@xberg-io/xberg-wasm 1.0.6 → 1.0.7

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 +50 -76
  2. package/package.json +1 -1
  3. package/pkg/web/README.md +50 -76
  4. package/pkg/web/package.json +1 -1
  5. package/pkg/web/xberg_wasm.d.ts +88 -88
  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 +88 -88
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 98+ 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 306 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
 
@@ -262,84 +262,58 @@ extractDocuments(fileBytes, mimes)
262
262
 
263
263
  ## Features
264
264
 
265
- ### Supported File Formats (98+)
266
-
267
- 98+ file formats across 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 |
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 |
287
279
  | **Vector** | `.svg` | DOM parsing, embedded text, graphics metadata |
288
280
 
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 (306 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
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
329
307
 
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
334
308
  - **OCR Support** - Integrate multiple OCR backends for scanned documents
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 [306 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
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
343
317
 
344
318
  ## OCR Support
345
319
 
@@ -483,7 +457,7 @@ Contributions are welcome! See [Contributing Guide](https://github.com/xberg-io/
483
457
 
484
458
  - [crawlberg](https://github.com/xberg-io/crawlberg) — web crawling and scraping with HTML→Markdown and headless-Chrome fallback.
485
459
  - [html-to-markdown](https://github.com/xberg-io/html-to-markdown) — fast, lossless HTML→Markdown engine.
486
- - [liter-llm](https://github.com/xberg-io/liter-llm) — universal LLM API client with native bindings for 14 languages and 143 providers.
460
+ - [liter-llm](https://github.com/xberg-io/liter-llm) — universal LLM API client with native bindings for 14 languages and 165 providers.
487
461
  - [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) — tree-sitter grammars and code-intelligence primitives.
488
462
  - [alef](https://github.com/xberg-io/alef) — the polyglot binding generator that produces this README and all per-language bindings.
489
463
  - [Discord](https://discord.gg/xt9WY3GnKR) — community, roadmap, announcements.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@xberg-io/xberg-wasm",
3
- "version": "1.0.6",
3
+ "version": "1.0.7",
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 98+ 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 306 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
 
@@ -262,84 +262,58 @@ extractDocuments(fileBytes, mimes)
262
262
 
263
263
  ## Features
264
264
 
265
- ### Supported File Formats (98+)
266
-
267
- 98+ file formats across 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 |
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 |
287
279
  | **Vector** | `.svg` | DOM parsing, embedded text, graphics metadata |
288
280
 
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 (306 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
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
329
307
 
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
334
308
  - **OCR Support** - Integrate multiple OCR backends for scanned documents
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 [306 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
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
343
317
 
344
318
  ## OCR Support
345
319
 
@@ -483,7 +457,7 @@ Contributions are welcome! See [Contributing Guide](https://github.com/xberg-io/
483
457
 
