html-to-markdown 3.16.0 → 3.17.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8de853571f92ca7f37774b63db6d2348c505919cd43b9fb25b250a2a21206b8f
-  data.tar.gz: efd5431417618cecdf86c8a4f00fd4456c71c441dae449d28e967244d6ef023e
+  metadata.gz: 19c0e5729d5ff67d2a50d6686184cbd9e7a374c5213a1eb3bf10d0b98867471c
+  data.tar.gz: c647a47e588b194566216244edf2868d1dbfeb3afd164486a291702e82569230
 SHA512:
-  metadata.gz: ae9e03f4cf1e5177f4ff86edf3bc88779a12eb4ed184128b5c4d2ed5d60bdceed2fb10df37075e1ab643c05c69229543216a5e267074ee95823e70f834188216
-  data.tar.gz: 337c1ada4ab8b534b62b822fef53e9ce605561a40483a975799a2b9afa9a304d8654f5b7dce3924c15bc22a3a823f008603a2840ca57321b8955acc89e7b3066
+  metadata.gz: b55477714ff3d6335d584bd1dec80a0393eaed59830c361daddb5ed88ee3e97b5f80c047a9b037a6e50fffa03abcdc6d966f59b8d0d832dfad61fd323c7acae4
+  data.tar.gz: bcbd361625d98d5e996c636166343544ef3b54808f0d92e8f95f87788779bbdb97aab2282629c2c642e82bbbf66e9de27dcfe90a0b098761f741de168c3a043b

data/ext/html_to_markdown_rb/native/Cargo.lock

@@ -252,7 +252,7 @@ checksum = "c9356095b4b41197bba32173600e1582792cda618f65d12f68e2e77d273413c5"
 
 [[package]]
 name = "html-to-markdown-rb"
-version = "3.16.0"
+version = "3.17.0"
 dependencies = [
  "async-trait",
  "html-to-markdown-rs",
@@ -266,9 +266,9 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-rs"
-version = "3.16.0"
+version = "3.17.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "74c4d770081ede1c6ec100e61a57cac65bf41d28ab5516b1eae907b9f3415778"
+checksum = "0db3dbcfadba084ba535dab4bacee745d8d819d3161be64bda0f5f49c7ad81ed"
 dependencies = [
  "ahash",
  "astral-tl",
@@ -454,9 +454,9 @@ checksum = "20870f649af7073d53e38067b2a84312175d56ea15217e1b15bc83506ec50afb"
 
 [[package]]
 name = "libc"
-version = "0.2.189"
+version = "0.2.190"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3eaf3ede3fee6db1a4c2ee091bf8a8b4dccdc6d17f656fb07896ee72867612f2"
+checksum = "ce5d3ddc6d3fa000eb1536d85e147bfe31aacaba692ed6a876f95cb7c855be78"
 
 [[package]]
 name = "libloading"
@@ -491,9 +491,9 @@ checksum = "f9f8bd3e56ce4dfc153cf470fffbfa98c7620958b312ca5c3a4b8d5181fd13c6"
 
 [[package]]
 name = "magnus"
-version = "0.9.1"
+version = "0.9.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4c01bef153b8d0f086991f0e166474a1d57b738221f9a9c11405d498ac95a44d"
+checksum = "b9068710e00d21be762edca59b9c4c8946dbdb29820d4f76ad559afea07982c9"
 dependencies = [
  "magnus-macros",
  "rb-sys",
@@ -1014,9 +1014,9 @@ dependencies = [
 
 [[package]]
 name = "tokio"
-version = "1.53.1"
+version = "1.53.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "202caea871b69668250d242070849eb495be178ed697a3e98aebce5bc81a0bed"
+checksum = "e95f91fcc7a621e8b030f6aa23c71fe9838ae2fb4d8118b75602a328f5144044"
 dependencies = [
  "pin-project-lite",
 ]

data/ext/html_to_markdown_rb/native/Cargo.toml

@@ -1,11 +1,11 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:b55cf956fd88e14111a54d783edfd335d4ab5ae57ebcc23cdddbb67a1ee0dc96
+# alef:hash:fc7b06dbf1f7f4f6c6b3789b4fa618fcfce8a0bf955f98fa408fed99085aaad5
 # To regenerate: alef generate
 # To verify freshness: alef verify
 
 [package]
 name = "html-to-markdown-rb"
-version = "3.16.0"
+version = "3.17.0"
 edition = "2024"
 license = "MIT"
 description = "High-performance HTML to Markdown converter"
@@ -30,7 +30,7 @@ visitor = ["html-to-markdown-rs/visitor"]
 
