control over extraction behavior
342
+ - **Six Output Formats** - Plain text, Markdown, Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
311
343
 
312
344
  ## OCR Support
313
345
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@xberg-io/xberg-wasm",
3
- "version": "1.0.10",
3
+ "version": "1.0.12",
4
4
  "private": false,
5
5
  "description": "High-performance document intelligence library",
6
6
  "license": "MIT",
package/pkg/web/README.md CHANGED
@@ -100,6 +100,7 @@ pnpm add @xberg-io/xberg-wasm
100
100
  ```
101
101
 
102
102
  ### System Requirements
103
+
103
104
  - Modern browser with WebAssembly support, or Deno 1.0+, or Cloudflare Workers
104
105
  - Optional: [Tesseract WASM](https://github.com/naptha/tesseract.js) for OCR functionality
105
106
 
@@ -261,53 +262,84 @@ extractDocuments(fileBytes, mimes)
261
262
 
262
263
  ## Features
263
264
 
264
- ### Supported File Formats (101 formats · 115 file extensions) 101 formats across 115 file extensions in 8 major
265
- categories with intelligent format detection and comprehensive metadata extraction. #### Office Documents | Category |
266
- Formats | Capabilities | |----------|---------|--------------| | **Word Processing** | `.docx`, `.docm`, `.doc`,
267
- `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles
268
- | | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers`
269
- | Sheet data, formulas, cell metadata, charts | | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`,
270
- `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata | | **PDF** | `.pdf` | Text, tables, images,
271
- metadata, OCR support | | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources | | **Database** |
272
- `.dbf` | Table data extraction, field type support | | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text
273
- extraction | #### Images (OCR-Enabled) | Category | Formats | Features | |----------|---------|----------| | **Raster**
274
- | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions,
275
- color space | | **Advanced** | `.jp2`, `.jpx`, `.jpm`, `.mj2`, `.jbig2`, `.jb2`, `.pnm`, `.pbm`, `.pgm`, `.ppm` | OCR
276
- via hayro-jpeg2000 (pure Rust decoder), JBIG2 support, table detection, format-specific metadata |
265
+ ### Supported File Formats (101 formats · 115 file extensions)
266
+
267
+ 101 formats across 115 file extensions in 8 major categories with intelligent format detection and comprehensive metadata extraction.
268
+
269
+ #### Office Documents
270
+
271
+ | Category | Formats | Capabilities |
272
+ |----------|---------|--------------|
273
+ | **Word Processing** | `.docx`, `.docm`, `.doc`, `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles |
274
+ | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers` | Sheet data, formulas, cell metadata, charts |
275
+ | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`, `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata |
276
+ | **PDF** | `.pdf` | Text, tables, images, metadata, OCR support |
277
+ | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources |
278
+ | **Database** | `.dbf` | Table data extraction, field type support |
279
+ | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text extraction |
280
+
281
+ #### Images (OCR-Enabled)
282
+
283
+ | Category | Formats | Features |
284
+ |----------|---------|----------|
285
+ | **Raster** | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions, color space |
286
+ | **Advanced** | `.jp2`, `.jpx`, `.jpm`, `.mj2`, `.jbig2`, `.jb2`, `.pnm`, `.pbm`, `.pgm`, `.ppm` | OCR via hayro-jpeg2000 (pure Rust decoder), JBIG2 support, table detection, format-specific metadata |
277
287
  | **Vector** | `.svg` | DOM parsing, embedded text, graphics metadata |
278
288
 
279
- #### Web & Data | Category | Formats | Features | |----------|---------|----------| | **Markup** | `.html`, `.htm`,
280
- `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction | | **Structured Data** |
281
- `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation | | **Text &
282
- Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX,
283
- reStructuredText, Org Mode | #### Email & Archives | Category | Formats | Features | |----------|---------|----------| |
284
- **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading | | **Archives** | `.zip`,
285
- `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
286
- #### Academic & Scientific | Category | Formats | Features | |----------|---------|----------| | **Citations** | `.bib`,
287
- `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
288
- | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS
289
- | | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML
290
- outlines | | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats | #### Code Intelligence
291
- (371 Languages) | Feature | Description | |---------|-------------| | **Structure Extraction** | Functions, classes,
292
- methods, structs, interfaces, enums | | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
293
- | **Symbol Extraction** | Variables, constants, type aliases, properties | | **Docstring Parsing** | Google, NumPy,
294
- Sphinx, JSDoc, RustDoc, and 10+ formats | | **Diagnostics** | Parse errors with line/column positions | | **Syntax-Aware
295
- Chunking** | Split code by semantic boundaries, not arbitrary byte offsets | Powered by
296
- [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) —
297
