html-to-markdown 3.15.0 → 3.17.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a687f5a155a438defe585a158de533340696013393a2f218052706919574c4b3
-  data.tar.gz: f153aee57818a7bb4324bc0509a5a3e08ea92e9cf370a639d43969548a9b8064
+  metadata.gz: 19c0e5729d5ff67d2a50d6686184cbd9e7a374c5213a1eb3bf10d0b98867471c
+  data.tar.gz: c647a47e588b194566216244edf2868d1dbfeb3afd164486a291702e82569230
 SHA512:
-  metadata.gz: e76e5f2fe222209b761fd75312c22741eeafff30f41d046f35fdc9bb86fb6a691f812c31cab936c751184504e03c3cb4ca67fae9ca60d0282772f411c4e03543
-  data.tar.gz: f91b991b15c70926b54d6b8aaf9b4978de3d5a311fc561ca512be9f9df474e03c7de90df4f594871cb02efd7ec100982c1dd65472ba1f323d2428315b426d215
+  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 = "3.15.0"
+version = "3.17.0"
 dependencies = [
  "async-trait",
  "html-to-markdown-rs",
@@ -266,9 +266,9 @@ dependencies = [
 
 [[package]]
 name = "html-to-markdown-rs"
-version = "3.15.0"
+version = "3.17.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "16c7356a39129a282aa7aa92bd8de8aba28cdeb74d6e7321bcc1c232ff123104"
+checksum = "0db3dbcfadba084ba535dab4bacee745d8d819d3161be64bda0f5f49c7ad81ed"
 dependencies = [
  "ahash",
  "astral-tl",
@@ -448,15 +448,15 @@ checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682"
 
 [[package]]
 name = "lazy_static"
-version = "1.5.0"
+version = "1.5.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe"
+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 = "0.9.0"
+version = "0.9.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "379d50b6eea4ca84074ccd3e551eb9b07f7865194d5b3198325317e9ef78e5e9"
+checksum = "b9068710e00d21be762edca59b9c4c8946dbdb29820d4f76ad559afea07982c9"
 dependencies = [
  "magnus-macros",
  "rb-sys",
@@ -503,9 +503,9 @@ dependencies = [
 
 [[package]]
 name = "magnus-macros"
-version = "0.9.0"
+version = "0.9.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4ca70566af98184bb92d4d9853daaf533c1a3418b9570d0547ea44970035f3fe"
+checksum = "9dafde2eeb66f090e0f326112323eb8324b12cb32b5d1210e541d3df6b0ce77e"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -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",
 ]
@@ -1131,9 +1131,9 @@ dependencies = [
 
 [[package]]
 name = "yoke-derive"
-version = "0.8.3"
+version = "0.8.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "33811428bee40dbceb6d545e95754741d17a6aef9a4849f0fd62e2ba4f412a78"
+checksum = "ec8ebde2db3681e8c9980cc27822030e68752690ddfa9473e739aeb4dbde6d71"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -1143,18 +1143,18 @@ dependencies = [
 
 [[package]]
 name = "zerocopy"
-version = "0.8.58"
+version = "0.8.59"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c17e8fafad82b542ff3717217ecdc736231b59e387768c9630123b4ce4d2db44"
+checksum = "6df92bf3d9227be3d53173901ddbffac2babc27ae50f397776ffd6dc33f800cb"
 dependencies = [
  "zerocopy-derive",
 ]
 
 [[package]]
 name = "zerocopy-derive"
-version = "0.8.58"
+version = "0.8.59"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "595f56e044df4f46a0c9a626f65c3d99eb8488f7e8a8baa12dd76326d9710bf2"
+checksum = "ac4f328cf2f05d084e496c3e9c3f33ed0a183656a16e1fcec4d464d8373aec82"
 dependencies = [
  "proc-macro2",
  "quote",

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:7c09d629fcfd674b7909043e4ee6d1a609cdd325f88aeb6590932703d33aa83c
+# alef:hash:fc7b06dbf1f7f4f6c6b3789b4fa618fcfce8a0bf955f98fa408fed99085aaad5
 # To regenerate: alef generate
 # To verify freshness: alef verify
 
