html-to-markdown 3.4.0.pre.rc.45-aarch64-linux → 3.4.1-aarch64-linux
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.
- checksums.yaml +4 -4
- data/ext/html_to_markdown_rb/Cargo.toml +23 -0
- data/ext/html_to_markdown_rb/Makefile +592 -0
- data/ext/html_to_markdown_rb/extconf.rb +11 -0
- data/ext/html_to_markdown_rb/native/Cargo.lock +943 -0
- data/ext/html_to_markdown_rb/native/Cargo.toml +31 -0
- data/ext/html_to_markdown_rb/src/lib.rs +6833 -0
- data/lib/html_to_markdown/native.rb +3 -1
- data/lib/html_to_markdown/version.rb +2 -2
- data/lib/html_to_markdown.rb +1 -1
- data/lib/html_to_markdown_rb.so +0 -0
- data/sig/types.rbs +20 -18
- metadata +27 -5
- data/lib/bin/html-to-markdown +0 -0
|
@@ -1,5 +1,5 @@
|
|
|
1
1
|
# This file is auto-generated by alef — DO NOT EDIT.
|
|
2
|
-
# alef:hash:
|
|
2
|
+
# alef:hash:8104ea3e0a2d7ef26ed519cf9d400a92c4ae7564f726fb32ed02076450a2d277
|
|
3
3
|
# To regenerate: alef generate
|
|
4
4
|
# To verify freshness: alef verify --exit-code
|
|
5
5
|
# Issues & docs: https://github.com/kreuzberg-dev/alef
|
|
@@ -23,6 +23,7 @@ end
|
|
|
23
23
|
class Hash
|
|
24
24
|
# Support internally-tagged enum accessors like format.excel, format.email, etc.
|
|
25
25
|
# Also support direct field access like format.sheet_count
|
|
26
|
+
# rubocop:disable Metrics/CyclomaticComplexity
|
|
26
27
|
def method_missing(method_name, *args, &block)
|
|
27
28
|
# Try symbol key first (how Magnus converts JSON keys)
|
|
28
29
|
return self[method_name] if key?(method_name)
|
|
@@ -44,6 +45,7 @@ class Hash
|
|
|
44
45
|
|
|
45
46
|
super
|
|
46
47
|
end
|
|
48
|
+
# rubocop:enable Metrics/CyclomaticComplexity
|
|
47
49
|
|
|
48
50
|
def respond_to_missing?(method_name, include_private = false)
|
|
49
51
|
return true if key?(method_name) || key?(method_name.to_s)
|
|
@@ -1,10 +1,10 @@
|
|
|
1
1
|
# This file is auto-generated by alef — DO NOT EDIT.
|
|
2
|
-
# alef:hash:
|
|
2
|
+
# alef:hash:ee219b60565c49b67427d4a5c825fcafe3db1061d9c81a06de306dd72495b28b
|
|
3
3
|
# To regenerate: alef generate
|
|
4
4
|
# To verify freshness: alef verify --exit-code
|
|
5
5
|
# Issues & docs: https://github.com/kreuzberg-dev/alef
|
|
6
6
|
# frozen_string_literal: true
|
|
7
7
|
|
|
8
8
|
module HtmlToMarkdown
|
|
9
|
-
VERSION = '3.4.
|
|
9
|
+
VERSION = '3.4.1'
|
|
10
10
|
end
|
data/lib/html_to_markdown.rb
CHANGED
|
@@ -1,5 +1,5 @@
|
|
|
1
1
|
# This file is auto-generated by alef — DO NOT EDIT.
|
|
2
|
-
# alef:hash:
|
|
2
|
+
# alef:hash:8697ef2add7937bd1fef3fb3a54b34edd14a9764276d0e9f4a53aab7ce9499e0
|
|
3
3
|
# To regenerate: alef generate
|
|
4
4
|
# To verify freshness: alef verify --exit-code
|
|
5
5
|
# Issues & docs: https://github.com/kreuzberg-dev/alef
|
data/lib/html_to_markdown_rb.so
CHANGED
|
Binary file
|
data/sig/types.rbs
CHANGED
|
@@ -1,5 +1,5 @@
|
|
|
1
1
|
# This file is auto-generated by alef — DO NOT EDIT.
|
|
2
|
-
# alef:hash:
|
|
2
|
+
# alef:hash:7e5aa912785759cbcad9e2b4750726cbd2e1d7d0717eb2078ffde17d2dd6344f
|
|
3
3
|
# To regenerate: alef generate
|
|
4
4
|
# To verify freshness: alef verify --exit-code
|
|
5
5
|
# Issues & docs: https://github.com/kreuzberg-dev/alef
|
|
@@ -456,9 +456,11 @@ module HtmlToMarkdown
|
|
|
456
456
|
end
|
|
457
457
|
|
|
458
458
|
class VisitorHandle
|
|
459
|
-
# Type alias for a visitor handle (
|
|
459
|
+
# Type alias for a visitor handle (`Arc`-wrapped `Mutex` for thread-safe shared mutation).
|
|
460
460
|
#
|
|
461
|
-
#
|
|
461
|
+
# `Send + Sync` so that types embedding a `VisitorHandle` (e.g. `ConversionOptions`)
|
|
462
|
+
# can be shared across threads — required by callers that stash configs inside
|
|
463
|
+
# axum/rmcp/tokio Send-bound contexts.
|
|
462
464
|
|
|
463
465
|
end
|
|
464
466
|
|
|
@@ -483,77 +485,77 @@ module HtmlToMarkdown
|
|
|
483
485
|
# Text directionality of document content.
|
|
484
486
|
#
|
|
485
487
|
# Corresponds to the HTML `dir` attribute and `bdi` element directionality.
|
|
486
|
-
type
|
|
488
|
+
type value = :left_to_right | :right_to_left | :auto
|
|
487
489
|
end
|
|
488
490
|
|
|
489
491
|
class LinkType
|
|
490
492
|
# Link classification based on href value and document context.
|
|
491
493
|
#
|
|
492
494
|
# Used to categorize links during extraction for filtering and analysis.
|
|
493
|
-
type
|
|
495
|
+
type value = :anchor | :internal | :external | :email | :phone | :other
|
|
494
496
|
end
|
|
495
497
|
|
|
496
498
|
class ImageType
|
|
497
499
|
# Image source classification for proper handling and processing.
|
|
498
500
|
#
|
|
499
501
|
# Determines whether an image is embedded (data URI), inline SVG, external, or relative.
|
|
500
|
-
type
|
|
502
|
+
type value = :data_uri | :inline_svg | :external | :relative
|
|
501
503
|
end
|
|
502
504
|
|
|
503
505
|
class StructuredDataType
|
|
504
506
|
# Structured data format type.
|
|
505
507
|
#
|
|
506
508
|
# Identifies the schema/format used for structured data markup.
|
|
507
|
-
type
|
|
509
|
+
type value = :json_ld | :microdata | :r_d_fa
|
|
508
510
|
end
|
|
509
511
|
|
|
510
512
|
class PreprocessingPreset
|
|
511
513
|
# HTML preprocessing aggressiveness level.
|
|
512
514
|
#
|
|
513
515
|
# Controls the extent of cleanup performed before conversion. Higher levels remove more elements.
|
|
514
|
-
type
|
|
516
|
+
type value = :minimal | :standard | :aggressive
|
|
515
517
|
end
|
|
516
518
|
|
|
517
519
|
class HeadingStyle
|
|
518
520
|
# Heading style options for Markdown output.
|
|
519
521
|
#
|
|
520
522
|
# Controls how headings (h1-h6) are rendered in the output Markdown.
|
|
521
|
-
type
|
|
523
|
+
type value = :underlined | :atx | :atx_closed
|
|
522
524
|
end
|
|
523
525
|
|
|
524
526
|
class ListIndentType
|
|
525
527
|
# List indentation character type.
|
|
526
528
|
#
|
|
527
529
|
# Controls whether list items are indented with spaces or tabs.
|
|
528
|
-
type
|
|
530
|
+
type value = :spaces | :tabs
|
|
529
531
|
end
|
|
530
532
|
|
|
531
533
|
class WhitespaceMode
|
|
532
534
|
# Whitespace handling strategy during conversion.
|
|
533
535
|
#
|
|
534
536
|
# Determines how sequences of whitespace characters (spaces, tabs, newlines) are processed.
|
|
535
|
-
type
|
|
537
|
+
type value = :normalized | :strict
|
|
536
538
|
end
|
|
537
539
|
|
|
538
540
|
class NewlineStyle
|
|
539
541
|
# Line break syntax in Markdown output.
|
|
540
542
|
#
|
|
541
543
|
# Controls how soft line breaks (from `<br>` or line breaks in source) are rendered.
|
|
542
|
-
type
|
|
544
|
+
type value = :spaces | :backslash
|
|
543
545
|
end
|
|
544
546
|
|
|
545
547
|
class CodeBlockStyle
|
|
546
548
|
# Code block fence style in Markdown output.
|
|
547
549
|
#
|
|
548
550
|
# Determines how code blocks (`<pre><code>`) are rendered in Markdown.
|
|
549
|
-
type
|
|
551
|
+
type value = :indented | :backticks | :tildes
|
|
550
552
|
end
|
|
551
553
|
|
|
552
554
|
class HighlightStyle
|
|
553
555
|
# Highlight rendering style for `<mark>` elements.
|
|
554
556
|
#
|
|
555
557
|
# Controls how highlighted text is rendered in Markdown output.
|
|
556
|
-
type
|
|
558
|
+
type value = :double_equal | :html | :bold | :none
|
|
557
559
|
end
|
|
558
560
|
|
|
559
561
|
class LinkStyle
|
|
@@ -561,14 +563,14 @@ module HtmlToMarkdown
|
|
|
561
563
|
#
|
|
562
564
|
# Controls whether links and images use inline `[text](url)` syntax or
|
|
563
565
|
# reference-style `[text][1]` syntax with definitions collected at the end.
|
|
564
|
-
type
|
|
566
|
+
type value = :inline | :reference
|
|
565
567
|
end
|
|
566
568
|
|
|
567
569
|
class OutputFormat
|
|
568
570
|
# Output format for conversion.
|
|
569
571
|
#
|
|
570
572
|
# Specifies the target markup language format for the conversion output.
|
|
571
|
-
type
|
|
573
|
+
type value = :markdown | :djot | :plain
|
|
572
574
|
end
|
|
573
575
|
|
|
574
576
|
class NodeContent
|
|
@@ -585,7 +587,7 @@ module HtmlToMarkdown
|
|
|
585
587
|
|
|
586
588
|
class WarningKind
|
|
587
589
|
# Categories of processing warnings.
|
|
588
|
-
type
|
|
590
|
+
type value = :image_extraction_failed | :encoding_fallback | :truncated_input | :malformed_html | :sanitization_applied | :depth_limit_exceeded
|
|
589
591
|
end
|
|
590
592
|
|
|
591
593
|
class NodeType
|
|
@@ -593,7 +595,7 @@ module HtmlToMarkdown
|
|
|
593
595
|
#
|
|
594
596
|
# This enum categorizes all HTML elements that the converter recognizes,
|
|
595
597
|
# providing a coarse-grained classification for visitor dispatch.
|
|
596
|
-
type
|
|
598
|
+
type value = :text | :element | :heading | :paragraph | :div | :blockquote | :pre | :hr | :list | :list_item | :definition_list | :definition_term | :definition_description | :table | :table_row | :table_cell | :table_header | :table_body | :table_head | :table_foot | :link | :image | :strong | :em | :code | :strikethrough | :underline | :subscript | :superscript | :mark | :small | :br | :span | :article | :section | :nav | :aside | :header | :footer | :main | :figure | :figcaption | :time | :details | :summary | :form | :input | :select | :option | :button | :textarea | :label | :fieldset | :legend | :audio | :video | :picture | :source | :iframe | :svg | :canvas | :ruby | :rt | :rp | :abbr | :kbd | :samp | :var | :cite | :q | :del | :ins | :data | :meter | :progress | :output | :template | :slot | :html | :head | :body | :title | :meta | :link_tag | :style | :script | :base | :custom
|
|
597
599
|
end
|
|
598
600
|
|
|
599
601
|
class VisitResult
|
metadata
CHANGED
|
@@ -1,15 +1,29 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: html-to-markdown
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 3.4.
|
|
4
|
+
version: 3.4.1
|
|
5
5
|
platform: aarch64-linux
|
|
6
6
|
authors:
|
|
7
7
|
- Kreuzberg Team
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: bin
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2026-05-
|
|
12
|
-
dependencies:
|
|
11
|
+
date: 2026-05-14 00:00:00.000000000 Z
|
|
12
|
+
dependencies:
|
|
13
|
+
- !ruby/object:Gem::Dependency
|
|
14
|
+
name: rb_sys
|
|
15
|
+
requirement: !ruby/object:Gem::Requirement
|
|
16
|
+
requirements:
|
|
17
|
+
- - "~>"
|
|
18
|
+
- !ruby/object:Gem::Version
|
|
19
|
+
version: '0.9'
|
|
20
|
+
type: :runtime
|
|
21
|
+
prerelease: false
|
|
22
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
23
|
+
requirements:
|
|
24
|
+
- - "~>"
|
|
25
|
+
- !ruby/object:Gem::Version
|
|
26
|
+
version: '0.9'
|
|
13
27
|
description: High-performance HTML to Markdown converter
|
|
14
28
|
email:
|
|
15
29
|
executables: []
|
|
@@ -17,7 +31,12 @@ extensions: []
|
|
|
17
31
|
extra_rdoc_files: []
|
|
18
32
|
files:
|
|
19
33
|
- Steepfile
|
|
20
|
-
-
|
|
34
|
+
- ext/html_to_markdown_rb/Cargo.toml
|
|
35
|
+
- ext/html_to_markdown_rb/Makefile
|
|
36
|
+
- ext/html_to_markdown_rb/extconf.rb
|
|
37
|
+
- ext/html_to_markdown_rb/native/Cargo.lock
|
|
38
|
+
- ext/html_to_markdown_rb/native/Cargo.toml
|
|
39
|
+
- ext/html_to_markdown_rb/src/lib.rs
|
|
21
40
|
- lib/html_to_markdown.rb
|
|
22
41
|
- lib/html_to_markdown/native.rb
|
|
23
42
|
- lib/html_to_markdown/version.rb
|
|
@@ -40,7 +59,10 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
|
40
59
|
requirements:
|
|
41
60
|
- - ">="
|
|
42
61
|
- !ruby/object:Gem::Version
|
|
43
|
-
version: 3.
|
|
62
|
+
version: '3.3'
|
|
63
|
+
- - "<"
|
|
64
|
+
- !ruby/object:Gem::Version
|
|
65
|
+
version: 3.4.dev
|
|
44
66
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
|
45
67
|
requirements:
|
|
46
68
|
- - ">="
|
data/lib/bin/html-to-markdown
DELETED
|
Binary file
|