@xberg-io/xberg-wasm 1.0.8 → 1.0.11

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 +75 -50
  2. package/package.json +1 -1
  3. package/pkg/web/README.md +75 -50
  4. package/pkg/web/package.json +1 -1
  5. package/pkg/web/xberg_wasm.d.ts +86 -86
  6. package/pkg/web/xberg_wasm.js +3 -3
  7. package/pkg/web/xberg_wasm_bg.wasm +0 -0
  8. package/pkg/web/xberg_wasm_bg.wasm.d.ts +86 -86
package/README.md CHANGED
@@ -80,7 +80,7 @@
80
80
  </a>
81
81
  </div>
82
82
 
83
- Extract text, tables, images, metadata, and code intelligence from 101 file formats and 306 programming languages including PDF, Office documents, and images. WebAssembly bindings for browsers, Deno, and Cloudflare Workers with portable deployment and multi-threading support.
83
+ Extract text, tables, images, metadata, and code intelligence from 101 file formats and 371 programming languages including PDF, Office documents, and images. WebAssembly bindings for browsers, Deno, and Cloudflare Workers with portable deployment and multi-threading support.
84
84
 
85
85
  ## What This Package Provides
86
86
 
@@ -100,7 +100,6 @@ pnpm add @xberg-io/xberg-wasm
100
100
  ```
101
101
 
102
102
  ### System Requirements
103
-
104
103
  - Modern browser with WebAssembly support, or Deno 1.0+, or Cloudflare Workers
105
104
  - Optional: [Tesseract WASM](https://github.com/naptha/tesseract.js) for OCR functionality
106
105
 
@@ -262,58 +261,84 @@ extractDocuments(fileBytes, mimes)
262
261
 
263
262
  ## Features
264
263
 
265
- ### Supported File Formats (101 formats · 115 file extensions) 101 formats across 115 file extensions in 8 major
266
-
267
- categories with intelligent format detection and comprehensive metadata extraction. #### Office Documents | Category |
268
- Formats | Capabilities | |----------|---------|--------------| | **Word Processing** | `.docx`, `.docm`, `.doc`,
269
- `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles
270
- | | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers`
271
- | Sheet data, formulas, cell metadata, charts | | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`,
272
- `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata | | **PDF** | `.pdf` | Text, tables, images,
273
- metadata, OCR support | | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources | | **Database** |
274
- `.dbf` | Table data extraction, field type support | | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text
275
- extraction | #### Images (OCR-Enabled) | Category | Formats | Features | |----------|---------|----------| | **Raster**
276
- | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions,
277
- color space | | **Advanced** | `.jp2`, `.jpx`, `.jpm`, `.mj2`, `.jbig2`, `.jb2`, `.pnm`, `.pbm`, `.pgm`, `.ppm` | OCR
278
- via hayro-jpeg2000 (pure Rust decoder), JBIG2 support, table detection, format-specific metadata |
264
+ ### Supported File Formats (101 formats · 115 file extensions)
265
+
266
+ 101 formats across 115 file extensions in 8 major categories with intelligent format detection and comprehensive metadata extraction.
267
+
268
+ #### Office Documents
269
+
270
+ | Category | Formats | Capabilities |
271
+ |----------|---------|--------------|
272
+ | **Word Processing** | `.docx`, `.docm`, `.doc`, `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles |
273
+ | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers` | Sheet data, formulas, cell metadata, charts |
274
+ | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`, `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata |
275
+ | **PDF** | `.pdf` | Text, tables, images, metadata, OCR support |
276
+ | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources |
277
+ | **Database** | `.dbf` | Table data extraction, field type support |
278
+ | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text extraction |
279
+
280
+ #### Images (OCR-Enabled)
281
+
282
+ | Category | Formats | Features |
283
+ |----------|---------|----------|
284
+ | **Raster** | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions, color space |
285
+ | **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 |
279
286
  | **Vector** | `.svg` | DOM parsing, embedded text, graphics metadata |
280
287
 
281
- #### Web & Data | Category | Formats | Features | |----------|---------|----------| | **Markup** | `.html`, `.htm`
282
-
283
- `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction | | **Structured Data** |
284
- `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation | | **Text &
285
- Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX,
286
- reStructuredText, Org Mode | #### Email & Archives | Category | Formats | Features | |----------|---------|----------| |
287
- **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading | | **Archives** | `.zip`,
288
- `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
289
-
290
- #### Academic & Scientific | Category | Formats | Features | |----------|---------|----------| | **Citations** | `.bib`
291
-
292
- `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
293
- | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS
294
- | | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML
295
- outlines | | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats | #### Code Intelligence
296
- (306 Languages) | Feature | Description | |---------|-------------| | **Structure Extraction** | Functions, classes,
297
- methods, structs, interfaces, enums | | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
298
- | **Symbol Extraction** | Variables, constants, type aliases, properties | | **Docstring Parsing** | Google, NumPy,
299
- Sphinx, JSDoc, RustDoc, and 10+ formats | | **Diagnostics** | Parse errors with line/column positions | | **Syntax-Aware
300
- Chunking** | Split code by semantic boundaries, not arbitrary byte offsets | Powered by
301
- [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) —
302
- [documentation](https://docs.tree-sitter-language-pack.xberg.io). **[Complete Format
303
- Reference](https://docs.xberg.io/reference/formats/)** ### Key Capabilities - **Text Extraction** - Extract all text
304
- content with position and formatting information - **Metadata Extraction** - Retrieve document properties, creation
305
- date, author, etc. - **Table Extraction** - Parse tables with structure and cell content preservation - **Image
306
- Extraction** - Extract embedded images and render page previews
288
+ #### Web & Data
289
+
290
+ | Category | Formats | Features |
291
+ |----------|---------|----------|
292
+ | **Markup** | `.html`, `.htm`, `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction |
293
+ | **Structured Data** | `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation |
294
+ | **Text & Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX, reStructuredText, Org Mode |
295
+
296
+ #### Email & Archives
297
+
298
+ | Category | Formats | Features |
299
+ |----------|---------|----------|
300
+ | **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading |
301
+ | **Archives** | `.zip`, `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
302
+
303
+ #### Academic & Scientific
304
+
305
+ | Category | Formats | Features |
306
+ |----------|---------|----------|
307
+ | **Citations** | `.bib`, `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
308
+ | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS |
309
+ | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML outlines |
310
+ | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats |
311
+
312
+ #### Code Intelligence (371 Languages)
313
+
314
+ | Feature | Description |
315
+ |---------|-------------|
316
+ | **Structure Extraction** | Functions, classes, methods, structs, interfaces, enums |
317
+ | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
318
+ | **Symbol Extraction** | Variables, constants, type aliases, properties |
319
+ | **Docstring Parsing** | Google, NumPy, Sphinx, JSDoc, RustDoc, and 10+ formats |
320
+ | **Diagnostics** | Parse errors with line/column positions |
321
+ | **Syntax-Aware Chunking** | Split code by semantic boundaries, not arbitrary byte offsets |
322
+
323
+ Powered by [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) — [documentation](https://docs.tree-sitter-language-pack.xberg.io).