 [package]
 name = "html-to-markdown-rb"
-version = "3.15.0"
+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 = "3.15.0", 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:adcf6160cdf7035806744fc3efe8b3116ce330a5c164ade234e6c2fe49fbe748
+// alef:hash:d5974922982c4a533f5e1f202d897cb89a0544519b7fbcb4ab5b198735e7fec1
 // Re-generate with: alef generate
 #![allow(dead_code, unused_imports, unused_variables)]
 #![allow(
@@ -94,12 +94,14 @@ pub struct ConversionOptions {
     strip_tags: Vec<String>,
     preserve_tags: Vec<String>,
     skip_images: bool,
+    inline_data_media: InlineDataMedia,
     url_escape_style: UrlEscapeStyle,
     link_style: LinkStyle,
     output_format: OutputFormat,
     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>,
@@ -437,6 +439,15 @@ impl ConversionOptions {
                 })?,
                 None => false,
             },
+            inline_data_media: match kwargs.get(ruby.to_symbol("inline_data_media")) {
+                Some(v) => InlineDataMedia::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `inline_data_media`: {}", e),
+                    )
+                })?,
+                None => InlineDataMedia::Keep,
+            },
             url_escape_style: match kwargs.get(ruby.to_symbol("url_escape_style")) {
                 Some(v) => UrlEscapeStyle::try_convert(v).map_err(|e| {
                     magnus::Error::new(
@@ -491,6 +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(
@@ -676,6 +696,10 @@ impl ConversionOptions {
         self.skip_images
     }
 
+    fn inline_data_media(&self) -> InlineDataMedia {
+        self.inline_data_media.clone()
+    }
+
     fn url_escape_style(&self) -> UrlEscapeStyle {
         self.url_escape_style.clone()
     }
@@ -700,6 +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
     }
@@ -761,12 +789,14 @@ pub struct ConversionOptionsUpdate {
     strip_tags: Option<Vec<String>>,
     preserve_tags: Option<Vec<String>>,
     skip_images: Option<bool>,
+    inline_data_media: Option<InlineDataMedia>,
     url_escape_style: Option<UrlEscapeStyle>,
     link_style: Option<LinkStyle>,
     output_format: Option<OutputFormat>,
     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>,
@@ -1107,6 +1137,15 @@ impl ConversionOptionsUpdate {
                 })?),
                 None => None,
             },
+            inline_data_media: match kwargs.get(ruby.to_symbol("inline_data_media")).filter(|v| !v.is_nil()) {
+                Some(v) => Some(InlineDataMedia::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `inline_data_media`: {}", e),
+                    )
+                })?),
+                None => None,
+            },
             url_escape_style: match kwargs.get(ruby.to_symbol("url_escape_style")).filter(|v| !v.is_nil()) {
                 Some(v) => Some(UrlEscapeStyle::try_convert(v).map_err(|e| {
                     magnus::Error::new(
@@ -1164,6 +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(
@@ -1349,6 +1397,10 @@ impl ConversionOptionsUpdate {
         self.skip_images
     }
 
+    fn inline_data_media(&self) -> Option<InlineDataMedia> {
+        self.inline_data_media.clone()
+    }
+
     fn url_escape_style(&self) -> Option<UrlEscapeStyle> {
         self.url_escape_style.clone()
     }
@@ -1373,6 +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
     }
@@ -3921,6 +3977,58 @@ impl magnus::TryConvert for ImageType {
 unsafe impl IntoValueFromNative for ImageType {}
 unsafe impl TryConvertOwned for ImageType {}
 