 [dependencies]
 async-trait = "0.1"
-html-to-markdown-rs = { version = "3.16.0", features = ["serde", "metadata", "visitor", "inline-images", "testkit"] }
+html-to-markdown-rs = { version = "3.17.0", features = ["serde", "metadata", "visitor", "inline-images", "testkit"] }
 magnus = "0.9"
 rb-sys = ">=0.9.130, <0.10"
 serde = { version = "1", features = ["derive"] }

data/ext/html_to_markdown_rb/src/lib.rs

@@ -1,5 +1,5 @@
 // This file is auto-generated by alef. DO NOT EDIT.
-// alef:hash:ef17e73c4a8ab3ab7874d1b469b3b0bfa02b61db141c8db6976ce5dd46217c8f
+// alef:hash:d5974922982c4a533f5e1f202d897cb89a0544519b7fbcb4ab5b198735e7fec1
 // Re-generate with: alef generate
 #![allow(dead_code, unused_imports, unused_variables)]
 #![allow(
@@ -101,6 +101,7 @@ pub struct ConversionOptions {
     include_document_structure: bool,
     extract_images: bool,
     max_image_size: u64,
+    max_input_size: Option<u64>,
     capture_svg: bool,
     infer_dimensions: bool,
     max_depth: Option<usize>,
@@ -501,6 +502,15 @@ impl ConversionOptions {
                 })?,
                 None => 5242880,
             },
+            max_input_size: match kwargs.get(ruby.to_symbol("max_input_size")).filter(|v| !v.is_nil()) {
+                Some(v) => Some(u64::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `max_input_size`: {}", e),
+                    )
+                })?),
+                None => None,
+            },
             capture_svg: match kwargs.get(ruby.to_symbol("capture_svg")) {
                 Some(v) => bool::try_convert(v).map_err(|e| {
                     magnus::Error::new(
@@ -714,6 +724,10 @@ impl ConversionOptions {
         self.max_image_size
     }
 
+    fn max_input_size(&self) -> Option<u64> {
+        self.max_input_size
+    }
+
     fn capture_svg(&self) -> bool {
         self.capture_svg
     }
@@ -782,6 +796,7 @@ pub struct ConversionOptionsUpdate {
     include_document_structure: Option<bool>,
     extract_images: Option<bool>,
     max_image_size: Option<u64>,
+    max_input_size: Option<u64>,
     capture_svg: Option<bool>,
     infer_dimensions: Option<bool>,
     max_depth: Option<usize>,
@@ -1188,6 +1203,15 @@ impl ConversionOptionsUpdate {
                 })?),
                 None => None,
             },
+            max_input_size: match kwargs.get(ruby.to_symbol("max_input_size")).filter(|v| !v.is_nil()) {
+                Some(v) => Some(u64::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `max_input_size`: {}", e),
+                    )
+                })?),
+                None => None,
+            },
             capture_svg: match kwargs.get(ruby.to_symbol("capture_svg")).filter(|v| !v.is_nil()) {
                 Some(v) => Some(bool::try_convert(v).map_err(|e| {
                     magnus::Error::new(
@@ -1401,6 +1425,10 @@ impl ConversionOptionsUpdate {
         self.max_image_size
     }
 
+    fn max_input_size(&self) -> Option<u64> {
+        self.max_input_size.clone()
+    }
+
     fn capture_svg(&self) -> Option<bool> {
         self.capture_svg
     }
@@ -6904,6 +6932,7 @@ impl From<ConversionOptions> for html_to_markdown_rs::options::ConversionOptions
             include_document_structure: val.include_document_structure,
             extract_images: val.extract_images,
             max_image_size: val.max_image_size,
+            max_input_size: val.max_input_size,
             capture_svg: val.capture_svg,
             infer_dimensions: val.infer_dimensions,
             max_depth: val.max_depth,
@@ -6958,6 +6987,7 @@ impl From<html_to_markdown_rs::options::ConversionOptions> for ConversionOptions
             include_document_structure: val.include_document_structure,
             extract_images: val.extract_images,
             max_image_size: val.max_image_size,
+            max_input_size: val.max_input_size,
             capture_svg: val.capture_svg,
             infer_dimensions: val.infer_dimensions,
             max_depth: val.max_depth,
@@ -7011,6 +7041,7 @@ impl From<html_to_markdown_rs::options::ConversionOptionsUpdate> for ConversionO
             include_document_structure: val.include_document_structure,
             extract_images: val.extract_images,
             max_image_size: val.max_image_size,
+            max_input_size: val.max_input_size.flatten(),
             capture_svg: val.capture_svg,
             infer_dimensions: val.infer_dimensions,
             max_depth: val.max_depth.flatten(),
@@ -8320,6 +8351,8 @@ fn ruby_init(ruby: &Ruby) -> Result<(), Error> {
 
     class.define_method("max_image_size", method!(ConversionOptions::max_image_size, 0))?;
 