484
458
  - [crawlberg](https://github.com/xberg-io/crawlberg) — web crawling and scraping with HTML→Markdown and headless-Chrome fallback.
485
459
  - [html-to-markdown](https://github.com/xberg-io/html-to-markdown) — fast, lossless HTML→Markdown engine.
486
- - [liter-llm](https://github.com/xberg-io/liter-llm) — universal LLM API client with native bindings for 14 languages and 143 providers.
460
+ - [liter-llm](https://github.com/xberg-io/liter-llm) — universal LLM API client with native bindings for 14 languages and 165 providers.
487
461
  - [tree-sitter-language-pack](https://github.com/xberg-io/tree-sitter-language-pack) — tree-sitter grammars and code-intelligence primitives.
488
462
  - [alef](https://github.com/xberg-io/alef) — the polyglot binding generator that produces this README and all per-language bindings.
489
463
  - [Discord](https://discord.gg/xt9WY3GnKR) — community, roadmap, announcements.
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": "1.0.6",
5
+ "version": "1.0.7",
6
6
  "license": "MIT",
7
7
  "repository": {
8
8
  "type": "git",
package/pkg/web/xberg_wasm.d.ts CHANGED
@@ -7040,8 +7040,8 @@ export interface InitOutput {
7040
7040
  readonly wasmpagehierarchy_default: () => number;
7041
7041
  readonly wasmqrboundingbox_default: () => number;
7042
7042
  readonly __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
7043
- readonly __wbg_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
7044
7043
  readonly __wbg_wasmtablecell_free: (a: number, b: number) => void;
7044
+ readonly __wbg_wasmvlmfallbackpolicy_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];
@@ -7072,78 +7072,105 @@ 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 wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
7076
+ readonly wasmdiffline_set_0: (a: number, b: number, c: number) => void;
7077
+ readonly wasmsupportedformat_mimeType: (a: number) => [number, number];
7078
+ readonly wasmdiffline_kind: (a: number) => [number, number];
7079
+ readonly wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
7080
+ readonly wasmdjotlink_url: (a: number) => [number, number];
7081
+ readonly wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
7082
+ readonly wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
7083
+ readonly wasmssrfpolicy_denyPrivate: (a: number) => number;
7075
7084
  readonly wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
7076
7085
  readonly wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
7077
7086
  readonly wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
7078
7087
  readonly wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
7079
- readonly __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
7080
- readonly wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
7081
- readonly wasmpropertychange_name: (a: number) => [number, number];
7088
+ readonly wasmsparseembeddingconfig_model: (a: number) => any;
7089
+ readonly wasmsparseembeddingconfig_acceleration: (a: number) => number;
7082
7090
  readonly wasmocrelementconfig_minConfidence: (a: number) => number;
7083
7091
  readonly wasmocrelementconfig_includeElements: (a: number) => number;
7084
7092
  readonly wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
7085
7093
  readonly wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
7086
- readonly __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
7087
- readonly wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
7088
- readonly wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7089
- readonly wasmpageboundary_set_pageNumber: (a: number, b: number) => void;
7090
- readonly wasmpageboundary_set_byteStart: (a: number, b: number) => void;
7091
- readonly wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
7092
- readonly wasmpageboundary_byteStart: (a: number) => number;
7094
+ readonly wasmtablecell_set_rowSpan: (a: number, b: number) => void;
7095
+ readonly wasmtablecell_set_isHeader: (a: number, b: number) => void;
7096
+ readonly wasmtablecell_set_content: (a: number, b: number, c: number) => void;
7097
+ readonly wasmtablecell_set_colSpan: (a: number, b: number) => void;
7098
+ readonly wasmtablecell_isHeader: (a: number) => number;
7099
+ readonly wasmtablecell_content: (a: number) => [number, number];
7100
+ readonly wasmrerankerconfig_set_topK: (a: number, b: number) => void;
7101
+ readonly wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
7102
+ readonly wasmrerankerconfig_set_model: (a: number, b: any) => void;
7103
+ readonly wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
7104
+ readonly wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
7105
+ readonly wasmrerankerconfig_topK: (a: number) => number;
7106
+ readonly wasmrerankerconfig_showDownloadProgress: (a: number) => number;
7107
+ readonly wasmrerankerconfig_model: (a: number) => any;
7108
+ readonly wasmrerankerconfig_acceleration: (a: number) => number;
7109
+ readonly wasmtablecell_rowSpan: (a: number) => number;