+#[derive(Clone, Copy, PartialEq, Eq, Hash, Debug, serde::Serialize, serde::Deserialize)]
+pub enum InlineDataMedia {
+    Keep,
+    AltTextOnly,
+    DropElement,
+}
+
+impl Default for InlineDataMedia {
+    fn default() -> Self {
+        Self::Keep
+    }
+}
+
+impl magnus::IntoValue for InlineDataMedia {
+    fn into_value_with(self, handle: &Ruby) -> magnus::Value {
+        let sym = match self {
+            InlineDataMedia::Keep => "Keep",
+            InlineDataMedia::AltTextOnly => "AltTextOnly",
+            InlineDataMedia::DropElement => "DropElement",
+        };
+        handle.to_symbol(sym).into_value_with(handle)
+    }
+}
+
+impl magnus::TryConvert for InlineDataMedia {
+    fn try_convert(val: magnus::Value) -> Result<Self, magnus::Error> {
+        // IntoValue emits Symbols, including when this enum is a Hash key.
+        // Accept that exact representation as well as String inputs; do not
+        // coerce arbitrary objects through to_s.
+        let s: String = if let Some(symbol) = magnus::Symbol::from_value(val) {
+            symbol.name()?.into_owned()
+        } else {
+            magnus::TryConvert::try_convert(val)?
+        };
+        // Accept the real serde wire value, the legacy always-snake_case symbol Magnus used to
+        // emit unconditionally, and the verbatim PascalCase Rust variant name, so fixtures and
+        // existing consumer code written in any of those styles keep working.
+        match s.as_str() {
+            "Keep" | "keep" => Ok(InlineDataMedia::Keep),
+            "AltTextOnly" | "alt_text_only" => Ok(InlineDataMedia::AltTextOnly),
+            "DropElement" | "drop_element" => Ok(InlineDataMedia::DropElement),
+            other => Err(magnus::Error::new(
+                unsafe { Ruby::get_unchecked() }.exception_arg_error(),
+                format!("invalid InlineDataMedia value: {other}"),
+            )),
+        }
+    }
+}
+
+unsafe impl IntoValueFromNative for InlineDataMedia {}
+unsafe impl TryConvertOwned for InlineDataMedia {}
+
 #[derive(Clone, Copy, PartialEq, Eq, Hash, Debug, serde::Serialize, serde::Deserialize)]
 pub enum LinkStyle {
     Inline,
@@ -6817,12 +6925,14 @@ impl From<ConversionOptions> for html_to_markdown_rs::options::ConversionOptions
             strip_tags: val.strip_tags.into_iter().collect(),
             preserve_tags: val.preserve_tags.into_iter().collect(),
             skip_images: val.skip_images,
+            inline_data_media: val.inline_data_media.into(),
             url_escape_style: val.url_escape_style.into(),
             link_style: val.link_style.into(),
             output_format: val.output_format.into(),
             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,
@@ -6870,12 +6980,14 @@ impl From<html_to_markdown_rs::options::ConversionOptions> for ConversionOptions
             strip_tags: val.strip_tags.into_iter().collect(),
             preserve_tags: val.preserve_tags.into_iter().collect(),
             skip_images: val.skip_images,
+            inline_data_media: val.inline_data_media.into(),
             url_escape_style: val.url_escape_style.into(),
             link_style: val.link_style.into(),
             output_format: val.output_format.into(),
             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,
@@ -6922,12 +7034,14 @@ impl From<html_to_markdown_rs::options::ConversionOptionsUpdate> for ConversionO
             strip_tags: val.strip_tags.map(|v| v.into_iter().collect()),
             preserve_tags: val.preserve_tags.map(|v| v.into_iter().collect()),
             skip_images: val.skip_images,
+            inline_data_media: val.inline_data_media.map(Into::into),
             url_escape_style: val.url_escape_style.map(Into::into),
             link_style: val.link_style.map(Into::into),
             output_format: val.output_format.map(Into::into),
             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(),
@@ -7559,6 +7673,26 @@ impl From<html_to_markdown_rs::metadata::ImageType> for ImageType {
     }
 }
 