- [documentation](https://docs.tree-sitter-language-pack.xberg.io). **[Complete Format
298
- Reference](https://docs.xberg.io/reference/formats/)** ### Key Capabilities - **Text Extraction** - Extract all text
299
- content with position and formatting information - **Metadata Extraction** - Retrieve document properties, creation
300
- date, author, etc. - **Table Extraction** - Parse tables with structure and cell content preservation - **Image
301
- Extraction** - Extract embedded images and render page previews
289
+ #### Web & Data
290
+
291
+ | Category | Formats | Features |
292
+ |----------|---------|----------|
293
+ | **Markup** | `.html`, `.htm`, `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction |
294
+ | **Structured Data** | `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation |
295
+ | **Text & Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX, reStructuredText, Org Mode |
296
+
297
+ #### Email & Archives
298
+
299
+ | Category | Formats | Features |
300
+ |----------|---------|----------|
301
+ | **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading |
302
+ | **Archives** | `.zip`, `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
303
+
304
+ #### Academic & Scientific
305
+
306
+ | Category | Formats | Features |
307
+ |----------|---------|----------|
308
+ | **Citations** | `.bib`, `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
309
+ | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS |
310
+ | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML outlines |
311
+ | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats |
312
+
313
+ #### Code Intelligence (371 Languages)
314
+
315
+ | Feature | Description |
316
+ |---------|-------------|
317
+ | **Structure Extraction** | Functions, classes, methods, structs, interfaces, enums |
318
+ | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
319
+ | **Symbol Extraction** | Variables, constants, type aliases, properties |
320
+ | **Docstring Parsing** | Google, NumPy, Sphinx, JSDoc, RustDoc, and 10+ formats |
321
+ | **Diagnostics** | Parse errors with line/column positions |
322
+ | **Syntax-Aware Chunking** | Split code by semantic boundaries, not arbitrary byte offsets |
323
+
324
+ Powered by [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) — [documentation](https://docs.tree-sitter-language-pack.xberg.io).
325
+
326
+ **[Complete Format Reference](https://docs.xberg.io/reference/formats/)**
327
+
328
+ ### Key Capabilities
329
+
330
+ - **Text Extraction** - Extract all text content with position and formatting information
331
+ - **Metadata Extraction** - Retrieve document properties, creation date, author, etc.
332
+ - **Table Extraction** - Parse tables with structure and cell content preservation
333
+ - **Image Extraction** - Extract embedded images and render page previews
302
334
  - **OCR Support** - Integrate multiple OCR backends for scanned documents
303
- - **Async/Await** - Non-blocking document processing with concurrent operations
304
- - **Plugin System** - Extensible post-processing for custom text transformation
305
- - **Batch Processing** - Efficiently process multiple documents in parallel - **Memory Efficient** - Stream large files
306
- without loading entirely into memory - **Language Detection** - Detect and support multiple languages in documents
307
- - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [371 programming
308
- languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
309
- - **Configuration** - Fine-grained control over extraction behavior - **Six Output Formats** - Plain text, Markdown,
310
- Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
335
+ - **Async/Await** - Non-blocking document processing with concurrent operations
336
+ - **Plugin System** - Extensible post-processing for custom text transformation
337
+ - **Batch Processing** - Efficiently process multiple documents in parallel
338
+ - **Memory Efficient** - Stream large files without loading entirely into memory
339
+ - **Language Detection** - Detect and support multiple languages in documents
340
+ - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [371 programming languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
341
+ - **Configuration** - Fine-grained control over extraction behavior
342
+ - **Six Output Formats** - Plain text, Markdown, Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
311
343
 
312
344
  ## OCR Support
313
345
 
package/pkg/web/package.json CHANGED
@@ -2,7 +2,7 @@
2
2
  "name": "xberg-wasm",
3
3
  "type": "module",
4
4
  "description": "High-performance document intelligence library",
5
- "version": "1.0.10",
5
+ "version": "1.0.12",
6
6
  "license": "MIT",
7
7
  "repository": {
8
8
  "type": "git",
package/pkg/web/xberg_wasm.d.ts CHANGED
@@ -7039,9 +7039,9 @@ export interface InitOutput {
7039
7039
  readonly wasmprocessingwarning_new: (a: number, b: number, c: number, d: number) => number;
7040
7040
  readonly wasmpagehierarchy_default: () => number;
7041
7041
  readonly wasmqrboundingbox_default: () => number;
7042
- readonly __wbg_wasmtablecell_free: (a: number, b: number) => void;
7043
- readonly __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
7044
7042
  readonly __wbg_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
7043
+ readonly __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
7044
+ readonly __wbg_wasmtablecell_free: (a: number, b: number) => void;
7045
7045