+    class.define_method("max_input_size", method!(ConversionOptions::max_input_size, 0))?;
+
     class.define_method("capture_svg", method!(ConversionOptions::capture_svg, 0))?;
 
     class.define_method("infer_dimensions", method!(ConversionOptions::infer_dimensions, 0))?;
@@ -8450,6 +8483,8 @@ fn ruby_init(ruby: &Ruby) -> Result<(), Error> {
 
     class.define_method("max_image_size", method!(ConversionOptionsUpdate::max_image_size, 0))?;
 
+    class.define_method("max_input_size", method!(ConversionOptionsUpdate::max_input_size, 0))?;
+
     class.define_method("capture_svg", method!(ConversionOptionsUpdate::capture_svg, 0))?;
 
     class.define_method(

data/lib/html_to_markdown/version.rb

@@ -1,10 +1,10 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:9cd1523beb892a6e5eef0c33a956855a8178865245dbf115fe62ed9e618d85ab
+# alef:hash:31cca426bed4cf070e9b4260d3264c73700befe6d2315f631edfc4885e1f3252
 # To regenerate: alef generate
 # To verify freshness: alef verify
 # frozen_string_literal: true
 
 module HtmlToMarkdown
   ## The version string for this package.
-  VERSION = "3.16.0"
+  VERSION = "3.17.0"
 end

data/sig/types.rbs

@@ -1,5 +1,5 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:7c6e1f7268fd0bfae41495747b5df560e98fa6c7b17ab6e17963ab6c5f8e4e53
+# alef:hash:622a87d08343ba4489b9f55f6d1f41fec2db28a92968c81276f59faf488ecbdb
 # To regenerate: alef generate
 # To verify freshness: alef verify
 
@@ -49,6 +49,7 @@ type json_value = Hash[String, untyped] | Array[untyped] | String | Integer | Fl
         attr_reader include_document_structure: bool
         attr_reader extract_images: bool
         attr_reader max_image_size: Integer
+        attr_reader max_input_size: Integer?
         attr_reader capture_svg: bool
         attr_reader infer_dimensions: bool
         attr_reader max_depth: Integer?
@@ -57,7 +58,7 @@ type json_value = Hash[String, untyped] | Array[untyped] | String | Integer | Fl
         attr_reader base_url: String?
         attr_reader visitor: VisitorHandle?
 
-def initialize: (?heading_style: enum_HeadingStyle, ?list_indent_type: enum_ListIndentType, ?list_indent_width: Integer, ?bullets: String, ?strong_em_symbol: String, ?escape_asterisks: bool, ?escape_underscores: bool, ?escape_misc: bool, ?escape_ascii: bool, ?code_language: String, ?autolinks: bool, ?default_title: bool, ?br_in_tables: bool, ?compact_tables: bool, ?highlight_style: enum_HighlightStyle, ?extract_metadata: bool, ?whitespace_mode: enum_WhitespaceMode, ?strip_newlines: bool, ?wrap: bool, ?wrap_width: Integer, ?convert_as_inline: bool, ?sub_symbol: String, ?sup_symbol: String, ?newline_style: enum_NewlineStyle, ?code_block_style: enum_CodeBlockStyle, ?keep_inline_images_in: Array[String], ?preprocessing: PreprocessingOptions, ?encoding: String, ?debug: bool, ?strip_tags: Array[String], ?preserve_tags: Array[String], ?skip_images: bool, ?inline_data_media: enum_InlineDataMedia, ?url_escape_style: enum_UrlEscapeStyle, ?link_style: enum_LinkStyle, ?output_format: enum_OutputFormat, ?include_document_structure: bool, ?extract_images: bool, ?max_image_size: Integer, ?capture_svg: bool, ?infer_dimensions: bool, ?max_depth: Integer, ?exclude_selectors: Array[String], ?tier_strategy: enum_TierStrategy, ?base_url: String, ?visitor: VisitorHandle) -> void
+def initialize: (?heading_style: enum_HeadingStyle, ?list_indent_type: enum_ListIndentType, ?list_indent_width: Integer, ?bullets: String, ?strong_em_symbol: String, ?escape_asterisks: bool, ?escape_underscores: bool, ?escape_misc: bool, ?escape_ascii: bool, ?code_language: String, ?autolinks: bool, ?default_title: bool, ?br_in_tables: bool, ?compact_tables: bool, ?highlight_style: enum_HighlightStyle, ?extract_metadata: bool, ?whitespace_mode: enum_WhitespaceMode, ?strip_newlines: bool, ?wrap: bool, ?wrap_width: Integer, ?convert_as_inline: bool, ?sub_symbol: String, ?sup_symbol: String, ?newline_style: enum_NewlineStyle, ?code_block_style: enum_CodeBlockStyle, ?keep_inline_images_in: Array[String], ?preprocessing: PreprocessingOptions, ?encoding: String, ?debug: bool, ?strip_tags: Array[String], ?preserve_tags: Array[String], ?skip_images: bool, ?inline_data_media: enum_InlineDataMedia, ?url_escape_style: enum_UrlEscapeStyle, ?link_style: enum_LinkStyle, ?output_format: enum_OutputFormat, ?include_document_structure: bool, ?extract_images: bool, ?max_image_size: Integer, ?max_input_size: Integer, ?capture_svg: bool, ?infer_dimensions: bool, ?max_depth: Integer, ?exclude_selectors: Array[String], ?tier_strategy: enum_TierStrategy, ?base_url: String, ?visitor: VisitorHandle) -> void
     end
 
     class ConversionOptionsUpdate
@@ -100,6 +101,7 @@ def initialize: (?heading_style: enum_HeadingStyle, ?list_indent_type: enum_List
         attr_reader include_document_structure: bool?
         attr_reader extract_images: bool?
         attr_reader max_image_size: Integer?
+        attr_reader max_input_size: Integer?
         attr_reader capture_svg: bool?
         attr_reader infer_dimensions: bool?
         attr_reader max_depth: Integer?
@@ -108,7 +110,7 @@ def initialize: (?heading_style: enum_HeadingStyle, ?list_indent_type: enum_List
         attr_reader base_url: String?
         attr_reader visitor: VisitorHandle?
 
-def initialize: (?heading_style: enum_HeadingStyle, ?list_indent_type: enum_ListIndentType, ?list_indent_width: Integer, ?bullets: String, ?strong_em_symbol: String, ?escape_asterisks: bool, ?escape_underscores: bool, ?escape_misc: bool, ?escape_ascii: bool, ?code_language: String, ?autolinks: bool, ?default_title: bool, ?br_in_tables: bool, ?compact_tables: bool, ?highlight_style: enum_HighlightStyle, ?extract_metadata: bool, ?whitespace_mode: enum_WhitespaceMode, ?strip_newlines: bool, ?wrap: bool, ?wrap_width: Integer, ?convert_as_inline: bool, ?sub_symbol: String, ?sup_symbol: String, ?newline_style: enum_NewlineStyle, ?code_block_style: enum_CodeBlockStyle, ?keep_inline_images_in: Array[String], ?preprocessing: PreprocessingOptionsUpdate, ?encoding: String, ?debug: bool, ?strip_tags: Array[String], ?preserve_tags: Array[String], ?skip_images: bool, ?inline_data_media: enum_InlineDataMedia, ?url_escape_style: enum_UrlEscapeStyle, ?link_style: enum_LinkStyle, ?output_format: enum_OutputFormat, ?include_document_structure: bool, ?extract_images: bool, ?max_image_size: Integer, ?capture_svg: bool, ?infer_dimensions: bool, ?max_depth: Integer?, ?exclude_selectors: Array[String], ?tier_strategy: enum_TierStrategy, ?base_url: String?, ?visitor: VisitorHandle) -> void
+def initialize: (?heading_style: enum_HeadingStyle, ?list_indent_type: enum_ListIndentType, ?list_indent_width: Integer, ?bullets: String, ?strong_em_symbol: String, ?escape_asterisks: bool, ?escape_underscores: bool, ?escape_misc: bool, ?escape_ascii: bool, ?code_language: String, ?autolinks: bool, ?default_title: bool, ?br_in_tables: bool, ?compact_tables: bool, ?highlight_style: enum_HighlightStyle, ?extract_metadata: bool, ?whitespace_mode: enum_WhitespaceMode, ?strip_newlines: bool, ?wrap: bool, ?wrap_width: Integer, ?convert_as_inline: bool, ?sub_symbol: String, ?sup_symbol: String, ?newline_style: enum_NewlineStyle, ?code_block_style: enum_CodeBlockStyle, ?keep_inline_images_in: Array[String], ?preprocessing: PreprocessingOptionsUpdate, ?encoding: String, ?debug: bool, ?strip_tags: Array[String], ?preserve_tags: Array[String], ?skip_images: bool, ?inline_data_media: enum_InlineDataMedia, ?url_escape_style: enum_UrlEscapeStyle, ?link_style: enum_LinkStyle, ?output_format: enum_OutputFormat, ?include_document_structure: bool, ?extract_images: bool, ?max_image_size: Integer, ?max_input_size: Integer?, ?capture_svg: bool, ?infer_dimensions: bool, ?max_depth: Integer?, ?exclude_selectors: Array[String], ?tier_strategy: enum_TierStrategy, ?base_url: String?, ?visitor: VisitorHandle) -> void
     end
 
     class ConversionResult

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: html-to-markdown
 version: !ruby/object:Gem::Version
-  version: 3.16.0
+  version: 3.17.0
 platform: ruby
 authors:
 - Na'aman Hirschfeld <naaman@xberg.io>
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-10-02 00:00:00.000000000 Z
+date: 2026-10-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys