html-to-markdown 3.2.1 → 3.2.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: dfb7a68ac2c1b5c3629d8a11dc85321b39a933426879e7d2cc4697e3e1d80adc
-  data.tar.gz: b5d4af84b1b75fa552a101651bd3c0b35af28b5554981f8b188abf7692e8e1df
+  metadata.gz: 3c711c865a3b470f6d6bc92636f34a5b7e68cde781425190fc0a44cfc29cd2ef
+  data.tar.gz: '097aa9f04565d4e9991080b2fddc602d1ee03e74290121fbafe19b315475a554'
 SHA512:
-  metadata.gz: 965b91313b2d28e3bc8039c4cf5ad037e0e064933b52ad328b77619a9500b6de3dee76472f1c3fb8b32a66fb996313aec4d9ed1ecaa2834da5c606d28803e14e
-  data.tar.gz: 8412594ec37d83ecc7a7099990279888232119292392500b6bd8ffcadcf75d1e0d3c7d30f48d49e5e4459566baf75492353dbe4f9d604f5cd495179c263ab3d8
+  metadata.gz: 8b7c260249126639ca0025994bf23f6b969b918f86c5eb1d628a0422295a9e66b5819ec2d6d88df294dfd9c5e49136fde88e16f2cc27193d8e023b01302cfefa
+  data.tar.gz: 4417306533d8cf18ae4e77aef7070cba23c987af138bd43b4245bd2f845c17b3ef05acb7e32c07a2eece06f8df8ee676883a0f0238643a28d93872ce836cd0d3

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    html-to-markdown (3.2.1)
+    html-to-markdown (3.2.3)
       rb_sys (>= 0.9, < 1.0)
 
 GEM
@@ -128,7 +128,7 @@ CHECKSUMS
   ffi (1.17.4-arm64-darwin) sha256=19071aaf1419251b0a46852abf960e77330a3b334d13a4ab51d58b31a937001b
   ffi (1.17.4-x86_64-linux-gnu) sha256=9d3db14c2eae074b382fa9c083fe95aec6e0a1451da249eab096c34002bc752d
   fileutils (1.8.0) sha256=8c6b1df54e2540bdb2f39258f08af78853aa70bad52b4d394bbc6424593c6e02
-  html-to-markdown (3.2.1)
+  html-to-markdown (3.2.3)
   json (2.19.3) sha256=289b0bb53052a1fa8c34ab33cc750b659ba14a5c45f3fcf4b18762dc67c78646
   language_server-protocol (3.17.0.5) sha256=fd1e39a51a28bf3eec959379985a72e296e9f9acfce46f6a79d31ca8760803cc
   lint_roller (1.1.0) sha256=2c0c845b632a7d172cb849cc90c1bce937a28c5c8ccccb50dfd46a485003cc87

data/ext/html_to_markdown_rb/src/lib.rs

@@ -40,9 +40,8 @@ fn json_to_ruby(handle: &Ruby, val: serde_json::Value) -> magnus::Value {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::MetadataConfig")]
-#[serde(default)]
 pub struct MetadataConfig {
     pub extract_document: bool,
     pub extract_headers: bool,
@@ -131,9 +130,8 @@ impl MetadataConfig {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::MetadataConfigUpdate")]
-#[serde(default)]
 pub struct MetadataConfigUpdate {
     pub extract_document: Option<bool>,
     pub extract_headers: Option<bool>,
@@ -210,9 +208,8 @@ impl MetadataConfigUpdate {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::DocumentMetadata")]
-#[serde(default)]
 pub struct DocumentMetadata {
     pub title: Option<String>,
     pub description: Option<String>,
@@ -237,24 +234,6 @@ impl magnus::TryConvert for DocumentMetadata {
 }
 unsafe impl TryConvertOwned for DocumentMetadata {}
 
-impl Default for DocumentMetadata {
-    fn default() -> Self {
-        Self {
-            title: Default::default(),
-            description: Default::default(),
-            keywords: Default::default(),
-            author: Default::default(),
-            canonical_url: Default::default(),
-            base_href: Default::default(),
-            language: Default::default(),
-            text_direction: Default::default(),
-            open_graph: Default::default(),
-            twitter_card: Default::default(),
-            meta_tags: Default::default(),
-        }
-    }
-}
-
 impl DocumentMetadata {
     fn new(
         title: Option<String>,
@@ -329,7 +308,7 @@ impl DocumentMetadata {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::HeaderMetadata")]
 pub struct HeaderMetadata {
     pub level: u8,
@@ -392,7 +371,7 @@ impl HeaderMetadata {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::LinkMetadata")]
 pub struct LinkMetadata {
     pub href: String,
@@ -457,7 +436,7 @@ impl LinkMetadata {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::ImageMetadata")]
 pub struct ImageMetadata {
     pub src: String,
@@ -522,7 +501,7 @@ impl ImageMetadata {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::StructuredData")]
 pub struct StructuredData {
     pub data_type: StructuredDataType,
@@ -562,9 +541,8 @@ impl StructuredData {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::HtmlMetadata")]
-#[serde(default)]
 pub struct HtmlMetadata {
     pub document: DocumentMetadata,
     pub headers: Vec<HeaderMetadata>,
@@ -583,18 +561,6 @@ impl magnus::TryConvert for HtmlMetadata {
 }
 unsafe impl TryConvertOwned for HtmlMetadata {}
 
-impl Default for HtmlMetadata {
-    fn default() -> Self {
-        Self {
-            document: Default::default(),
-            headers: Default::default(),
-            links: Default::default(),
-            images: Default::default(),
-            structured_data: Default::default(),
-        }
-    }
-}
-
 impl HtmlMetadata {
     fn new(
         document: Option<DocumentMetadata>,
@@ -633,9 +599,8 @@ impl HtmlMetadata {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::ConversionOptions")]
-#[serde(default)]
 pub struct ConversionOptions {
     pub heading_style: HeadingStyle,
     pub list_indent_type: ListIndentType,
@@ -687,51 +652,6 @@ impl magnus::TryConvert for ConversionOptions {
 }
 unsafe impl TryConvertOwned for ConversionOptions {}
 
-impl Default for ConversionOptions {
-    fn default() -> Self {
-        Self {
-            heading_style: Default::default(),
-            list_indent_type: Default::default(),
-            list_indent_width: Default::default(),
-            bullets: Default::default(),
-            strong_em_symbol: Default::default(),
-            escape_asterisks: Default::default(),
-            escape_underscores: Default::default(),
-            escape_misc: Default::default(),
-            escape_ascii: Default::default(),
-            code_language: Default::default(),
-            autolinks: Default::default(),
-            default_title: Default::default(),
-            br_in_tables: Default::default(),
-            highlight_style: Default::default(),
-            extract_metadata: Default::default(),
-            whitespace_mode: Default::default(),
-            strip_newlines: Default::default(),
-            wrap: Default::default(),
-            wrap_width: Default::default(),
-            convert_as_inline: Default::default(),
-            sub_symbol: Default::default(),
-            sup_symbol: Default::default(),
-            newline_style: Default::default(),
-            code_block_style: Default::default(),
-            keep_inline_images_in: Default::default(),
-            preprocessing: Default::default(),
-            encoding: Default::default(),
-            debug: Default::default(),
-            strip_tags: Default::default(),
-            preserve_tags: Default::default(),
-            skip_images: Default::default(),
-            link_style: Default::default(),
-            output_format: Default::default(),
-            include_document_structure: Default::default(),
-            extract_images: Default::default(),
-            max_image_size: Default::default(),
-            capture_svg: Default::default(),
-            infer_dimensions: Default::default(),
-        }
-    }
-}
-
 impl ConversionOptions {
     fn new(kwargs: magnus::RHash) -> Result<Self, magnus::Error> {
         let ruby = unsafe { magnus::Ruby::get_unchecked() };
@@ -1090,9 +1010,8 @@ impl ConversionOptionsBuilder {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::ConversionOptionsUpdate")]
-#[serde(default)]
 pub struct ConversionOptionsUpdate {
     pub heading_style: Option<HeadingStyle>,
     pub list_indent_type: Option<ListIndentType>,
@@ -1144,51 +1063,6 @@ impl magnus::TryConvert for ConversionOptionsUpdate {
 }
 unsafe impl TryConvertOwned for ConversionOptionsUpdate {}
 
-impl Default for ConversionOptionsUpdate {
-    fn default() -> Self {
-        Self {
-            heading_style: Default::default(),
-            list_indent_type: Default::default(),
-            list_indent_width: Default::default(),
-            bullets: Default::default(),
-            strong_em_symbol: Default::default(),
-            escape_asterisks: Default::default(),
-            escape_underscores: Default::default(),
-            escape_misc: Default::default(),
-            escape_ascii: Default::default(),
-            code_language: Default::default(),
-            autolinks: Default::default(),
-            default_title: Default::default(),
-            br_in_tables: Default::default(),
-            highlight_style: Default::default(),
-            extract_metadata: Default::default(),
-            whitespace_mode: Default::default(),
-            strip_newlines: Default::default(),
-            wrap: Default::default(),
-            wrap_width: Default::default(),
-            convert_as_inline: Default::default(),
-            sub_symbol: Default::default(),
-            sup_symbol: Default::default(),
-            newline_style: Default::default(),
-            code_block_style: Default::default(),
-            keep_inline_images_in: Default::default(),
-            preprocessing: Default::default(),
-            encoding: Default::default(),
-            debug: Default::default(),
-            strip_tags: Default::default(),
-            preserve_tags: Default::default(),
-            skip_images: Default::default(),
-            link_style: Default::default(),
-            output_format: Default::default(),
-            include_document_structure: Default::default(),
-            extract_images: Default::default(),
-            max_image_size: Default::default(),
-            capture_svg: Default::default(),
-            infer_dimensions: Default::default(),
-        }
-    }
-}
-
 impl ConversionOptionsUpdate {
     fn new(kwargs: magnus::RHash) -> Result<Self, magnus::Error> {
         let ruby = unsafe { magnus::Ruby::get_unchecked() };
@@ -1463,9 +1337,8 @@ impl ConversionOptionsUpdate {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::PreprocessingOptions")]
-#[serde(default)]
 pub struct PreprocessingOptions {
     pub enabled: bool,
     pub preset: PreprocessingPreset,
@@ -1483,17 +1356,6 @@ impl magnus::TryConvert for PreprocessingOptions {
 }
 unsafe impl TryConvertOwned for PreprocessingOptions {}
 
-impl Default for PreprocessingOptions {
-    fn default() -> Self {
-        Self {
-            enabled: Default::default(),
-            preset: Default::default(),
-            remove_navigation: Default::default(),
-            remove_forms: Default::default(),
-        }
-    }
-}
-
 impl PreprocessingOptions {
     fn new(
         enabled: Option<bool>,
@@ -1526,9 +1388,8 @@ impl PreprocessingOptions {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::PreprocessingOptionsUpdate")]
-#[serde(default)]
 pub struct PreprocessingOptionsUpdate {
     pub enabled: Option<bool>,
     pub preset: Option<PreprocessingPreset>,
@@ -1546,17 +1407,6 @@ impl magnus::TryConvert for PreprocessingOptionsUpdate {
 }
 unsafe impl TryConvertOwned for PreprocessingOptionsUpdate {}
 
-impl Default for PreprocessingOptionsUpdate {
-    fn default() -> Self {
-        Self {
-            enabled: Default::default(),
-            preset: Default::default(),
-            remove_navigation: Default::default(),
-            remove_forms: Default::default(),
-        }
-    }
-}
-
 impl PreprocessingOptionsUpdate {
     fn new(
         enabled: Option<bool>,
@@ -1721,9 +1571,8 @@ impl TextAnnotation {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::ConversionResult")]
-#[serde(default)]
 pub struct ConversionResult {
     pub content: Option<String>,
     pub document: Option<DocumentStructure>,
@@ -1743,19 +1592,6 @@ impl magnus::TryConvert for ConversionResult {
 }
 unsafe impl TryConvertOwned for ConversionResult {}
 
-impl Default for ConversionResult {
-    fn default() -> Self {
-        Self {
-            content: Default::default(),
-            document: Default::default(),
-            metadata: Default::default(),
-            tables: Default::default(),
-            images: Default::default(),
-            warnings: Default::default(),
-        }
-    }
-}
-
 impl ConversionResult {
     fn new(
         content: Option<String>,
@@ -1800,7 +1636,7 @@ impl ConversionResult {
     }
 }
 
-#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+#[derive(Clone, Debug, serde::Serialize, serde::Deserialize, Default)]
 #[magnus::wrap(class = "HtmlToMarkdownRs::TableGrid")]
 #[serde(default)]
 pub struct TableGrid {
@@ -1819,16 +1655,6 @@ impl magnus::TryConvert for TableGrid {
 }
 unsafe impl TryConvertOwned for TableGrid {}
 
-impl Default for TableGrid {
-    fn default() -> Self {
-        Self {
-            rows: Default::default(),
-            cols: Default::default(),
-            cells: Default::default(),
-        }
-    }
-}
-
 impl TableGrid {
     fn new(rows: Option<u32>, cols: Option<u32>, cells: Option<Vec<GridCell>>) -> Self {
         Self {
@@ -2594,7 +2420,8 @@ impl magnus::IntoValue for NodeContent {
 impl magnus::TryConvert for NodeContent {
     fn try_convert(val: magnus::Value) -> Result<Self, magnus::Error> {
         let s: String = magnus::TryConvert::try_convert(val)?;
-        serde_json::from_str(&s).map_err(|e| magnus::Error::new(magnus::exception::type_error(), e.to_string()))
+        serde_json::from_str(&s)
+            .map_err(|e| magnus::Error::new(unsafe { Ruby::get_unchecked() }.exception_type_error(), e.to_string()))
     }
 }
 
@@ -2633,7 +2460,8 @@ impl magnus::IntoValue for AnnotationKind {
 impl magnus::TryConvert for AnnotationKind {
     fn try_convert(val: magnus::Value) -> Result<Self, magnus::Error> {
         let s: String = magnus::TryConvert::try_convert(val)?;
-        serde_json::from_str(&s).map_err(|e| magnus::Error::new(magnus::exception::type_error(), e.to_string()))
+        serde_json::from_str(&s)
+            .map_err(|e| magnus::Error::new(unsafe { Ruby::get_unchecked() }.exception_type_error(), e.to_string()))
     }
 }
 
@@ -2693,13 +2521,18 @@ fn convert(html: String, options: Option<String>) -> Result<ConversionResult, Er
         .as_deref()
         .filter(|s| *s != "nil")
         .map(|s| {
-            let core: html_to_markdown_rs::ConversionOptions = serde_json::from_str(s)
-                .map_err(|e| magnus::Error::new(magnus::exception::type_error(), e.to_string()))?;
+            let core: html_to_markdown_rs::ConversionOptions = serde_json::from_str(s).map_err(|e| {
+                magnus::Error::new(unsafe { Ruby::get_unchecked() }.exception_type_error(), e.to_string())
+            })?;
             Ok::<_, magnus::Error>(core.into())
         })
         .transpose()?;
-    let result = html_to_markdown_rs::convert(&html, options.map(Into::into))
-        .map_err(|e| magnus::Error::new(magnus::exception::runtime_error(), e.to_string()))?;
+    let result = html_to_markdown_rs::convert(&html, options.map(Into::into)).map_err(|e| {
+        magnus::Error::new(
+            unsafe { Ruby::get_unchecked() }.exception_runtime_error(),
+            e.to_string(),
+        )
+    })?;
     Ok(result.into())
 }
 
@@ -3640,7 +3473,7 @@ impl From<html_to_markdown_rs::WarningKind> for WarningKind {
 #[allow(dead_code)]
 fn conversion_error_to_magnus_err(e: html_to_markdown_rs::error::ConversionError) -> magnus::Error {
     let msg = e.to_string();
-    magnus::Error::new(magnus::exception::runtime_error(), msg)
+    magnus::Error::new(unsafe { magnus::Ruby::get_unchecked() }.exception_runtime_error(), msg)
 }
 
 #[magnus::init]

data/lib/html_to_markdown/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module HtmlToMarkdown
-  VERSION = '3.2.1'
+  VERSION = '3.2.3'
 end

data/vendor/Cargo.toml

@@ -3,7 +3,7 @@ members = ["html-to-markdown-rs"]
 resolver = "2"
 
 [workspace.package]
-version = "3.2.1"
+version = "3.2.3"
 edition = "2024"
 rust-version = "1.85"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]

data/vendor/html-to-markdown-rs/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "html-to-markdown-rs"
-version = "3.2.1"
+version = "3.2.3"
 edition = "2024"
 authors = ["Na'aman Hirschfeld <naaman@kreuzberg.dev>"]
 license = "MIT"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: html-to-markdown
 version: !ruby/object:Gem::Version
-  version: 3.2.1
+  version: 3.2.3
 platform: ruby
 authors:
 - Na'aman Hirschfeld
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-04-16 00:00:00.000000000 Z
+date: 2026-04-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys
@@ -51,7 +51,6 @@ files:
 - Rakefile
 - Steepfile
 - exe/html-to-markdown
-- ext/html-to-markdown_rb/Cargo.toml
 - ext/html_to_markdown_rb/Cargo.toml
 - ext/html_to_markdown_rb/extconf.rb
 - ext/html_to_markdown_rb/src/html_to_markdown_rs.rb

data/ext/html-to-markdown_rb/Cargo.toml

@@ -1,14 +0,0 @@
-[package]
-name = "html-to-markdown-rb"
-version = "3.2.0"
-edition = "2024"
-license = "MIT"
-
-[lib]
-crate-type = ["cdylib"]
-
-[dependencies]
-html-to-markdown-rs = { path = "../../../../crates/html-to-markdown", features = ["full", "metadata", "visitor", "serde", "inline-images"] }
-magnus = "0.8"
-serde = { version = "1", features = ["derive"] }
-serde_json = "1"