7110
+ readonly wasmtablecell_colSpan: (a: number) => number;
7111
+ readonly wasmprocessingwarning_message: (a: number) => [number, number];
7112
+ readonly wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
7113
+ readonly wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
7093
7114
  readonly wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
7094
7115
  readonly wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
7095
7116
  readonly wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
7096
7117
  readonly wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
7097
- readonly wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
7098
- readonly wasmdiffline_set_0: (a: number, b: number, c: number) => void;
7099
- readonly wasmdiffline_kind: (a: number) => [number, number];
7100
- readonly wasmdiffline_0: (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 wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
7122
+ readonly wasmdbffieldinfo_name: (a: number) => [number, number];
7123
+ readonly wasmerrormetadata_message: (a: number) => [number, number];
7124
+ readonly __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
7125
+ readonly wasmqrboundingbox_set_y: (a: number, b: number) => void;
7126
+ readonly wasmqrboundingbox_set_x: (a: number, b: number) => void;
7127
+ readonly wasmqrboundingbox_set_width: (a: number, b: number) => void;
7128
+ readonly wasmqrboundingbox_set_height: (a: number, b: number) => void;
7129
+ readonly wasmqrboundingbox_x: (a: number) => number;
7130
+ readonly wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
7101
7131
  readonly wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
7102
7132
  readonly wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
7133
+ readonly wasmpropertychange_name: (a: number) => [number, number];
7103
7134
  readonly wasmproxyconfig_username: (a: number) => [number, number];
7104
7135
  readonly wasmproxyconfig_password: (a: number) => [number, number];
7136
+ readonly wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
7137
+ readonly wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
7138
+ readonly wasmdjotlink_title: (a: number) => [number, number];
7139
+ readonly wasmdjotlink_text: (a: number) => [number, number];
7140
+ readonly __wbg_wasmdjotlink_free: (a: number, b: number) => void;
7141
+ readonly wasmpatternmatch_start: (a: number) => number;
7142
+ readonly __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
7143
+ readonly __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
7144
+ readonly wasmpreprocessingoptions_enabled: (a: number) => number;
7145
+ readonly wasmpageboundary_set_pageNumber: (a: number, b: number) => void;
7146
+ readonly wasmpageboundary_set_byteStart: (a: number, b: number) => void;
7147
+ readonly wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
7148
+ readonly wasmpageboundary_byteStart: (a: number) => number;
7105
7149
  readonly wasmpageboundary_byteEnd: (a: number) => number;
7106
7150
  readonly wasmpageboundary_pageNumber: (a: number) => number;
7107
7151
  readonly __wbg_wasmpageboundary_free: (a: number, b: number) => void;
7108
- readonly wasmpreprocessingoptions_enabled: (a: number) => number;
7109
- readonly wasmsparseembeddingconfig_model: (a: number) => any;
7152
+ readonly wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
7153
+ readonly wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
7154
+ readonly wasmpropertychange_to: (a: number) => [number, number];
7155
+ readonly wasmpropertychange_from: (a: number) => [number, number];
7156
+ readonly __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
7157
+ readonly wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
7158
+ readonly wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7159
+ readonly __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
7160
+ readonly wasmdocumentrelationship_target: (a: number) => number;
7161
+ readonly wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
7162
+ readonly wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7110
7163
  readonly wasmsparseembeddingconfig_maxLength: (a: number) => number;
7111
7164
  readonly wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
7112
7165
  readonly wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
7113
7166
  readonly wasmsparseembeddingconfig_batchSize: (a: number) => number;
7114
- readonly wasmsparseembeddingconfig_acceleration: (a: number) => number;
7115
- readonly wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
7116
- readonly wasmprocessingwarning_message: (a: number) => [number, number];
7117
- readonly __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
7118
- readonly wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
7119
- readonly wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
7120
- readonly wasmerrormetadata_message: (a: number) => [number, number];
7121
- readonly __wbg_wasmdjotlink_free: (a: number, b: number) => void;
7122
- readonly wasmredactionpattern_pattern: (a: number) => [number, number];
7123
- readonly __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
7124
- readonly wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
7125
- readonly wasmtablecell_set_rowSpan: (a: number, b: number) => void;
7126
- readonly wasmtablecell_set_isHeader: (a: number, b: number) => void;
7127
- readonly wasmtablecell_set_content: (a: number, b: number, c: number) => void;
7128
- readonly wasmtablecell_set_colSpan: (a: number, b: number) => void;
7129
- readonly wasmtablecell_rowSpan: (a: number) => number;
7130
- readonly wasmtablecell_isHeader: (a: number) => number;
7131
- readonly wasmtablecell_content: (a: number) => [number, number];
7132
- readonly wasmtablecell_colSpan: (a: number) => number;
7133
- readonly wasmqrboundingbox_set_y: (a: number, b: number) => void;
7134
- readonly wasmqrboundingbox_set_x: (a: number, b: number) => void;
7135
- readonly wasmqrboundingbox_set_width: (a: number, b: number) => void;
7136
- readonly wasmqrboundingbox_set_height: (a: number, b: number) => void;
7137
- readonly wasmqrboundingbox_x: (a: number) => number;
7138
- readonly wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
7139
- readonly wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
7140
- readonly wasmpageconfig_markerFormat: (a: number) => [number, number];
7141
- readonly wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
7142
7167
  readonly __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
7143
- readonly wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
7144
- readonly wasmdbffieldinfo_name: (a: number) => [number, number];
7145
- readonly wasmdbffieldinfo_fieldType: (a: number) => [number, number];
7146
- readonly __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
7168
+ readonly __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
7169
+ readonly wasmqrboundingbox_y: (a: number) => number;
7170
+ readonly wasmqrboundingbox_width: (a: number) => number;
7171
+ readonly wasmqrboundingbox_height: (a: number) => number;
7172
+ readonly wasmdiffline_0: (a: number) => [number, number];
7173
+ readonly __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
7147
7174
  readonly wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
7148
7175
  readonly wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
7149
7176
  readonly wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
@@ -7151,52 +7178,25 @@ export interface InitOutput {
7151
7178
  readonly wasmredactionterm_label: (a: number) => [number, number];
7152
7179
  readonly wasmredactionterm_caseSensitive: (a: number) => number;
7153
7180
  readonly __wbg_wasmredactionterm_free: (a: number, b: number) => void;
7154
- readonly __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
7155
- readonly wasmqrboundingbox_y: (a: number) => number;
7156
- readonly wasmqrboundingbox_width: (a: number) => number;
7157
- readonly wasmqrboundingbox_height: (a: number) => number;
7158
- readonly wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
7159
- readonly wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
7160
- readonly wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
7161
- readonly wasmdjotlink_url: (a: number) => [number, number];
7162
- readonly wasmdjotlink_title: (a: number) => [number, number];
7163
- readonly wasmdjotlink_text: (a: number) => [number, number];
7164
- readonly wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
7165
- readonly wasmssrfpolicy_denyPrivate: (a: number) => number;
7166
- readonly wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
7167
- readonly wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
7168
- readonly wasmpropertychange_to: (a: number) => [number, number];
7169
- readonly wasmpropertychange_from: (a: number) => [number, number];
7170
- readonly __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
7171
- readonly wasmsupportedformat_mimeType: (a: number) => [number, number];
7172
- readonly wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
7181
+ readonly wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
7182
+ readonly wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
7183
+ readonly wasmpageconfig_markerFormat: (a: number) => [number, number];
7184
+ readonly wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
7173
7185
  readonly __wbg_wasmpageconfig_free: (a: number, b: number) => void;
7174
7186
  readonly wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
7175
- readonly wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
7176
- readonly wasmocrrotation_confidence: (a: number) => [number, number];
7177
7187
  readonly wasmocrrotation_angleDegrees: (a: number) => number;
7188
+ readonly wasmocrrotation_confidence: (a: number) => [number, number];
7178
7189
  readonly __wbg_wasmocrrotation_free: (a: number, b: number) => void;
7190
+ readonly wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
7191
+ readonly wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
7192
+ readonly wasmdbffieldinfo_fieldType: (a: number) => [number, number];
7193
+ readonly __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
7179
7194
  readonly wasmredactionfinding_start: (a: number) => number;
7180
7195
  readonly wasmredactionfinding_end: (a: number) => number;
7181
7196
  readonly __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
7182
- readonly wasmrerankerconfig_set_topK: (a: number, b: number) => void;
7183
- readonly wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
7184
- readonly wasmrerankerconfig_set_model: (a: number, b: any) => void;
7185
- readonly wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
7186
- readonly wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
7187
- readonly wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
7188
- readonly wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
7189
- readonly wasmrerankerconfig_topK: (a: number) => number;
7190
- readonly wasmrerankerconfig_showDownloadProgress: (a: number) => number;
7191
- readonly wasmrerankerconfig_model: (a: number) => any;
7192
- readonly wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
7193
- readonly wasmrerankerconfig_cacheDir: (a: number) => [number, number];
7194
- readonly wasmrerankerconfig_batchSize: (a: number) => number;
7195
- readonly wasmrerankerconfig_acceleration: (a: number) => number;
7196
- readonly __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
7197
- readonly wasmdocumentrelationship_target: (a: number) => number;
7198
- readonly wasmpatternmatch_start: (a: number) => number;
7199
- readonly __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
7197
+ readonly wasmredactionpattern_pattern: (a: number) => [number, number];
7198
+ readonly __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
7199
+ readonly wasmredactionpattern_set_pattern: (a: number, b: number, c: 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_58d677af4b34c028: function(arg0, arg1) {
28127
+ __wbg_warn_f6f2ddda37728ad7: 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
@@ -2420,8 +2420,8 @@ export const wasmprocessingwarning_new: (a: number, b: number, c: number, d: num
2420
2420
  export const wasmpagehierarchy_default: () => number;
2421
2421
  export const wasmqrboundingbox_default: () => number;
2422
2422
  export const __wbg_wasmsparseembeddingconfig_free: (a: number, b: number) => void;
2423
- export const __wbg_wasmvlmfallbackpolicy_free: (a: number, b: number) => void;
2424
2423
  export const __wbg_wasmtablecell_free: (a: number, b: number) => void;
2424
+ export const __wbg_wasmvlmfallbackpolicy_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];
@@ -2452,78 +2452,105 @@ 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 wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
2456
+ export const wasmdiffline_set_0: (a: number, b: number, c: number) => void;
2457
+ export const wasmsupportedformat_mimeType: (a: number) => [number, number];
2458
+ export const wasmdiffline_kind: (a: number) => [number, number];
2459
+ export const wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
2460
+ export const wasmdjotlink_url: (a: number) => [number, number];
2461
+ export const wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
2462
+ export const wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
2463
+ export const wasmssrfpolicy_denyPrivate: (a: number) => number;
2455
2464
  export const wasmsparseembeddingconfig_set_model: (a: number, b: any) => void;
2456
2465
  export const wasmsparseembeddingconfig_set_maxLength: (a: number, b: number) => void;
2457
2466
  export const wasmsparseembeddingconfig_set_batchSize: (a: number, b: number) => void;
2458
2467
  export const wasmsparseembeddingconfig_set_acceleration: (a: number, b: number) => void;
2459
- export const __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
2460
- export const wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
2461
- export const wasmpropertychange_name: (a: number) => [number, number];
2468
+ export const wasmsparseembeddingconfig_model: (a: number) => any;
2469
+ export const wasmsparseembeddingconfig_acceleration: (a: number) => number;
2462
2470
  export const wasmocrelementconfig_minConfidence: (a: number) => number;
2463
2471
  export const wasmocrelementconfig_includeElements: (a: number) => number;
2464
2472
  export const wasmocrelementconfig_set_minConfidence: (a: number, b: number) => void;
2465
2473
  export const wasmocrelementconfig_set_includeElements: (a: number, b: number) => void;
2466
- export const __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
2467
- export const wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
2468
- export const wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2469
- export const wasmpageboundary_set_pageNumber: (a: number, b: number) => void;
2470
- export const wasmpageboundary_set_byteStart: (a: number, b: number) => void;
2471
- export const wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
2472
- export const wasmpageboundary_byteStart: (a: number) => number;
2474
+ export const wasmtablecell_set_rowSpan: (a: number, b: number) => void;
2475
+ export const wasmtablecell_set_isHeader: (a: number, b: number) => void;
2476
+ export const wasmtablecell_set_content: (a: number, b: number, c: number) => void;
2477
+ export const wasmtablecell_set_colSpan: (a: number, b: number) => void;
2478
+ export const wasmtablecell_isHeader: (a: number) => number;
2479
+ export const wasmtablecell_content: (a: number) => [number, number];
2480
+ export const wasmrerankerconfig_set_topK: (a: number, b: number) => void;
2481
+ export const wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
2482
+ export const wasmrerankerconfig_set_model: (a: number, b: any) => void;
2483
+ export const wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
2484
+ export const wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
2485
+ export const wasmrerankerconfig_topK: (a: number) => number;
2486
+ export const wasmrerankerconfig_showDownloadProgress: (a: number) => number;
2487
+ export const wasmrerankerconfig_model: (a: number) => any;
2488
+ export const wasmrerankerconfig_acceleration: (a: number) => number;
2489
+ export const wasmtablecell_rowSpan: (a: number) => number;
2490
+ export const wasmtablecell_colSpan: (a: number) => number;
2491
+ export const wasmprocessingwarning_message: (a: number) => [number, number];
2492
+ export const wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
2493
+ export const wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
2473
2494
  export const wasmvlmfallbackpolicy_set_mode: (a: number, b: number, c: number) => void;
2474
2495
  export const wasmvlmfallbackpolicy_set_qualityThreshold: (a: number, b: number, c: number) => void;
2475
2496
  export const wasmvlmfallbackpolicy_mode: (a: number) => [number, number];
2476
2497
  export const wasmvlmfallbackpolicy_qualityThreshold: (a: number) => [number, number];
2477
- export const wasmdiffline_set_kind: (a: number, b: number, c: number) => void;
2478
- export const wasmdiffline_set_0: (a: number, b: number, c: number) => void;
2479
- export const wasmdiffline_kind: (a: number) => [number, number];
2480
- export const wasmdiffline_0: (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 wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
2502
+ export const wasmdbffieldinfo_name: (a: number) => [number, number];
2503
+ export const wasmerrormetadata_message: (a: number) => [number, number];
2504
+ export const __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
2505
+ export const wasmqrboundingbox_set_y: (a: number, b: number) => void;
2506
+ export const wasmqrboundingbox_set_x: (a: number, b: number) => void;
2507
+ export const wasmqrboundingbox_set_width: (a: number, b: number) => void;
2508
+ export const wasmqrboundingbox_set_height: (a: number, b: number) => void;
2509
+ export const wasmqrboundingbox_x: (a: number) => number;
2510
+ export const wasmpropertychange_set_name: (a: number, b: number, c: number) => void;
2481
2511
  export const wasmproxyconfig_set_username: (a: number, b: number, c: number) => void;
2482
2512
  export const wasmproxyconfig_set_password: (a: number, b: number, c: number) => void;
2513
+ export const wasmpropertychange_name: (a: number) => [number, number];
2483
2514
  export const wasmproxyconfig_username: (a: number) => [number, number];
2484
2515
  export const wasmproxyconfig_password: (a: number) => [number, number];
2516
+ export const wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
2517
+ export const wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
2518
+ export const wasmdjotlink_title: (a: number) => [number, number];
2519
+ export const wasmdjotlink_text: (a: number) => [number, number];
2520
+ export const __wbg_wasmdjotlink_free: (a: number, b: number) => void;
2521
+ export const wasmpatternmatch_start: (a: number) => number;
2522
+ export const __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
2523
+ export const __wbg_wasmpreprocessingoptions_free: (a: number, b: number) => void;
2524
+ export const wasmpreprocessingoptions_enabled: (a: number) => number;
2525
+ export const wasmpageboundary_set_pageNumber: (a: number, b: number) => void;
2526
+ export const wasmpageboundary_set_byteStart: (a: number, b: number) => void;
2527
+ export const wasmpageboundary_set_byteEnd: (a: number, b: number) => void;
2528
+ export const wasmpageboundary_byteStart: (a: number) => number;
2485
2529
  export const wasmpageboundary_byteEnd: (a: number) => number;
2486
2530
  export const wasmpageboundary_pageNumber: (a: number) => number;
2487
2531
  export const __wbg_wasmpageboundary_free: (a: number, b: number) => void;
2488
- export const wasmpreprocessingoptions_enabled: (a: number) => number;
2489
- export const wasmsparseembeddingconfig_model: (a: number) => any;
2532
+ export const wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
2533
+ export const wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
2534
+ export const wasmpropertychange_to: (a: number) => [number, number];
2535
+ export const wasmpropertychange_from: (a: number) => [number, number];
2536
+ export const __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
2537
+ export const wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
2538
+ export const wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2539
+ export const __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
2540
+ export const wasmdocumentrelationship_target: (a: number) => number;
2541
+ export const wasmsparseembeddingconfig_set_maxEmbedDurationSecs: (a: number, b: number, c: bigint) => void;
2542
+ export const wasmsparseembeddingconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2490
2543
  export const wasmsparseembeddingconfig_maxLength: (a: number) => number;
2491
2544
  export const wasmsparseembeddingconfig_maxEmbedDurationSecs: (a: number) => [number, bigint];
2492
2545
  export const wasmsparseembeddingconfig_cacheDir: (a: number) => [number, number];
2493
2546
  export const wasmsparseembeddingconfig_batchSize: (a: number) => number;
2494
- export const wasmsparseembeddingconfig_acceleration: (a: number) => number;
2495
- export const wasmprocessingwarning_set_message: (a: number, b: number, c: number) => void;
2496
- export const wasmprocessingwarning_message: (a: number) => [number, number];
2497
- export const __wbg_wasmprocessingwarning_free: (a: number, b: number) => void;
2498
- export const wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
2499
- export const wasmerrormetadata_set_message: (a: number, b: number, c: number) => void;
2500
- export const wasmerrormetadata_message: (a: number) => [number, number];
2501
- export const __wbg_wasmdjotlink_free: (a: number, b: number) => void;
2502
- export const wasmredactionpattern_pattern: (a: number) => [number, number];
2503
- export const __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
2504
- export const wasmredactionpattern_set_pattern: (a: number, b: number, c: number) => void;
2505
- export const wasmtablecell_set_rowSpan: (a: number, b: number) => void;
2506
- export const wasmtablecell_set_isHeader: (a: number, b: number) => void;
2507
- export const wasmtablecell_set_content: (a: number, b: number, c: number) => void;
2508
- export const wasmtablecell_set_colSpan: (a: number, b: number) => void;
2509
- export const wasmtablecell_rowSpan: (a: number) => number;
2510
- export const wasmtablecell_isHeader: (a: number) => number;
2511
- export const wasmtablecell_content: (a: number) => [number, number];
2512
- export const wasmtablecell_colSpan: (a: number) => number;
2513
- export const wasmqrboundingbox_set_y: (a: number, b: number) => void;
2514
- export const wasmqrboundingbox_set_x: (a: number, b: number) => void;
2515
- export const wasmqrboundingbox_set_width: (a: number, b: number) => void;
2516
- export const wasmqrboundingbox_set_height: (a: number, b: number) => void;
2517
- export const wasmqrboundingbox_x: (a: number) => number;
2518
- export const wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
2519
- export const wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
2520
- export const wasmpageconfig_markerFormat: (a: number) => [number, number];
2521
- export const wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
2522
2547
  export const __wbg_wasmtokenreductionoptions_free: (a: number, b: number) => void;
2523
- export const wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
2524
- export const wasmdbffieldinfo_name: (a: number) => [number, number];
2525
- export const wasmdbffieldinfo_fieldType: (a: number) => [number, number];
2526
- export const __wbg_wasmerrormetadata_free: (a: number, b: number) => void;
2548
+ export const __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
2549
+ export const wasmqrboundingbox_y: (a: number) => number;
2550
+ export const wasmqrboundingbox_width: (a: number) => number;
2551
+ export const wasmqrboundingbox_height: (a: number) => number;
2552
+ export const wasmdiffline_0: (a: number) => [number, number];
2553
+ export const __wbg_wasmocrelementconfig_free: (a: number, b: number) => void;
2527
2554
  export const wasmredactionterm_set_value: (a: number, b: number, c: number) => void;
2528
2555
  export const wasmredactionterm_set_label: (a: number, b: number, c: number) => void;
2529
2556
  export const wasmredactionterm_set_caseSensitive: (a: number, b: number) => void;
@@ -2531,52 +2558,25 @@ export const wasmredactionterm_value: (a: number) => [number, number];
2531
2558
  export const wasmredactionterm_label: (a: number) => [number, number];
2532
2559
  export const wasmredactionterm_caseSensitive: (a: number) => number;
2533
2560
  export const __wbg_wasmredactionterm_free: (a: number, b: number) => void;
2534
- export const __wbg_wasmqrboundingbox_free: (a: number, b: number) => void;
2535
- export const wasmqrboundingbox_y: (a: number) => number;
2536
- export const wasmqrboundingbox_width: (a: number) => number;
2537
- export const wasmqrboundingbox_height: (a: number) => number;
2538
- export const wasmdjotlink_set_url: (a: number, b: number, c: number) => void;
2539
- export const wasmdjotlink_set_title: (a: number, b: number, c: number) => void;
2540
- export const wasmdjotlink_set_text: (a: number, b: number, c: number) => void;
2541
- export const wasmdjotlink_url: (a: number) => [number, number];
2542
- export const wasmdjotlink_title: (a: number) => [number, number];
2543
- export const wasmdjotlink_text: (a: number) => [number, number];
2544
- export const wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
2545
- export const wasmssrfpolicy_denyPrivate: (a: number) => number;
2546
- export const wasmpropertychange_set_to: (a: number, b: number, c: number) => void;
2547
- export const wasmpropertychange_set_from: (a: number, b: number, c: number) => void;
2548
- export const wasmpropertychange_to: (a: number) => [number, number];
2549
- export const wasmpropertychange_from: (a: number) => [number, number];
2550
- export const __wbg_wasmproxyconfig_free: (a: number, b: number) => void;
2551
- export const wasmsupportedformat_mimeType: (a: number) => [number, number];
2552
- export const wasmsupportedformat_set_mimeType: (a: number, b: number, c: number) => void;
2561
+ export const wasmpageconfig_set_markerFormat: (a: number, b: number, c: number) => void;
2562
+ export const wasmtokenreductionoptions_set_preserveImportantWords: (a: number, b: number) => void;
2563
+ export const wasmpageconfig_markerFormat: (a: number) => [number, number];
2564
+ export const wasmtokenreductionoptions_preserveImportantWords: (a: number) => number;
2553
2565
  export const __wbg_wasmpageconfig_free: (a: number, b: number) => void;
2554
2566
  export const wasmocrrotation_set_confidence: (a: number, b: number, c: number) => void;
2555
- export const wasmocrrotation_set_angleDegrees: (a: number, b: number) => void;
2556
- export const wasmocrrotation_confidence: (a: number) => [number, number];
2557
2567
  export const wasmocrrotation_angleDegrees: (a: number) => number;
2568
+ export const wasmocrrotation_confidence: (a: number) => [number, number];
2558
2569
  export const __wbg_wasmocrrotation_free: (a: number, b: number) => void;
2570
+ export const wasmdbffieldinfo_set_name: (a: number, b: number, c: number) => void;
2571
+ export const wasmdbffieldinfo_set_fieldType: (a: number, b: number, c: number) => void;
2572
+ export const wasmdbffieldinfo_fieldType: (a: number) => [number, number];
2573
+ export const __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
2559
2574
  export const wasmredactionfinding_start: (a: number) => number;
2560
2575
  export const wasmredactionfinding_end: (a: number) => number;
2561
2576
  export const __wbg_wasmredactionfinding_free: (a: number, b: number) => void;
2562
- export const wasmrerankerconfig_set_topK: (a: number, b: number) => void;
2563
- export const wasmrerankerconfig_set_showDownloadProgress: (a: number, b: number) => void;
2564
- export const wasmrerankerconfig_set_model: (a: number, b: any) => void;
2565
- export const wasmrerankerconfig_set_maxRerankDurationSecs: (a: number, b: number, c: bigint) => void;
2566
- export const wasmrerankerconfig_set_cacheDir: (a: number, b: number, c: number) => void;
2567
- export const wasmrerankerconfig_set_batchSize: (a: number, b: number) => void;
2568
- export const wasmrerankerconfig_set_acceleration: (a: number, b: number) => void;
2569
- export const wasmrerankerconfig_topK: (a: number) => number;
2570
- export const wasmrerankerconfig_showDownloadProgress: (a: number) => number;
2571
- export const wasmrerankerconfig_model: (a: number) => any;
2572
- export const wasmrerankerconfig_maxRerankDurationSecs: (a: number) => [number, bigint];
2573
- export const wasmrerankerconfig_cacheDir: (a: number) => [number, number];
2574
- export const wasmrerankerconfig_batchSize: (a: number) => number;
2575
- export const wasmrerankerconfig_acceleration: (a: number) => number;
2576
- export const __wbg_wasmdocumentrelationship_free: (a: number, b: number) => void;
2577
- export const wasmdocumentrelationship_target: (a: number) => number;
2578
- export const wasmpatternmatch_start: (a: number) => number;
2579
- export const __wbg_wasmdbffieldinfo_free: (a: number, b: number) => void;
2577
+ export const wasmredactionpattern_pattern: (a: number) => [number, number];
2578
+ export const __wbg_wasmredactionpattern_free: (a: number, b: number) => void;
2579
+ export const wasmredactionpattern_set_pattern: (a: number, b: number, c: 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;