+impl From<InlineDataMedia> for html_to_markdown_rs::options::InlineDataMedia {
+    fn from(val: InlineDataMedia) -> Self {
+        match val {
+            InlineDataMedia::Keep => Self::Keep,
+            InlineDataMedia::AltTextOnly => Self::AltTextOnly,
+            InlineDataMedia::DropElement => Self::DropElement,
+        }
+    }
+}
+
+impl From<html_to_markdown_rs::options::InlineDataMedia> for InlineDataMedia {
+    fn from(val: html_to_markdown_rs::options::InlineDataMedia) -> Self {
+        match val {
+            html_to_markdown_rs::options::InlineDataMedia::Keep => Self::Keep,
+            html_to_markdown_rs::options::InlineDataMedia::AltTextOnly => Self::AltTextOnly,
+            html_to_markdown_rs::options::InlineDataMedia::DropElement => Self::DropElement,
+        }
+    }
+}
+
 impl From<LinkStyle> for html_to_markdown_rs::options::LinkStyle {
     fn from(val: LinkStyle) -> Self {
         match val {
@@ -8200,6 +8334,8 @@ fn ruby_init(ruby: &Ruby) -> Result<(), Error> {
 
     class.define_method("skip_images", method!(ConversionOptions::skip_images, 0))?;
 
+    class.define_method("inline_data_media", method!(ConversionOptions::inline_data_media, 0))?;
+
     class.define_method("url_escape_style", method!(ConversionOptions::url_escape_style, 0))?;
 
     class.define_method("link_style", method!(ConversionOptions::link_style, 0))?;
@@ -8215,6 +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))?;
@@ -8322,6 +8460,11 @@ fn ruby_init(ruby: &Ruby) -> Result<(), Error> {
 
     class.define_method("skip_images", method!(ConversionOptionsUpdate::skip_images, 0))?;
 
+    class.define_method(
+        "inline_data_media",
+        method!(ConversionOptionsUpdate::inline_data_media, 0),
+    )?;
+
     class.define_method(
         "url_escape_style",
         method!(ConversionOptionsUpdate::url_escape_style, 0),
@@ -8340,6 +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:d3db8f2a519dfe0fdc4abb8529bf0e8c4990d6581fa1a04122cce92a730c65d5
+# 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 = "3.15.0"
+  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:c5ae0579cda5776e922483b37b6e2a326aa2211bf97a44c9155bb8478aa502d5
+# alef:hash:622a87d08343ba4489b9f55f6d1f41fec2db28a92968c81276f59faf488ecbdb
 # To regenerate: alef generate
 # To verify freshness: alef verify
 
@@ -42,12 +42,14 @@ type json_value = Hash[String, untyped] | Array[untyped] | String | Integer | Fl
     attr_reader strip_tags: Array[String]
     attr_reader preserve_tags: Array[String]
         attr_reader skip_images: bool
+        attr_reader inline_data_media: enum_InlineDataMedia
         attr_reader url_escape_style: enum_UrlEscapeStyle
         attr_reader link_style: enum_LinkStyle
         attr_reader output_format: enum_OutputFormat
         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?
@@ -56,7 +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, ?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
@@ -92,12 +94,14 @@ def initialize: (?heading_style: enum_HeadingStyle, ?list_indent_type: enum_List
     attr_reader strip_tags: Array[String]?
     attr_reader preserve_tags: Array[String]?
         attr_reader skip_images: bool?
+        attr_reader inline_data_media: enum_InlineDataMedia?
         attr_reader url_escape_style: enum_UrlEscapeStyle?
         attr_reader link_style: enum_LinkStyle?
         attr_reader output_format: enum_OutputFormat?
         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?
@@ -106,7 +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, ?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
@@ -304,6 +308,8 @@ def initialize: (?rows: Integer, ?cols: Integer, ?cells: Array[GridCell]) -> voi
 
     type enum_ImageType = :data_uri | :inline_svg | :external | :relative
 
+    type enum_InlineDataMedia = :Keep | :AltTextOnly | :DropElement
+
     type enum_LinkStyle = :Inline | :Reference
 
     type enum_LinkType = :anchor | :internal | :external | :email | :phone | :other

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: html-to-markdown
 version: !ruby/object:Gem::Version
-  version: 3.15.0
+  version: 3.17.0
 platform: ruby
 authors:
 - Na'aman Hirschfeld <naaman@xberg.io>
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-09-26 00:00:00.000000000 Z
+date: 2026-10-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys