@kreuzberg/html-to-markdown-wasm 3.6.0-rc.7 → 3.6.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/package.json +1 -1
- package/pkg/bundler/html_to_markdown_wasm.d.ts +74 -10
- package/pkg/bundler/html_to_markdown_wasm.js +1 -1
- package/pkg/bundler/html_to_markdown_wasm_bg.js +473 -79
- package/pkg/bundler/html_to_markdown_wasm_bg.wasm +0 -0
- package/pkg/bundler/html_to_markdown_wasm_bg.wasm.d.ts +35 -22
- package/pkg/bundler/package.json +1 -1
- package/pkg/deno/html_to_markdown_wasm.d.ts +74 -10
- package/pkg/deno/html_to_markdown_wasm.js +473 -79
- package/pkg/deno/html_to_markdown_wasm_bg.wasm +0 -0
- package/pkg/deno/html_to_markdown_wasm_bg.wasm.d.ts +35 -22
- package/pkg/nodejs/html_to_markdown_wasm.d.ts +74 -10
- package/pkg/nodejs/html_to_markdown_wasm.js +475 -79
- package/pkg/nodejs/html_to_markdown_wasm_bg.wasm +0 -0
- package/pkg/nodejs/html_to_markdown_wasm_bg.wasm.d.ts +35 -22
- package/pkg/nodejs/package.json +1 -1
- package/pkg/web/html_to_markdown_wasm.d.ts +109 -32
- package/pkg/web/html_to_markdown_wasm.js +473 -79
- package/pkg/web/html_to_markdown_wasm_bg.wasm +0 -0
- package/pkg/web/html_to_markdown_wasm_bg.wasm.d.ts +35 -22
- package/pkg/web/package.json +1 -1
package/package.json
CHANGED
|
@@ -225,7 +225,6 @@ export class WasmConversionResult {
|
|
|
225
225
|
set content(value: string | null | undefined);
|
|
226
226
|
get document(): WasmDocumentStructure | undefined;
|
|
227
227
|
set document(value: WasmDocumentStructure | null | undefined);
|
|
228
|
-
images: string[];
|
|
229
228
|
metadata: WasmHtmlMetadata;
|
|
230
229
|
tables: WasmTableData[];
|
|
231
230
|
warnings: WasmProcessingWarning[];
|
|
@@ -387,6 +386,23 @@ export class WasmHtmlMetadata {
|
|
|
387
386
|
structuredData: WasmStructuredData[];
|
|
388
387
|
}
|
|
389
388
|
|
|
389
|
+
/**
|
|
390
|
+
* Image dimensions in pixels.
|
|
391
|
+
*
|
|
392
|
+
* Binding-safe replacement for `(u32, u32)` tuples, which degrade to
|
|
393
|
+
* `Vec<Vec<String>>` when sanitized for cross-language binding generation.
|
|
394
|
+
* Used by both `ImageMetadata` and
|
|
395
|
+
* `InlineImage`.
|
|
396
|
+
*/
|
|
397
|
+
export class WasmImageDimensions {
|
|
398
|
+
free(): void;
|
|
399
|
+
[Symbol.dispose](): void;
|
|
400
|
+
static default(): WasmImageDimensions;
|
|
401
|
+
constructor(width: number, height: number);
|
|
402
|
+
height: number;
|
|
403
|
+
width: number;
|
|
404
|
+
}
|
|
405
|
+
|
|
390
406
|
/**
|
|
391
407
|
* Image metadata with source and dimensions.
|
|
392
408
|
*
|
|
@@ -399,12 +415,12 @@ export class WasmImageMetadata {
|
|
|
399
415
|
free(): void;
|
|
400
416
|
[Symbol.dispose](): void;
|
|
401
417
|
static default(): WasmImageMetadata;
|
|
402
|
-
constructor(src: string, imageType: WasmImageType, attributes: any, alt?: string | null, title?: string | null, dimensions?:
|
|
418
|
+
constructor(src: string, imageType: WasmImageType, attributes: any, alt?: string | null, title?: string | null, dimensions?: WasmImageDimensions | null);
|
|
403
419
|
get alt(): string | undefined;
|
|
404
420
|
set alt(value: string | null | undefined);
|
|
405
421
|
attributes: any;
|
|
406
|
-
get dimensions():
|
|
407
|
-
set dimensions(value:
|
|
422
|
+
get dimensions(): WasmImageDimensions | undefined;
|
|
423
|
+
set dimensions(value: WasmImageDimensions | null | undefined);
|
|
408
424
|
get imageType(): string;
|
|
409
425
|
set imageType(value: WasmImageType);
|
|
410
426
|
src: string;
|
|
@@ -481,6 +497,22 @@ export enum WasmListIndentType {
|
|
|
481
497
|
Tabs = 1,
|
|
482
498
|
}
|
|
483
499
|
|
|
500
|
+
/**
|
|
501
|
+
* A single key-value metadata entry from `<head>` meta tags.
|
|
502
|
+
*
|
|
503
|
+
* Binding-safe replacement for `(String, String)` tuples used in
|
|
504
|
+
* `NodeContent.MetadataBlock`. Tuple pairs cannot be represented
|
|
505
|
+
* across language boundaries without lossy degradation.
|
|
506
|
+
*/
|
|
507
|
+
export class WasmMetadataEntry {
|
|
508
|
+
free(): void;
|
|
509
|
+
[Symbol.dispose](): void;
|
|
510
|
+
static default(): WasmMetadataEntry;
|
|
511
|
+
constructor(key: string, value: string);
|
|
512
|
+
key: string;
|
|
513
|
+
value: string;
|
|
514
|
+
}
|
|
515
|
+
|
|
484
516
|
/**
|
|
485
517
|
* Line break syntax in Markdown output.
|
|
486
518
|
*
|
|
@@ -507,8 +539,8 @@ export class WasmNodeContent {
|
|
|
507
539
|
set definition(value: string | null | undefined);
|
|
508
540
|
get description(): string | undefined;
|
|
509
541
|
set description(value: string | null | undefined);
|
|
510
|
-
get entries():
|
|
511
|
-
set entries(value:
|
|
542
|
+
get entries(): WasmMetadataEntry[] | undefined;
|
|
543
|
+
set entries(value: WasmMetadataEntry[] | null | undefined);
|
|
512
544
|
get format(): string | undefined;
|
|
513
545
|
set format(value: string | null | undefined);
|
|
514
546
|
get grid(): WasmTableGrid | undefined;
|
|
@@ -540,14 +572,46 @@ export class WasmNodeContent {
|
|
|
540
572
|
* Context information passed to all visitor methods.
|
|
541
573
|
*
|
|
542
574
|
* Provides comprehensive metadata about the current node being visited,
|
|
543
|
-
* including its type,
|
|
575
|
+
* including its type, tag name, position in the DOM tree, and parent context.
|
|
576
|
+
*
|
|
577
|
+
* ## Attributes
|
|
578
|
+
*
|
|
579
|
+
* Access attributes via `NodeContext.attributes`, which returns
|
|
580
|
+
* `&BTreeMap<String, String>`. When the context was built with
|
|
581
|
+
* `NodeContext.with_lazy_attributes` (the hot path inside the converter),
|
|
582
|
+
* the map is only materialized on the first call — if the visitor never reads
|
|
583
|
+
* attributes, the allocation is skipped.
|
|
584
|
+
*
|
|
585
|
+
* ## Lifetimes
|
|
586
|
+
*
|
|
587
|
+
* String fields use `Cow<'_, str>` so the converter can pass slices directly
|
|
588
|
+
* out of the parsed DOM without allocating. Visitor implementations that need
|
|
589
|
+
* to outlive the callback should call `NodeContext.into_owned`.
|
|
544
590
|
*/
|
|
545
591
|
export class WasmNodeContext {
|
|
546
592
|
free(): void;
|
|
547
593
|
[Symbol.dispose](): void;
|
|
594
|
+
/**
|
|
595
|
+
* Return a reference to the attribute map.
|
|
596
|
+
*
|
|
597
|
+
* If the context was built with `NodeContext.with_lazy_attributes`, the
|
|
598
|
+
* map is materialized on the first call and cached for subsequent calls.
|
|
599
|
+
* If this method is never called, no allocation occurs for attributes.
|
|
600
|
+
*/
|
|
601
|
+
attributes(): any;
|
|
548
602
|
static default(): WasmNodeContext;
|
|
549
|
-
|
|
550
|
-
|
|
603
|
+
/**
|
|
604
|
+
* Promote any borrowed fields into owned storage so the context can outlive `'a`.
|
|
605
|
+
*/
|
|
606
|
+
intoOwned(): WasmNodeContext;
|
|
607
|
+
constructor(nodeType: WasmNodeType, tagName: string, depth: number, indexInParent: number, isInline: boolean, parentTag?: string | null);
|
|
608
|
+
/**
|
|
609
|
+
* Construct a `NodeContext` with an owned attribute map.
|
|
610
|
+
*
|
|
611
|
+
* Prefer `NodeContext.with_lazy_attributes` (pub(crate)) inside the
|
|
612
|
+
* converter to avoid the eager `collect_tag_attributes` allocation.
|
|
613
|
+
*/
|
|
614
|
+
static withOwnedAttributes(node_type: WasmNodeType, tag_name: string, attributes: any, depth: number, index_in_parent: number, parent_tag: string | null | undefined, is_inline: boolean): WasmNodeContext;
|
|
551
615
|
depth: number;
|
|
552
616
|
indexInParent: number;
|
|
553
617
|
isInline: boolean;
|
|
@@ -742,7 +806,7 @@ export class WasmProcessingWarning {
|
|
|
742
806
|
}
|
|
743
807
|
|
|
744
808
|
/**
|
|
745
|
-
* Structured data block (JSON-LD, Microdata, or RDFa).
|
|
809
|
+
* Structured data block (JSON-LD, Microdata, or `RDFa`).
|
|
746
810
|
*
|
|
747
811
|
* Represents machine-readable structured data found in the document.
|
|
748
812
|
* JSON-LD blocks are collected as raw JSON strings for flexibility.
|
|
@@ -5,5 +5,5 @@ import { __wbg_set_wasm } from "./html_to_markdown_wasm_bg.js";
|
|
|
5
5
|
__wbg_set_wasm(wasm);
|
|
6
6
|
|
|
7
7
|
export {
|
|
8
|
-
WasmAnnotationKind, WasmCodeBlockStyle, WasmConversionOptions, WasmConversionOptionsUpdate, WasmConversionResult, WasmDocumentMetadata, WasmDocumentNode, WasmDocumentStructure, WasmGridCell, WasmHeaderMetadata, WasmHeadingStyle, WasmHighlightStyle, WasmHtmlMetadata, WasmImageMetadata, WasmImageType, WasmLinkMetadata, WasmLinkStyle, WasmLinkType, WasmListIndentType, WasmNewlineStyle, WasmNodeContent, WasmNodeContext, WasmNodeType, WasmOutputFormat, WasmPreprocessingOptions, WasmPreprocessingOptionsUpdate, WasmPreprocessingPreset, WasmProcessingWarning, WasmStructuredData, WasmStructuredDataType, WasmTableData, WasmTableGrid, WasmTextAnnotation, WasmTextDirection, WasmTierStrategy, WasmUrlEscapeStyle, WasmVisitResult, WasmVisitorHandle, WasmWarningKind, WasmWhitespaceMode, convert
|
|
8
|
+
WasmAnnotationKind, WasmCodeBlockStyle, WasmConversionOptions, WasmConversionOptionsUpdate, WasmConversionResult, WasmDocumentMetadata, WasmDocumentNode, WasmDocumentStructure, WasmGridCell, WasmHeaderMetadata, WasmHeadingStyle, WasmHighlightStyle, WasmHtmlMetadata, WasmImageDimensions, WasmImageMetadata, WasmImageType, WasmLinkMetadata, WasmLinkStyle, WasmLinkType, WasmListIndentType, WasmMetadataEntry, WasmNewlineStyle, WasmNodeContent, WasmNodeContext, WasmNodeType, WasmOutputFormat, WasmPreprocessingOptions, WasmPreprocessingOptionsUpdate, WasmPreprocessingPreset, WasmProcessingWarning, WasmStructuredData, WasmStructuredDataType, WasmTableData, WasmTableGrid, WasmTextAnnotation, WasmTextDirection, WasmTierStrategy, WasmUrlEscapeStyle, WasmVisitResult, WasmVisitorHandle, WasmWarningKind, WasmWhitespaceMode, convert
|
|
9
9
|
} from "./html_to_markdown_wasm_bg.js";
|