324
+
325
+ **[Complete Format Reference](https://docs.xberg.io/reference/formats/)**
326
+
327
+ ### Key Capabilities
307
328
 
329
+ - **Text Extraction** - Extract all text content with position and formatting information
330
+ - **Metadata Extraction** - Retrieve document properties, creation date, author, etc.
331
+ - **Table Extraction** - Parse tables with structure and cell content preservation
332
+ - **Image Extraction** - Extract embedded images and render page previews
308
333
  - **OCR Support** - Integrate multiple OCR backends for scanned documents
309
- - **Async/Await** - Non-blocking document processing with concurrent operations
310
- - **Plugin System** - Extensible post-processing for custom text transformation
311
- - **Batch Processing** - Efficiently process multiple documents in parallel - **Memory Efficient** - Stream large files
312
- without loading entirely into memory - **Language Detection** - Detect and support multiple languages in documents
313
- - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [306 programming
314
- languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
315
- - **Configuration** - Fine-grained control over extraction behavior - **Six Output Formats** - Plain text, Markdown,
316
- Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
334
+ - **Async/Await** - Non-blocking document processing with concurrent operations
335
+ - **Plugin System** - Extensible post-processing for custom text transformation
336
+ - **Batch Processing** - Efficiently process multiple documents in parallel
337
+ - **Memory Efficient** - Stream large files without loading entirely into memory
338
+ - **Language Detection** - Detect and support multiple languages in documents
339
+ - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [371 programming languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
340
+ - **Configuration** - Fine-grained control over extraction behavior
341
+ - **Six Output Formats** - Plain text, Markdown, Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
317
342
 
318
343
  ## OCR Support
319
344
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@xberg-io/xberg-wasm",
3
- "version": "1.0.8",
3
+ "version": "1.0.11",
4
4
  "private": false,
5
5
  "description": "High-performance document intelligence library",
6
6
  "license": "MIT",
package/pkg/web/README.md CHANGED
@@ -80,7 +80,7 @@
80
80
  </a>
81
81
  </div>
82
82
 
83
- Extract text, tables, images, metadata, and code intelligence from 101 file formats and 306 programming languages including PDF, Office documents, and images. WebAssembly bindings for browsers, Deno, and Cloudflare Workers with portable deployment and multi-threading support.
83
+ Extract text, tables, images, metadata, and code intelligence from 101 file formats and 371 programming languages including PDF, Office documents, and images. WebAssembly bindings for browsers, Deno, and Cloudflare Workers with portable deployment and multi-threading support.
84
84
 
85
85
  ## What This Package Provides
86
86
 
@@ -100,7 +100,6 @@ pnpm add @xberg-io/xberg-wasm
100
100
  ```
101
101
 
102
102
  ### System Requirements
103
-
104
103
  - Modern browser with WebAssembly support, or Deno 1.0+, or Cloudflare Workers
105
104
  - Optional: [Tesseract WASM](https://github.com/naptha/tesseract.js) for OCR functionality
106
105
 
@@ -262,58 +261,84 @@ extractDocuments(fileBytes, mimes)
262
261
 
263
262
  ## Features
264
263
 
265
- ### Supported File Formats (101 formats · 115 file extensions) 101 formats across 115 file extensions in 8 major
266
-
267
- categories with intelligent format detection and comprehensive metadata extraction. #### Office Documents | Category |
268
- Formats | Capabilities | |----------|---------|--------------| | **Word Processing** | `.docx`, `.docm`, `.doc`,
269
- `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles
270
- | | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers`
271
- | Sheet data, formulas, cell metadata, charts | | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`,
272
- `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata | | **PDF** | `.pdf` | Text, tables, images,
273
- metadata, OCR support | | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources | | **Database** |
274
- `.dbf` | Table data extraction, field type support | | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text
275
- extraction | #### Images (OCR-Enabled) | Category | Formats | Features | |----------|---------|----------| | **Raster**
276
- | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions,
277
- color space | | **Advanced** | `.jp2`, `.jpx`, `.jpm`, `.mj2`, `.jbig2`, `.jb2`, `.pnm`, `.pbm`, `.pgm`, `.ppm` | OCR
278
- via hayro-jpeg2000 (pure Rust decoder), JBIG2 support, table detection, format-specific metadata |
264
+ ### Supported File Formats (101 formats · 115 file extensions)
265
+
266
+ 101 formats across 115 file extensions in 8 major categories with intelligent format detection and comprehensive metadata extraction.