  readonly wasmvlmfallbackpolicy_default: () => number;
7046
7046
  readonly wasmproxyconfig_url: (a: number) => [number, number];
7047
7047
  readonly wasmsupportedformat_extension: (a: number) => [number, number];
@@ -7072,38 +7072,88 @@ export interface InitOutput {
7072
7072
  readonly wasmpatternmatch_set_start: (a: number, b: number) => void;
7073
7073
  readonly wasmpatternmatch_set_text: (a: number, b: number, c: number) => void;
7074
7074
  readonly wasmerrormetadata_errorType: (a: number) => [number, number];
7075
+ readonly wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
7075
7076
  readonly wasmpageboundary_set_pageNumber: (a: number, b: number) => void;
7076
7077
  readonly wasmpageboundary_set_byteStart: (a: number, b: number) => void;
7077
7078
  readonly wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
7078
7079
  readonly wasmpageboundary_byteStart: (a: number) => number;
7079
- readonly wasmocrelementconfig_minConfidence: (a: number) => number;
7080
- readonly wasmocrelementconfig_includeElements: (a: number) => number;
7080
+ readonly wasmpageboundary_byteEnd: (a: number) => number;
7081
+ readonly __wbg_wasmpageboundary_free: (a: number, b: number) => void;
7082
+ readonly wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
7083
+ readonly wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
7084
+ readonly wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
7085
+ readonly wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
7086
+ readonly wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
7087
+ readonly wasmdocumentrelationship_target: (a: number) => number;
7088
+ readonly wasmpageboundary_pageNumber: (a: number) => number;
7089
+ readonly __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
7090
+ readonly wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
7081
7091
  readonly wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
7082
7092
  readonly wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
7093
+ readonly wasmredactionterm_value: (a: number) => [number, number];
7083
7094
  readonly wasmredactionterm_label: (a: number) => [number, number];
7084
7095
  readonly wasmredactionterm_caseSensitive: (a: number) => number;
7096
+ readonly __wbg_wasmredactionterm_free: (a: number, b: number) => void;
7097
+ readonly wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
7098
+ readonly wasmocrrotation_angleDegrees: (a: number) => number;
7099
+ readonly wasmocrrotation_confidence: (a: number) => [number, number];
7100
+ readonly __wbg_wasmocrrotation_free: (a: number, b: number) => void;
7101
+ readonly wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
7102
+ readonly wasmdbffieldinfo_name: (a: number) => [number, number];
7103
+ readonly wasmdbffieldinfo_fieldType: (a: number) => [number, number];
7104
+ readonly wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
7105
+ readonly wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
7106
+ readonly wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
7107
+ readonly wasmdjotlink_url: (a: number) => [number, number];
7108
+ readonly wasmdjotlink_title: (a: number) => [number, number];
7109
+ readonly wasmdjotlink_text: (a: number) => [number, number];
7110
+ readonly __wbg_wasmdjotlink_free: (a: number, b: number) => void;
7111
+ readonly wasmocrelementconfig_minConfidence: (a: number) => number;
7112
+ readonly wasmocrelementconfig_includeElements: (a: number) => number;
7113
+ readonly __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
7085
7114
  readonly wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
7086
7115
  readonly wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
7087
7116
  readonly wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
7088
7117
  readonly wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
7089
7118
  readonly wasmpageconfig_markerFormat: (a: number) => [number, number];
7090
7119
  readonly wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
7091
- readonly __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
7120
+ readonly wasmredactionpattern_pattern: (a: number) => [number, number];
7121
+ readonly __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
7122
+ readonly wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
7123
+ readonly wasmprocessingwarning_message: (a: number) => [number, number];
7124
+ readonly __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
7125
+ readonly wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
7126
+ readonly wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
7127
+ readonly wasmerrormetadata_message: (a: number) => [number, number];
7128
+ readonly __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
7129
+ readonly __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
7130
+ readonly wasmpreprocessingoptions_enabled: (a: number) => number;
7131
+ readonly __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
7132
+ readonly wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
7133
+ readonly wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
7134
+ readonly wasmproxyconfig_password: (a: number) => [number, number];
7135
+ readonly wasmproxyconfig_username: (a: number) => [number, number];
7136
+ readonly __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
7137
+ readonly wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
7138
+ readonly wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
7139
+ readonly wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
7140
+ readonly wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
7141
+ readonly wasmsparseembeddingconfig_model: (a: number) => any;
7142
+ readonly wasmsparseembeddingconfig_acceleration: (a: number) => number;
7092
7143
  readonly wasmrerankerconfig_set_topK: (a: number, b: number) => void;
7093
7144
  readonly wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
7094
7145
  readonly wasmrerankerconfig_set_model: (a: number, b: any) => void;
7146
+ readonly wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
7147
+ readonly wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7095
7148
  readonly wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
7096
7149
  readonly wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
7097
7150
  readonly wasmrerankerconfig_topK: (a: number) => number;
7098
7151
  readonly wasmrerankerconfig_showDownloadProgress: (a: number) => number;
7099
7152
  readonly wasmrerankerconfig_model: (a: number) => any;
7153
+ readonly wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
7154
+ readonly wasmrerankerconfig_cacheDir: (a: number) => [number, number];
7155
+ readonly wasmrerankerconfig_batchSize: (a: number) => number;
7100
7156
  readonly wasmrerankerconfig_acceleration: (a: number) => number;
7101
- readonly wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
7102
- readonly wasmssrfpolicy_denyPrivate: (a: number) => number;
7103
- readonly wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
7104
- readonly wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
7105
- readonly wasmredactionfinding_start: (a: number) => number;
7106
- readonly __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
7107
7157
  readonly wasmqrboundingbox_set_y: (a: number, b: number) => void;
7108
7158
  readonly wasmqrboundingbox_set_x: (a: number, b: number) => void;
7109
7159
  readonly wasmqrboundingbox_set_width: (a: number, b: number) => void;
@@ -7113,8 +7163,17 @@ export interface InitOutput {
7113
7163
  readonly wasmqrboundingbox_width: (a: number) => number;
7114
7164
  readonly wasmqrboundingbox_height: (a: number) => number;
7115
7165
  readonly __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
7116
- readonly __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
7117
- readonly wasmpreprocessingoptions_enabled: (a: number) => number;
7166
+ readonly wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
7167
+ readonly wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7168
+ readonly wasmsparseembeddingconfig_maxLength: (a: number) => number;
7169
+ readonly wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
7170
+ readonly wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
7171
+ readonly wasmsparseembeddingconfig_batchSize: (a: number) => number;
7172
+ readonly wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
7173
+ readonly wasmdiffline_set_0: (a: number, b: number, c: number) => void;
7174
+ readonly wasmdiffline_kind: (a: number) => [number, number];
7175
+ readonly wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
7176
+ readonly wasmssrfpolicy_denyPrivate: (a: number) => number;
7118
7177
  readonly wasmtablecell_set_rowSpan: (a: number, b: number) => void;
7119
7178
  readonly wasmtablecell_set_isHeader: (a: number, b: number) => void;
7120
7179
  readonly wasmtablecell_set_content: (a: number, b: number, c: number) => void;
@@ -7123,78 +7182,19 @@ export interface InitOutput {
7123
7182
  readonly wasmtablecell_isHeader: (a: number) => number;
7124
7183
  readonly wasmtablecell_content: (a: number) => [number, number];
7125
7184
  readonly wasmtablecell_colSpan: (a: number) => number;
7126
- readonly __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
7127
- readonly __wbg_wasmpageconfig_free: (a: number, b: number) => void;
7128
- readonly wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
7129
- readonly wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
7130
- readonly wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
7131
- readonly wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7132
- readonly wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
7133
- readonly wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
7134
- readonly wasmsparseembeddingconfig_model: (a: number) => any;
7135
- readonly wasmsparseembeddingconfig_maxLength: (a: number) => number;
7136
- readonly wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
7137
- readonly wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
7138
- readonly wasmsparseembeddingconfig_batchSize: (a: number) => number;
7139
- readonly wasmsparseembeddingconfig_acceleration: (a: number) => number;
7140
- readonly wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
7141
- readonly wasmrerankerconfig_cacheDir: (a: number) => [number, number];
7142
- readonly wasmrerankerconfig_batchSize: (a: number) => number;
7143
7185
  readonly wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
7186
+ readonly wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
7187
+ readonly wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
7144
7188
  readonly wasmpropertychange_name: (a: number) => [number, number];
7145
- readonly __wbg_wasmdjotlink_free: (a: number, b: number) => void;
7146
- readonly wasmpatternmatch_start: (a: number) => number;
7147
- readonly wasmredactionfinding_end: (a: number) => number;
7148
- readonly wasmredactionterm_value: (a: number) => [number, number];
7149
- readonly wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
7150
- readonly wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
7151
- readonly wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
7152
- readonly wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
7153
- readonly wasmdjotlink_url: (a: number) => [number, number];
7154
- readonly wasmdjotlink_title: (a: number) => [number, number];
7155
- readonly wasmdjotlink_text: (a: number) => [number, number];
7156
- readonly wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
7157
- readonly wasmerrormetadata_message: (a: number) => [number, number];
7158
- readonly __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
7159
- readonly wasmprocessingwarning_message: (a: number) => [number, number];
7160
- readonly __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
7161
- readonly wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
7162
- readonly wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
7163
- readonly wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7164
- readonly wasmpageboundary_byteEnd: (a: number) => number;
7165
- readonly wasmpageboundary_pageNumber: (a: number) => number;
7166
- readonly __wbg_wasmpageboundary_free: (a: number, b: number) => void;
7167
- readonly wasmredactionpattern_pattern: (a: number) => [number, number];
7168
- readonly __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
7169
- readonly wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
7170
- readonly wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
7171
- readonly wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
7172
- readonly wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
7173
- readonly wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
7174
- readonly wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
7175
- readonly wasmdbffieldinfo_name: (a: number) => [number, number];
7176
- readonly wasmdbffieldinfo_fieldType: (a: number) => [number, number];
7189
+ readonly wasmpropertychange_from: (a: number) => [number, number];
7190
+ readonly wasmpropertychange_to: (a: number) => [number, number];
7191
+ readonly wasmredactionfinding_start: (a: number) => number;
7177
7192
  readonly __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
7178
- readonly wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
7179
- readonly wasmdiffline_set_0: (a: number, b: number, c: number) => void;
7180
- readonly wasmdiffline_kind: (a: number) => [number, number];
7193
+ readonly __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
7181
7194
  readonly wasmdiffline_0: (a: number) => [number, number];
7182
- readonly wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
7183
- readonly wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
7184
- readonly wasmproxyconfig_username: (a: number) => [number, number];
7185
- readonly wasmproxyconfig_password: (a: number) => [number, number];
7186
- readonly wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
7187
- readonly wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
7188
- readonly wasmpropertychange_to: (a: number) => [number, number];
7189
- readonly wasmpropertychange_from: (a: number) => [number, number];
7190
- readonly __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
7191
- readonly wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
7192
- readonly wasmocrrotation_angleDegrees: (a: number) => number;
7193
- readonly wasmocrrotation_confidence: (a: number) => [number, number];
7194
- readonly __wbg_wasmocrrotation_free: (a: number, b: number) => void;
7195
- readonly __wbg_wasmredactionterm_free: (a: number, b: number) => void;
7196
- readonly __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
7197
- readonly wasmdocumentrelationship_target: (a: number) => number;
7195
+ readonly wasmpatternmatch_start: (a: number) => number;
7196
+ readonly wasmredactionfinding_end: (a: number) => number;
7197
+ readonly __wbg_wasmpageconfig_free: (a: number, b: number) => void;
7198
7198
  readonly wasmsupportedformat_mimeType: (a: number) => [number, number];
7199
7199
  readonly wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
7200
7200
  readonly __wbg_wasmsparseembeddingmodeltype_free: (a: number, b: number) => void;
package/pkg/web/xberg_wasm.js CHANGED
@@ -28124,7 +28124,7 @@ function __wbg_get_imports() {
28124
28124
  const ret = arg0.value;
28125
28125
  return ret;
28126
28126
  },
28127
- __wbg_warn_34b4e7d26f9b931b: function(arg0, arg1) {
28127
+ __wbg_warn_77d0951ec538e39e: function(arg0, arg1) {
28128
28128
  console.warn(getStringFromWasm0(arg0, arg1));
28129
28129
  },
28130
28130
  __wbg_wasmarchiveentry_new: function(arg0) {
package/pkg/web/xberg_wasm_bg.wasm CHANGED
Binary file
package/pkg/web/xberg_wasm_bg.wasm.d.ts CHANGED
@@ -2419,9 +2419,9 @@ export const wasmredactionterm_default: () => number;
2419
2419
  export const wasmprocessingwarning_new: (a: number, b: number, c: number, d: number) => number;
2420
2420
  export const wasmpagehierarchy_default: () => number;
2421
2421
  export const wasmqrboundingbox_default: () => number;
2422
- export const __wbg_wasmtablecell_free: (a: number, b: number) => void;
2423
- export const __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
2424
2422
  export const __wbg_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
2423
+ export const __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
2424
+ export const __wbg_wasmtablecell_free: (a: number, b: number) => void;
2425
2425
  export const wasmvlmfallbackpolicy_default: () => number;
2426
2426
  export const wasmproxyconfig_url: (a: number) => [number, number];
2427
2427
  export const wasmsupportedformat_extension: (a: number) => [number, number];
@@ -2452,38 +2452,88 @@ export const wasmtokenreductionoptions_set_mode: (a: number, b: number, c: numbe
2452
2452
  export const wasmpatternmatch_set_start: (a: number, b: number) => void;
2453
2453
  export const wasmpatternmatch_set_text: (a: number, b: number, c: number) => void;
2454
2454
  export const wasmerrormetadata_errorType: (a: number) => [number, number];
2455
+ export const wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
2455
2456
  export const wasmpageboundary_set_pageNumber: (a: number, b: number) => void;
2456
2457
  export const wasmpageboundary_set_byteStart: (a: number, b: number) => void;
2457
2458
  export const wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
2458
2459
  export const wasmpageboundary_byteStart: (a: number) => number;
2459
- export const wasmocrelementconfig_minConfidence: (a: number) => number;
2460
- export const wasmocrelementconfig_includeElements: (a: number) => number;
2460
+ export const wasmpageboundary_byteEnd: (a: number) => number;
2461
+ export const __wbg_wasmpageboundary_free: (a: number, b: number) => void;
2462
+ export const wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
2463
+ export const wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
2464
+ export const wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
2465
+ export const wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
2466
+ export const wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
2467
+ export const wasmdocumentrelationship_target: (a: number) => number;
2468
+ export const wasmpageboundary_pageNumber: (a: number) => number;
2469
+ export const __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
2470
+ export const wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
2461
2471
  export const wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
2462
2472
  export const wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
2473
+ export const wasmredactionterm_value: (a: number) => [number, number];
2463
2474
  export const wasmredactionterm_label: (a: number) => [number, number];
2464
2475
  export const wasmredactionterm_caseSensitive: (a: number) => number;
2476
+ export const __wbg_wasmredactionterm_free: (a: number, b: number) => void;
2477
+ export const wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
2478
+ export const wasmocrrotation_angleDegrees: (a: number) => number;
2479
+ export const wasmocrrotation_confidence: (a: number) => [number, number];
2480
+ export const __wbg_wasmocrrotation_free: (a: number, b: number) => void;
2481
+ export const wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
2482
+ export const wasmdbffieldinfo_name: (a: number) => [number, number];
2483
+ export const wasmdbffieldinfo_fieldType: (a: number) => [number, number];
2484
+ export const wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
2485
+ export const wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
2486
+ export const wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
2487
+ export const wasmdjotlink_url: (a: number) => [number, number];
2488
+ export const wasmdjotlink_title: (a: number) => [number, number];
2489
+ export const wasmdjotlink_text: (a: number) => [number, number];
2490
+ export const __wbg_wasmdjotlink_free: (a: number, b: number) => void;
2491
+ export const wasmocrelementconfig_minConfidence: (a: number) => number;
2492
+ export const wasmocrelementconfig_includeElements: (a: number) => number;
2493
+ export const __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
2465
2494
  export const wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
2466
2495
  export const wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
2467
2496
  export const wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
2468
2497
  export const wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
2469
2498
  export const wasmpageconfig_markerFormat: (a: number) => [number, number];
2470
2499
  export const wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
2471
- export const __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
2500
+ export const wasmredactionpattern_pattern: (a: number) => [number, number];
2501
+ export const __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
2502
+ export const wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
2503
+ export const wasmprocessingwarning_message: (a: number) => [number, number];
2504
+ export const __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
2505
+ export const wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
2506
+ export const wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
2507
+ export const wasmerrormetadata_message: (a: number) => [number, number];
2508
+ export const __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
2509
+ export const __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
2510
+ export const wasmpreprocessingoptions_enabled: (a: number) => number;
2511
+ export const __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
2512
+ export const wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
2513
+ export const wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
2514
+ export const wasmproxyconfig_password: (a: number) => [number, number];
2515
+ export const wasmproxyconfig_username: (a: number) => [number, number];
2516
+ export const __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
2517
+ export const wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
2518
+ export const wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
2519
+ export const wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
2520
+ export const wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
2521
+ export const wasmsparseembeddingconfig_model: (a: number) => any;
2522
+ export const wasmsparseembeddingconfig_acceleration: (a: number) => number;
2472
2523
  export const wasmrerankerconfig_set_topK: (a: number, b: number) => void;
2473
2524
  export const wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
2474
2525
  export const wasmrerankerconfig_set_model: (a: number, b: any) => void;
2526
+ export const wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
2527
+ export const wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2475
2528
  export const wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
2476
2529
  export const wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
2477
2530
  export const wasmrerankerconfig_topK: (a: number) => number;
2478
2531
  export const wasmrerankerconfig_showDownloadProgress: (a: number) => number;
2479
2532
  export const wasmrerankerconfig_model: (a: number) => any;
2533
+ export const wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
2534
+ export const wasmrerankerconfig_cacheDir: (a: number) => [number, number];
2535
+ export const wasmrerankerconfig_batchSize: (a: number) => number;
2480
2536
  export const wasmrerankerconfig_acceleration: (a: number) => number;
2481
- export const wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
2482
- export const wasmssrfpolicy_denyPrivate: (a: number) => number;
2483
- export const wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
2484
- export const wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
2485
- export const wasmredactionfinding_start: (a: number) => number;
2486
- export const __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
2487
2537
  export const wasmqrboundingbox_set_y: (a: number, b: number) => void;
2488
2538
  export const wasmqrboundingbox_set_x: (a: number, b: number) => void;
2489
2539
  export const wasmqrboundingbox_set_width: (a: number, b: number) => void;
@@ -2493,8 +2543,17 @@ export const wasmqrboundingbox_x: (a: number) => number;
2493
2543
  export const wasmqrboundingbox_width: (a: number) => number;
2494
2544
  export const wasmqrboundingbox_height: (a: number) => number;
2495
2545
  export const __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
2496
- export const __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
2497
- export const wasmpreprocessingoptions_enabled: (a: number) => number;
2546
+ export const wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
2547
+ export const wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2548
+ export const wasmsparseembeddingconfig_maxLength: (a: number) => number;
2549
+ export const wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
2550
+ export const wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
2551
+ export const wasmsparseembeddingconfig_batchSize: (a: number) => number;
2552
+ export const wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
2553
+ export const wasmdiffline_set_0: (a: number, b: number, c: number) => void;
2554
+ export const wasmdiffline_kind: (a: number) => [number, number];
2555
+ export const wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
2556
+ export const wasmssrfpolicy_denyPrivate: (a: number) => number;
2498
2557
  export const wasmtablecell_set_rowSpan: (a: number, b: number) => void;
2499
2558
  export const wasmtablecell_set_isHeader: (a: number, b: number) => void;
2500
2559
  export const wasmtablecell_set_content: (a: number, b: number, c: number) => void;
@@ -2503,78 +2562,19 @@ export const wasmtablecell_rowSpan: (a: number) => number;
2503
2562
  export const wasmtablecell_isHeader: (a: number) => number;
2504
2563
  export const wasmtablecell_content: (a: number) => [number, number];
2505
2564
  export const wasmtablecell_colSpan: (a: number) => number;
2506
- export const __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
2507
- export const __wbg_wasmpageconfig_free: (a: number, b: number) => void;
2508
- export const wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
2509
- export const wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
2510
- export const wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
2511
- export const wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2512
- export const wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
2513
- export const wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
2514
- export const wasmsparseembeddingconfig_model: (a: number) => any;
2515
- export const wasmsparseembeddingconfig_maxLength: (a: number) => number;
2516
- export const wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
2517
- export const wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
2518
- export const wasmsparseembeddingconfig_batchSize: (a: number) => number;
2519
- export const wasmsparseembeddingconfig_acceleration: (a: number) => number;
2520
- export const wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
2521
- export const wasmrerankerconfig_cacheDir: (a: number) => [number, number];
2522
- export const wasmrerankerconfig_batchSize: (a: number) => number;
2523
2565
  export const wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
2566
+ export const wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
2567
+ export const wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
2524
2568
  export const wasmpropertychange_name: (a: number) => [number, number];
2525
- export const __wbg_wasmdjotlink_free: (a: number, b: number) => void;
2526
- export const wasmpatternmatch_start: (a: number) => number;
2527
- export const wasmredactionfinding_end: (a: number) => number;
2528
- export const wasmredactionterm_value: (a: number) => [number, number];
2529
- export const wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
2530
- export const wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
2531
- export const wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
2532
- export const wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
2533
- export const wasmdjotlink_url: (a: number) => [number, number];
2534
- export const wasmdjotlink_title: (a: number) => [number, number];
2535
- export const wasmdjotlink_text: (a: number) => [number, number];
2536
- export const wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
2537
- export const wasmerrormetadata_message: (a: number) => [number, number];
2538
- export const __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
2539
- export const wasmprocessingwarning_message: (a: number) => [number, number];
2540
- export const __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
2541
- export const wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
2542
- export const wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
2543
- export const wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2544
- export const wasmpageboundary_byteEnd: (a: number) => number;
2545
- export const wasmpageboundary_pageNumber: (a: number) => number;
2546
- export const __wbg_wasmpageboundary_free: (a: number, b: number) => void;
2547
- export const wasmredactionpattern_pattern: (a: number) => [number, number];
2548
- export const __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
2549
- export const wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
2550
- export const wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
2551
- export const wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
2552
- export const wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
2553
- export const wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
2554
- export const wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
2555
- export const wasmdbffieldinfo_name: (a: number) => [number, number];
2556
- export const wasmdbffieldinfo_fieldType: (a: number) => [number, number];
2569
+ export const wasmpropertychange_from: (a: number) => [number, number];
2570
+ export const wasmpropertychange_to: (a: number) => [number, number];
2571
+ export const wasmredactionfinding_start: (a: number) => number;
2557
2572
  export const __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
2558
- export const wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
2559
- export const wasmdiffline_set_0: (a: number, b: number, c: number) => void;
2560
- export const wasmdiffline_kind: (a: number) => [number, number];
2573
+ export const __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
2561
2574
  export const wasmdiffline_0: (a: number) => [number, number];
2562
- export const wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
2563
- export const wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
2564
- export const wasmproxyconfig_username: (a: number) => [number, number];
2565
- export const wasmproxyconfig_password: (a: number) => [number, number];
2566
- export const wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
2567
- export const wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
2568
- export const wasmpropertychange_to: (a: number) => [number, number];
2569
- export const wasmpropertychange_from: (a: number) => [number, number];
2570
- export const __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
2571
- export const wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
2572
- export const wasmocrrotation_angleDegrees: (a: number) => number;
2573
- export const wasmocrrotation_confidence: (a: number) => [number, number];
2574
- export const __wbg_wasmocrrotation_free: (a: number, b: number) => void;
2575
- export const __wbg_wasmredactionterm_free: (a: number, b: number) => void;
2576
- export const __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
2577
- export const wasmdocumentrelationship_target: (a: number) => number;
2575
+ export const wasmpatternmatch_start: (a: number) => number;
2576
+ export const wasmredactionfinding_end: (a: number) => number;
2577
+ export const __wbg_wasmpageconfig_free: (a: number, b: number) => void;
2578
2578
  export const wasmsupportedformat_mimeType: (a: number) => [number, number];
2579
2579
  export const wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
2580
2580
  export const __wbg_wasmsparseembeddingmodeltype_free: (a: number, b: number) => void;