267
+
268
+ #### Office Documents
269
+
270
+ | Category | Formats | Capabilities |
271
+ |----------|---------|--------------|
272
+ | **Word Processing** | `.docx`, `.docm`, `.doc`, `.dotx`, `.dotm`, `.dot`, `.odt`, `.pages`, `.wpd`, `.wp`, `.wp5`, `.wp6` | Full text, tables, images, metadata, styles |
273
+ | **Spreadsheets** | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`, `.xla`, `.xlam`, `.xltm`, `.xltx`, `.xlt`, `.ods`, `.numbers` | Sheet data, formulas, cell metadata, charts |
274
+ | **Presentations** | `.pptx`, `.pptm`, `.ppt`, `.ppsx`, `.potx`, `.potm`, `.pot`, `.odp`, `.key` | Slides, speaker notes, images, metadata |
275
+ | **PDF** | `.pdf` | Text, tables, images, metadata, OCR support |
276
+ | **eBooks** | `.epub`, `.fb2` | Chapters, metadata, embedded resources |
277
+ | **Database** | `.dbf` | Table data extraction, field type support |
278
+ | **Hangul** | `.hwp`, `.hwpx` | Korean document format, text extraction |
279
+
280
+ #### Images (OCR-Enabled)
281
+
282
+ | Category | Formats | Features |
283
+ |----------|---------|----------|
284
+ | **Raster** | `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.bmp`, `.tiff`, `.tif` | OCR, table detection, EXIF metadata, dimensions, color space |
285
+ | **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 |
279
286
  | **Vector** | `.svg` | DOM parsing, embedded text, graphics metadata |
280
287
 
281
- #### Web & Data | Category | Formats | Features | |----------|---------|----------| | **Markup** | `.html`, `.htm`
282
-
283
- `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction | | **Structured Data** |
284
- `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation | | **Text &
285
- Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX,
286
- reStructuredText, Org Mode | #### Email & Archives | Category | Formats | Features | |----------|---------|----------| |
287
- **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading | | **Archives** | `.zip`,
288
- `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
289
-
290
- #### Academic & Scientific | Category | Formats | Features | |----------|---------|----------| | **Citations** | `.bib`
291
-
292
- `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
293
- | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS
294
- | | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML
295
- outlines | | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats | #### Code Intelligence
296
- (306 Languages) | Feature | Description | |---------|-------------| | **Structure Extraction** | Functions, classes,
297
- methods, structs, interfaces, enums | | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
298
- | **Symbol Extraction** | Variables, constants, type aliases, properties | | **Docstring Parsing** | Google, NumPy,
299
- Sphinx, JSDoc, RustDoc, and 10+ formats | | **Diagnostics** | Parse errors with line/column positions | | **Syntax-Aware
300
- Chunking** | Split code by semantic boundaries, not arbitrary byte offsets | Powered by
301
- [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) —
302
- [documentation](https://docs.tree-sitter-language-pack.xberg.io). **[Complete Format
303
- Reference](https://docs.xberg.io/reference/formats/)** ### Key Capabilities - **Text Extraction** - Extract all text
304
- content with position and formatting information - **Metadata Extraction** - Retrieve document properties, creation
305
- date, author, etc. - **Table Extraction** - Parse tables with structure and cell content preservation - **Image
306
- Extraction** - Extract embedded images and render page previews
288
+ #### Web & Data
289
+
290
+ | Category | Formats | Features |
291
+ |----------|---------|----------|
292
+ | **Markup** | `.html`, `.htm`, `.xhtml`, `.xml`, `.svg` | DOM parsing, metadata (Open Graph, Twitter Card), link extraction |
293
+ | **Structured Data** | `.json`, `.yaml`, `.yml`, `.toml`, `.csv`, `.tsv` | Schema detection, nested structures, validation |
294
+ | **Text & Markdown** | `.txt`, `.md`, `.markdown`, `.djot`, `.mdx`, `.rst`, `.org`, `.rtf` | CommonMark, GFM, Djot, MDX, reStructuredText, Org Mode |
295
+
296
+ #### Email & Archives
297
+
298
+ | Category | Formats | Features |
299
+ |----------|---------|----------|
300
+ | **Email** | `.eml`, `.msg`, `.pst` | Headers, body (HTML/plain), attachments, threading |
301
+ | **Archives** | `.zip`, `.tar`, `.tgz`, `.gz`, `.7z` | Recursive extraction of nested archives, file listing, metadata, zip-bomb protection |
302
+
303
+ #### Academic & Scientific
304
+
305
+ | Category | Formats | Features |
306
+ |----------|---------|----------|
307
+ | **Citations** | `.bib`, `.ris`, `.nbib`, `.enw` | Structured parsing: RIS, PubMed/MEDLINE, EndNote XML, BibTeX/BibLaTeX, CSL JSON by MIME type |
308
+ | **Scientific** | `.tex`, `.latex`, `.typ`, `.typst`, `.jats`, `.ipynb` | LaTeX, Typst, Jupyter notebooks, PubMed JATS |
309
+ | **Publishing** | `.fb2`, `.docbook`, `.dbk`, `.docbook4`, `.docbook5`, `.opml` | FictionBook, DocBook XML, OPML outlines |
310
+ | **Documentation** | MIME-only POD, mdoc, troff | Technical documentation formats |
311
+
312
+ #### Code Intelligence (371 Languages)
313
+
314
+ | Feature | Description |
315
+ |---------|-------------|
316
+ | **Structure Extraction** | Functions, classes, methods, structs, interfaces, enums |
317
+ | **Import/Export Analysis** | Module dependencies, re-exports, wildcard imports |
318
+ | **Symbol Extraction** | Variables, constants, type aliases, properties |
319
+ | **Docstring Parsing** | Google, NumPy, Sphinx, JSDoc, RustDoc, and 10+ formats |
320
+ | **Diagnostics** | Parse errors with line/column positions |
321
+ | **Syntax-Aware Chunking** | Split code by semantic boundaries, not arbitrary byte offsets |
322
+
323
+ Powered by [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) — [documentation](https://docs.tree-sitter-language-pack.xberg.io).
324
+
325
+ **[Complete Format Reference](https://docs.xberg.io/reference/formats/)**
326
+
327
+ ### Key Capabilities
307
328
 
329
+ - **Text Extraction** - Extract all text content with position and formatting information
330
+ - **Metadata Extraction** - Retrieve document properties, creation date, author, etc.
331
+ - **Table Extraction** - Parse tables with structure and cell content preservation
332
+ - **Image Extraction** - Extract embedded images and render page previews
308
333
  - **OCR Support** - Integrate multiple OCR backends for scanned documents
309
- - **Async/Await** - Non-blocking document processing with concurrent operations
310
- - **Plugin System** - Extensible post-processing for custom text transformation
311
- - **Batch Processing** - Efficiently process multiple documents in parallel - **Memory Efficient** - Stream large files
312
- without loading entirely into memory - **Language Detection** - Detect and support multiple languages in documents
313
- - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [306 programming
314
- languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
315
- - **Configuration** - Fine-grained control over extraction behavior - **Six Output Formats** - Plain text, Markdown,
316
- Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
334
+ - **Async/Await** - Non-blocking document processing with concurrent operations
335
+ - **Plugin System** - Extensible post-processing for custom text transformation
336
+ - **Batch Processing** - Efficiently process multiple documents in parallel
337
+ - **Memory Efficient** - Stream large files without loading entirely into memory
338
+ - **Language Detection** - Detect and support multiple languages in documents
339
+ - **Code Intelligence** - Extract structure, imports, exports, symbols, and docstrings from [371 programming languages](https://docs.tree-sitter-language-pack.xberg.io) via tree-sitter
340
+ - **Configuration** - Fine-grained control over extraction behavior
341
+ - **Six Output Formats** - Plain text, Markdown, Djot, HTML, JSON tree structure, or Structured JSON with OCR metadata
317
342
 
318
343
  ## OCR Support
319
344
 
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": "1.0.8",
5
+ "version": "1.0.11",
6
6
  "license": "MIT",
7
7
  "repository": {
8
8
  "type": "git",
package/pkg/web/xberg_wasm.d.ts CHANGED
@@ -7039,8 +7039,8 @@ 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_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
7043
7042
  readonly __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
7043
+ readonly __wbg_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
7044
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];
@@ -7076,66 +7076,105 @@ export interface InitOutput {
7076
7076
  readonly wasmpageboundary_set_byteStart: (a: number, b: number) => void;
7077
7077
  readonly wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
7078
7078
  readonly wasmpageboundary_byteStart: (a: number) => number;
7079
- readonly __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
7080
- readonly wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
7079
+ readonly wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
7080
+ readonly wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
7081
+ readonly wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
7082
+ readonly wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
7083
+ readonly wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
7084
+ readonly wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7081
7085
  readonly wasmdbffieldinfo_name: (a: number) => [number, number];
7082
- readonly wasmprocessingwarning_message: (a: number) => [number, number];
7086
+ readonly wasmsupportedformat_mimeType: (a: number) => [number, number];
7087
+ readonly wasmredactionpattern_pattern: (a: number) => [number, number];
7088
+ readonly __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
7089
+ readonly wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
7090
+ readonly __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
7091
+ readonly wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
7092
+ readonly wasmredactionfinding_start: (a: number) => number;
7093
+ readonly wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
7094
+ readonly wasmdiffline_set_0: (a: number, b: number, c: number) => void;
7095
+ readonly wasmdiffline_kind: (a: number) => [number, number];
7096
+ readonly wasmdiffline_0: (a: number) => [number, number];
7097
+ readonly wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
7098
+ readonly wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
7099
+ readonly wasmdbffieldinfo_fieldType: (a: number) => [number, number];
7100
+ readonly __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
7101
+ readonly __wbg_wasmpageboundary_free: (a: number, b: number) => void;
7102
+ readonly wasmpageboundary_byteEnd: (a: number) => number;
7103
+ readonly wasmsparseembeddingconfig_model: (a: number) => any;
7104
+ readonly wasmsparseembeddingconfig_maxLength: (a: number) => number;
7105
+ readonly wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
7106
+ readonly wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
7107
+ readonly wasmsparseembeddingconfig_batchSize: (a: number) => number;
7108
+ readonly wasmsparseembeddingconfig_acceleration: (a: number) => number;
7083
7109
  readonly wasmrerankerconfig_set_topK: (a: number, b: number) => void;
7084
7110
  readonly wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
7085
7111
  readonly wasmrerankerconfig_set_model: (a: number, b: any) => void;
7112
+ readonly wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
7113
+ readonly wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7086
7114
  readonly wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
7087
7115
  readonly wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
7088
7116
  readonly wasmrerankerconfig_topK: (a: number) => number;
7089
7117
  readonly wasmrerankerconfig_showDownloadProgress: (a: number) => number;
7090
7118
  readonly wasmrerankerconfig_model: (a: number) => any;
7119
+ readonly wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
7120
+ readonly wasmrerankerconfig_cacheDir: (a: number) => [number, number];
7121
+ readonly wasmrerankerconfig_batchSize: (a: number) => number;
7091
7122
  readonly wasmrerankerconfig_acceleration: (a: number) => number;
7092
- readonly wasmpreprocessingoptions_enabled: (a: number) => number;
7123
+ readonly wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
7124
+ readonly wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
7125
+ readonly wasmocrrotation_angleDegrees: (a: number) => number;
7126
+ readonly wasmocrrotation_confidence: (a: number) => [number, number];
7127
+ readonly __wbg_wasmocrrotation_free: (a: number, b: number) => void;
7128
+ readonly __wbg_wasmdjotlink_free: (a: number, b: number) => void;
7129
+ readonly wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
7130
+ readonly wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
7131
+ readonly wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
7132
+ readonly wasmredactionterm_value: (a: number) => [number, number];
7133
+ readonly wasmredactionterm_label: (a: number) => [number, number];
7134
+ readonly wasmredactionterm_caseSensitive: (a: number) => number;
7135
+ readonly __wbg_wasmredactionterm_free: (a: number, b: number) => void;
7093
7136
  readonly wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
7137
+ readonly wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
7138
+ readonly wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
7094
7139
  readonly wasmdjotlink_url: (a: number) => [number, number];
7095
- readonly wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
7096
- readonly __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
7097
- readonly wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
7098
- readonly wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
7099
- readonly wasmpageconfig_markerFormat: (a: number) => [number, number];
7100
- readonly wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
7101
- readonly __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
7102
- readonly wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
7103
- readonly wasmerrormetadata_message: (a: number) => [number, number];
7140
+ readonly wasmdjotlink_title: (a: number) => [number, number];
7141
+ readonly wasmdjotlink_text: (a: number) => [number, number];
7142
+ readonly wasmocrelementconfig_minConfidence: (a: number) => number;
7143
+ readonly wasmocrelementconfig_includeElements: (a: number) => number;
7144
+ readonly wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
7145
+ readonly wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
7146
+ readonly wasmprocessingwarning_message: (a: number) => [number, number];
7104
7147
  readonly __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
7105
- readonly wasmredactionfinding_start: (a: number) => number;
7106
- readonly wasmredactionfinding_end: (a: number) => number;
7107
- readonly wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
7108
- readonly wasmssrfpolicy_denyPrivate: (a: number) => number;
7109
- readonly wasmpageboundary_byteEnd: (a: number) => number;
7148
+ readonly wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
7149
+ readonly wasmdocumentrelationship_target: (a: number) => number;
7110
7150
  readonly wasmpageboundary_pageNumber: (a: number) => number;
7111
- readonly __wbg_wasmpageboundary_free: (a: number, b: number) => void;
7151
+ readonly __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
7112
7152
  readonly wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
7113
- readonly wasmpropertychange_name: (a: number) => [number, number];
7114
- readonly wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
7115
7153
  readonly wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
7116
- readonly wasmpropertychange_to: (a: number) => [number, number];
7154
+ readonly wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
7155
+ readonly wasmpropertychange_name: (a: number) => [number, number];
7117
7156
  readonly wasmpropertychange_from: (a: number) => [number, number];
7118
- readonly wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
7119
- readonly wasmrerankerconfig_cacheDir: (a: number) => [number, number];
7120
- readonly wasmrerankerconfig_batchSize: (a: number) => number;
7121
- readonly wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
7122
- readonly wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7123
- readonly wasmpatternmatch_start: (a: number) => number;
7157
+ readonly wasmpropertychange_to: (a: number) => [number, number];
7158
+ readonly wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
7159
+ readonly wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
7160
+ readonly wasmproxyconfig_username: (a: number) => [number, number];
7161
+ readonly wasmproxyconfig_password: (a: number) => [number, number];
7162
+ readonly __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
7163
+ readonly wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
7164
+ readonly wasmerrormetadata_message: (a: number) => [number, number];
7124
7165
  readonly __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
7125
- readonly wasmocrelementconfig_minConfidence: (a: number) => number;
7126
- readonly wasmocrelementconfig_includeElements: (a: number) => number;
7127
- readonly __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
7128
- readonly wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
7129
- readonly wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
7130
- readonly wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
7131
- readonly wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
7132
7166
  readonly wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
7133
- readonly wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
7134
- readonly wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
7135
- readonly wasmdbffieldinfo_fieldType: (a: number) => [number, number];
7136
- readonly __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
7137
- readonly wasmsupportedformat_mimeType: (a: number) => [number, number];
7138
- readonly wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
7167
+ readonly wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
7168
+ readonly wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
7169
+ readonly wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
7170
+ readonly wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
7171
+ readonly wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
7172
+ readonly wasmpageconfig_markerFormat: (a: number) => [number, number];
7173
+ readonly wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
7174
+ readonly __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
7175
+ readonly __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
7176
+ readonly wasmpreprocessingoptions_enabled: (a: number) => number;
7177
+ readonly __wbg_wasmpageconfig_free: (a: number, b: number) => void;
7139
7178
  readonly wasmqrboundingbox_set_y: (a: number, b: number) => void;
7140
7179
  readonly wasmqrboundingbox_set_x: (a: number, b: number) => void;
7141
7180
  readonly wasmqrboundingbox_set_width: (a: number, b: number) => void;
@@ -7145,45 +7184,11 @@ export interface InitOutput {
7145
7184
  readonly wasmqrboundingbox_width: (a: number) => number;
7146
7185
  readonly wasmqrboundingbox_height: (a: number) => number;
7147
7186
  readonly __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
7148
- readonly wasmredactionpattern_pattern: (a: number) => [number, number];
7149
- readonly __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
7150
- readonly wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
7151
- readonly __wbg_wasmpageconfig_free: (a: number, b: number) => void;
7152
- readonly wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
7153
- readonly wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
7154
- readonly wasmdjotlink_title: (a: number) => [number, number];
7155
- readonly wasmdjotlink_text: (a: number) => [number, number];
7156
- readonly __wbg_wasmdjotlink_free: (a: number, b: number) => void;
7157
- readonly wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
7158
- readonly wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
7159
- readonly wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
7160
- readonly wasmredactionterm_value: (a: number) => [number, number];
7161
- readonly wasmredactionterm_label: (a: number) => [number, number];
7162
- readonly wasmredactionterm_caseSensitive: (a: number) => number;
7163
- readonly __wbg_wasmredactionterm_free: (a: number, b: number) => void;
7164
- readonly wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
7165
- readonly wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
7166
- readonly wasmproxyconfig_username: (a: number) => [number, number];
7167
- readonly wasmproxyconfig_password: (a: number) => [number, number];
7168
- readonly __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
7169
- readonly wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
7170
- readonly wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
7171
- readonly wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
7172
- readonly wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7173
- readonly wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
7174
- readonly wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
7175
- readonly wasmsparseembeddingconfig_model: (a: number) => any;
7176
- readonly wasmsparseembeddingconfig_maxLength: (a: number) => number;
7177
- readonly wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
7178
- readonly wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
7179
- readonly wasmsparseembeddingconfig_batchSize: (a: number) => number;
7180
- readonly wasmsparseembeddingconfig_acceleration: (a: number) => number;
7181
- readonly wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
7182
- readonly wasmdiffline_set_0: (a: number, b: number, c: number) => void;
7183
- readonly wasmdiffline_kind: (a: number) => [number, number];
7184
- readonly wasmdiffline_0: (a: number) => [number, number];
7185
- readonly __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
7186
- readonly wasmdocumentrelationship_target: (a: number) => number;
7187
+ readonly wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
7188
+ readonly wasmssrfpolicy_denyPrivate: (a: number) => number;
7189
+ readonly wasmpatternmatch_start: (a: number) => number;
7190
+ readonly wasmredactionfinding_end: (a: number) => number;
7191
+ readonly __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
7187
7192
  readonly wasmtablecell_set_rowSpan: (a: number, b: number) => void;
7188
7193
  readonly wasmtablecell_set_isHeader: (a: number, b: number) => void;
7189
7194
  readonly wasmtablecell_set_content: (a: number, b: number, c: number) => void;
@@ -7192,11 +7197,6 @@ export interface InitOutput {
7192
7197
  readonly wasmtablecell_isHeader: (a: number) => number;
7193
7198
  readonly wasmtablecell_content: (a: number) => [number, number];
7194
7199
  readonly wasmtablecell_colSpan: (a: number) => number;
7195
- readonly wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
7196
- readonly wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
7197
- readonly wasmocrrotation_confidence: (a: number) => [number, number];
7198
- readonly wasmocrrotation_angleDegrees: (a: number) => number;
7199
- readonly __wbg_wasmocrrotation_free: (a: number, b: number) => void;
7200
7200
  readonly __wbg_wasmsparseembeddingmodeltype_free: (a: number, b: number) => void;
7201
7201
  readonly wasmsparseembeddingmodeltype_set_type: (a: number, b: number, c: number) => void;
7202
7202
  readonly wasmsparseembeddingmodeltype_set_name: (a: number, b: number, c: 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_efc892689d7050b1: function(arg0, arg1) {
28127
+ __wbg_warn_4e6f746520c99d57: function(arg0, arg1) {
28128
28128
  console.warn(getStringFromWasm0(arg0, arg1));
28129
28129
  },
28130
28130
  __wbg_wasmarchiveentry_new: function(arg0) {
@@ -28556,12 +28556,12 @@ function __wbg_get_imports() {
28556
28556
  return ret;
28557
28557
  },
28558
28558
  __wbindgen_cast_0000000000000001: function(arg0, arg1) {
28559
- // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [Externref], shim_idx: 1689, ret: Result(Unit), inner_ret: Some(Result(Unit)) }, mutable: true }) -> Externref`.
28559
+ // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [Externref], shim_idx: 1708, ret: Result(Unit), inner_ret: Some(Result(Unit)) }, mutable: true }) -> Externref`.
28560
28560
  const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__hc53a447a5735fa3a);
28561
28561
  return ret;
28562
28562
  },
28563
28563
  __wbindgen_cast_0000000000000002: function(arg0, arg1) {
28564
- // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [], shim_idx: 4551, ret: Unit, inner_ret: Some(Unit) }, mutable: true }) -> Externref`.
28564
+ // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [], shim_idx: 4568, ret: Unit, inner_ret: Some(Unit) }, mutable: true }) -> Externref`.
28565
28565
  const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__h66e27dec856adeef);
28566
28566
  return ret;
28567
28567
  },
package/pkg/web/xberg_wasm_bg.wasm CHANGED
Binary file
package/pkg/web/xberg_wasm_bg.wasm.d.ts CHANGED
@@ -2419,8 +2419,8 @@ 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_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
2423
2422
  export const __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
2423
+ export const __wbg_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
2424
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];
@@ -2456,66 +2456,105 @@ export const wasmpageboundary_set_pageNumber: (a: number, b: number) => void;
2456
2456
  export const wasmpageboundary_set_byteStart: (a: number, b: number) => void;
2457
2457
  export const wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
2458
2458
  export const wasmpageboundary_byteStart: (a: number) => number;
2459
- export const __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
2460
- export const wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
2459
+ export const wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
2460
+ export const wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
2461
+ export const wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
2462
+ export const wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
2463
+ export const wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
2464
+ export const wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2461
2465
  export const wasmdbffieldinfo_name: (a: number) => [number, number];
2462
- export const wasmprocessingwarning_message: (a: number) => [number, number];
2466
+ export const wasmsupportedformat_mimeType: (a: number) => [number, number];
2467
+ export const wasmredactionpattern_pattern: (a: number) => [number, number];
2468
+ export const __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
2469
+ export const wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
2470
+ export const __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
2471
+ export const wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
2472
+ export const wasmredactionfinding_start: (a: number) => number;
2473
+ export const wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
2474
+ export const wasmdiffline_set_0: (a: number, b: number, c: number) => void;
2475
+ export const wasmdiffline_kind: (a: number) => [number, number];
2476
+ export const wasmdiffline_0: (a: number) => [number, number];
2477
+ export const wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
2478
+ export const wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
2479
+ export const wasmdbffieldinfo_fieldType: (a: number) => [number, number];
2480
+ export const __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
2481
+ export const __wbg_wasmpageboundary_free: (a: number, b: number) => void;
2482
+ export const wasmpageboundary_byteEnd: (a: number) => number;
2483
+ export const wasmsparseembeddingconfig_model: (a: number) => any;
2484
+ export const wasmsparseembeddingconfig_maxLength: (a: number) => number;
2485
+ export const wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
2486
+ export const wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
2487
+ export const wasmsparseembeddingconfig_batchSize: (a: number) => number;
2488
+ export const wasmsparseembeddingconfig_acceleration: (a: number) => number;
2463
2489
  export const wasmrerankerconfig_set_topK: (a: number, b: number) => void;
2464
2490
  export const wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
2465
2491
  export const wasmrerankerconfig_set_model: (a: number, b: any) => void;
2492
+ export const wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
2493
+ export const wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2466
2494
  export const wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
2467
2495
  export const wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
2468
2496
  export const wasmrerankerconfig_topK: (a: number) => number;
2469
2497
  export const wasmrerankerconfig_showDownloadProgress: (a: number) => number;
2470
2498
  export const wasmrerankerconfig_model: (a: number) => any;
2499
+ export const wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
2500
+ export const wasmrerankerconfig_cacheDir: (a: number) => [number, number];
2501
+ export const wasmrerankerconfig_batchSize: (a: number) => number;
2471
2502
  export const wasmrerankerconfig_acceleration: (a: number) => number;
2472
- export const wasmpreprocessingoptions_enabled: (a: number) => number;
2503
+ export const wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
2504
+ export const wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
2505
+ export const wasmocrrotation_angleDegrees: (a: number) => number;
2506
+ export const wasmocrrotation_confidence: (a: number) => [number, number];
2507
+ export const __wbg_wasmocrrotation_free: (a: number, b: number) => void;
2508
+ export const __wbg_wasmdjotlink_free: (a: number, b: number) => void;
2509
+ export const wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
2510
+ export const wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
2511
+ export const wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
2512
+ export const wasmredactionterm_value: (a: number) => [number, number];
2513
+ export const wasmredactionterm_label: (a: number) => [number, number];
2514
+ export const wasmredactionterm_caseSensitive: (a: number) => number;
2515
+ export const __wbg_wasmredactionterm_free: (a: number, b: number) => void;
2473
2516
  export const wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
2517
+ export const wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
2518
+ export const wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
2474
2519
  export const wasmdjotlink_url: (a: number) => [number, number];
2475
- export const wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
2476
- export const __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
2477
- export const wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
2478
- export const wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
2479
- export const wasmpageconfig_markerFormat: (a: number) => [number, number];
2480
- export const wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
2481
- export const __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
2482
- export const wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
2483
- export const wasmerrormetadata_message: (a: number) => [number, number];
2520
+ export const wasmdjotlink_title: (a: number) => [number, number];
2521
+ export const wasmdjotlink_text: (a: number) => [number, number];
2522
+ export const wasmocrelementconfig_minConfidence: (a: number) => number;
2523
+ export const wasmocrelementconfig_includeElements: (a: number) => number;
2524
+ export const wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
2525
+ export const wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
2526
+ export const wasmprocessingwarning_message: (a: number) => [number, number];
2484
2527
  export const __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
2485
- export const wasmredactionfinding_start: (a: number) => number;
2486
- export const wasmredactionfinding_end: (a: number) => number;
2487
- export const wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
2488
- export const wasmssrfpolicy_denyPrivate: (a: number) => number;
2489
- export const wasmpageboundary_byteEnd: (a: number) => number;
2528
+ export const wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
2529
+ export const wasmdocumentrelationship_target: (a: number) => number;
2490
2530
  export const wasmpageboundary_pageNumber: (a: number) => number;
2491
- export const __wbg_wasmpageboundary_free: (a: number, b: number) => void;
2531
+ export const __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
2492
2532
  export const wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
2493
- export const wasmpropertychange_name: (a: number) => [number, number];
2494
- export const wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
2495
2533
  export const wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
2496
- export const wasmpropertychange_to: (a: number) => [number, number];
2534
+ export const wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
2535
+ export const wasmpropertychange_name: (a: number) => [number, number];
2497
2536
  export const wasmpropertychange_from: (a: number) => [number, number];
2498
- export const wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
2499
- export const wasmrerankerconfig_cacheDir: (a: number) => [number, number];
2500
- export const wasmrerankerconfig_batchSize: (a: number) => number;
2501
- export const wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
2502
- export const wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2503
- export const wasmpatternmatch_start: (a: number) => number;
2537
+ export const wasmpropertychange_to: (a: number) => [number, number];
2538
+ export const wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
2539
+ export const wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
2540
+ export const wasmproxyconfig_username: (a: number) => [number, number];
2541
+ export const wasmproxyconfig_password: (a: number) => [number, number];
2542
+ export const __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
2543
+ export const wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
2544
+ export const wasmerrormetadata_message: (a: number) => [number, number];
2504
2545
  export const __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
2505
- export const wasmocrelementconfig_minConfidence: (a: number) => number;
2506
- export const wasmocrelementconfig_includeElements: (a: number) => number;
2507
- export const __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
2508
- export const wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
2509
- export const wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
2510
- export const wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
2511
- export const wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
2512
2546
  export const wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
2513
- export const wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
2514
- export const wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
2515
- export const wasmdbffieldinfo_fieldType: (a: number) => [number, number];
2516
- export const __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
2517
- export const wasmsupportedformat_mimeType: (a: number) => [number, number];
2518
- export const wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
2547
+ export const wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
2548
+ export const wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
2549
+ export const wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
2550
+ export const wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
2551
+ export const wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
2552
+ export const wasmpageconfig_markerFormat: (a: number) => [number, number];
2553
+ export const wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
2554
+ export const __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
2555
+ export const __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
2556
+ export const wasmpreprocessingoptions_enabled: (a: number) => number;
2557
+ export const __wbg_wasmpageconfig_free: (a: number, b: number) => void;
2519
2558
  export const wasmqrboundingbox_set_y: (a: number, b: number) => void;
2520
2559
  export const wasmqrboundingbox_set_x: (a: number, b: number) => void;
2521
2560
  export const wasmqrboundingbox_set_width: (a: number, b: number) => void;
@@ -2525,45 +2564,11 @@ export const wasmqrboundingbox_x: (a: number) => number;
2525
2564
  export const wasmqrboundingbox_width: (a: number) => number;
2526
2565
  export const wasmqrboundingbox_height: (a: number) => number;
2527
2566
  export const __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
2528
- export const wasmredactionpattern_pattern: (a: number) => [number, number];
2529
- export const __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
2530
- export const wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
2531
- export const __wbg_wasmpageconfig_free: (a: number, b: number) => void;
2532
- export const wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
2533
- export const wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
2534
- export const wasmdjotlink_title: (a: number) => [number, number];
2535
- export const wasmdjotlink_text: (a: number) => [number, number];
2536
- export const __wbg_wasmdjotlink_free: (a: number, b: number) => void;
2537
- export const wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
2538
- export const wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
2539
- export const wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
2540
- export const wasmredactionterm_value: (a: number) => [number, number];
2541
- export const wasmredactionterm_label: (a: number) => [number, number];
2542
- export const wasmredactionterm_caseSensitive: (a: number) => number;
2543
- export const __wbg_wasmredactionterm_free: (a: number, b: number) => void;
2544
- export const wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
2545
- export const wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
2546
- export const wasmproxyconfig_username: (a: number) => [number, number];
2547
- export const wasmproxyconfig_password: (a: number) => [number, number];
2548
- export const __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
2549
- export const wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
2550
- export const wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
2551
- export const wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
2552
- export const wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2553
- export const wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
2554
- export const wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
2555
- export const wasmsparseembeddingconfig_model: (a: number) => any;
2556
- export const wasmsparseembeddingconfig_maxLength: (a: number) => number;
2557
- export const wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
2558
- export const wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
2559
- export const wasmsparseembeddingconfig_batchSize: (a: number) => number;
2560
- export const wasmsparseembeddingconfig_acceleration: (a: number) => number;
2561
- export const wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
2562
- export const wasmdiffline_set_0: (a: number, b: number, c: number) => void;
2563
- export const wasmdiffline_kind: (a: number) => [number, number];
2564
- export const wasmdiffline_0: (a: number) => [number, number];
2565
- export const __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
2566
- export const wasmdocumentrelationship_target: (a: number) => number;
2567
+ export const wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
2568
+ export const wasmssrfpolicy_denyPrivate: (a: number) => number;
2569
+ export const wasmpatternmatch_start: (a: number) => number;
2570
+ export const wasmredactionfinding_end: (a: number) => number;
2571
+ export const __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
2567
2572
  export const wasmtablecell_set_rowSpan: (a: number, b: number) => void;
2568
2573
  export const wasmtablecell_set_isHeader: (a: number, b: number) => void;
2569
2574
  export const wasmtablecell_set_content: (a: number, b: number, c: number) => void;
@@ -2572,11 +2577,6 @@ export const wasmtablecell_rowSpan: (a: number) => number;
2572
2577
  export const wasmtablecell_isHeader: (a: number) => number;
2573
2578
  export const wasmtablecell_content: (a: number) => [number, number];
2574
2579
  export const wasmtablecell_colSpan: (a: number) => number;
2575
- export const wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
2576
- export const wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
2577
- export const wasmocrrotation_confidence: (a: number) => [number, number];
2578
- export const wasmocrrotation_angleDegrees: (a: number) => number;
2579
- export const __wbg_wasmocrrotation_free: (a: number, b: number) => void;
2580
2580
  export const __wbg_wasmsparseembeddingmodeltype_free: (a: number, b: number) => void;
2581
2581
  export const wasmsparseembeddingmodeltype_set_type: (a: number, b: number, c: number) => void;
2582
2582
  export const wasmsparseembeddingmodeltype_set_name: (a: number, b: number, c: number) => void;