html-to-markdown 2.27.3 → 2.28.1

data/rust-vendor/uuid/Cargo.toml.orig

@@ -29,7 +29,7 @@ homepage = "https://github.com/uuid-rs/uuid"
 name = "uuid"
 readme = "README.md"
 repository = "https://github.com/uuid-rs/uuid"
-version = "1.21.0" # remember to update html_root_url in lib.rs
+version = "1.22.0" # remember to update html_root_url in lib.rs
 rust-version = "1.85.0"
 
 [package.metadata.docs.rs]
@@ -86,7 +86,7 @@ borsh = ["dep:borsh", "dep:borsh-derive"]
 
 # Public: Used in trait impls on `Uuid`
 [dependencies.bytemuck]
-version = "1.21.0"
+version = "1.22.0"
 optional = true
 features = ["derive"]
 
@@ -140,13 +140,13 @@ optional = true
 [target.'cfg(all(target_arch = "wasm32", any(target_os = "unknown", target_os = "none")))'.dependencies.uuid-rng-internal-lib]
 # Work-around lack of support for both `dep:x` and `x/` in MSRV
 package = "uuid-rng-internal"
-version = "1.21.0"
+version = "1.22.0"
 path = "rng"
 optional = true
 
 # Private
 [target.'cfg(not(all(target_arch = "wasm32", any(target_os = "unknown", target_os = "none"))))'.dependencies.rand]
-version = "0.9"
+version = "0.10"
 optional = true
 
 # Private

data/rust-vendor/uuid/README.md

@@ -28,7 +28,7 @@ Add the following to your `Cargo.toml`:
 
 ```toml
 [dependencies.uuid]
-version = "1.21.0"
+version = "1.22.0"
 # Lets you generate random UUIDs
 features = [
     "v4",
@@ -64,11 +64,11 @@ assert_eq!(Some(Version::Random), my_uuid.get_version());
 If you'd like to parse UUIDs _really_ fast, check out the [`uuid-simd`](https://github.com/nugine/uuid-simd)
 library.
 
-For more details on using `uuid`, [see the library documentation](https://docs.rs/uuid/1.21.0/uuid).
+For more details on using `uuid`, [see the library documentation](https://docs.rs/uuid/1.22.0/uuid).
 
 ## References
 
-* [`uuid` library docs](https://docs.rs/uuid/1.21.0/uuid).
+* [`uuid` library docs](https://docs.rs/uuid/1.22.0/uuid).
 * [Wikipedia: Universally Unique Identifier](http://en.wikipedia.org/wiki/Universally_unique_identifier).
 * [RFC 9562: Universally Unique IDentifiers (UUID)](https://www.ietf.org/rfc/rfc9562.html).
 

data/rust-vendor/uuid/src/lib.rs

@@ -38,7 +38,7 @@
 //!
 //! ```toml
 //! [dependencies.uuid]
-//! version = "1.21.0"
+//! version = "1.22.0"
 //! # Lets you generate random UUIDs
 //! features = [
 //!     "v4",
@@ -138,7 +138,7 @@
 //!
 //! ```toml
 //! [dependencies.uuid]
-//! version = "1.21.0"
+//! version = "1.22.0"
 //! features = [
 //!     "v4",
 //!     "v7",
@@ -153,7 +153,7 @@
 //!
 //! ```toml
 //! [dependencies.uuid]
-//! version = "1.21.0"
+//! version = "1.22.0"
 //! default-features = false
 //! ```
 //!
@@ -211,7 +211,7 @@
 #![doc(
     html_logo_url = "https://www.rust-lang.org/logos/rust-logo-128x128-blk-v2.png",
     html_favicon_url = "https://www.rust-lang.org/favicon.ico",
-    html_root_url = "https://docs.rs/uuid/1.21.0"
+    html_root_url = "https://docs.rs/uuid/1.22.0"
 )]
 
 #[cfg(any(feature = "std", test))]

data/sig/html_to_markdown.rbs

@@ -185,6 +185,18 @@ module HtmlToMarkdown
     structured_data: Array[structured_data]
   }
 
+  type table_data = {
+    cells: Array[Array[String]],
+    markdown: String,
+    is_header_row: Array[bool]
+  }
+
+  type table_extraction_result = {
+    content: String,
+    metadata: extended_metadata?,
+    tables: Array[table_data]
+  }
+
   # Native methods (implemented in Rust via Magnus/rb-sys)
   # These are aliased from the Rust extension and available as both module and instance methods
   private
@@ -217,6 +229,11 @@ module HtmlToMarkdown
     conversion_options? options,
     visitor? visitor
   ) -> String
+  def self.native_convert_with_tables: (
+    String html,
+    conversion_options? options,
+    metadata_config? metadata_config
+  ) -> table_extraction_result
 
   def native_convert: (String html, conversion_options? options) -> String
   def native_options: (conversion_options? options_hash) -> Options
@@ -246,6 +263,11 @@ module HtmlToMarkdown
     conversion_options? options,
     visitor? visitor
   ) -> String
+  def native_convert_with_tables: (
+    String html,
+    conversion_options? options,
+    metadata_config? metadata_config
+  ) -> table_extraction_result
 
   # Visitor interface for customizing conversion behavior
   type visitor = Object
@@ -422,6 +444,25 @@ module HtmlToMarkdown
     ?metadata_config metadata_config
   ) -> [String, extended_metadata]
 
+  # Convert HTML and extract tables as structured data
+  #
+  # Args:
+  #   html: HTML string to convert
+  #   options: Optional conversion configuration
+  #   metadata_config: Optional metadata extraction configuration
+  #
+  # Returns:
+  #   table_extraction_result: Hash containing content, metadata, and tables array
+  #
+  # Example:
+  #   result = HtmlToMarkdown.convert_with_tables(html)
+  #   puts result[:tables].length
+  def self.convert_with_tables: (
+    String html,
+    ?conversion_options options,
+    ?metadata_config metadata_config
+  ) -> table_extraction_result
+
   # Instance method versions (created by module_function)
   def convert: (String html, ?conversion_options options, ?visitor visitor) -> String
   def options: (?conversion_options options_hash) -> Options
@@ -449,4 +490,9 @@ module HtmlToMarkdown
     Options options_handle,
     ?metadata_config metadata_config
   ) -> [String, extended_metadata]
+  def convert_with_tables: (
+    String html,
+    ?conversion_options options,
+    ?metadata_config metadata_config
+  ) -> table_extraction_result
 end

data/spec/convert_with_tables_spec.rb

@@ -0,0 +1,194 @@
+# frozen_string_literal: true
+
+require 'spec_helper'
+
+RSpec.describe HtmlToMarkdown do
+  describe '.convert_with_tables' do
+    it 'returns a hash with content, metadata, and tables keys' do
+      html = '<table><tr><td>Cell</td></tr></table>'
+      result = described_class.convert_with_tables(html)
+
+      expect(result).to be_a(Hash)
+      expect(result).to include(:content, :metadata, :tables)
+    end
+
+    context 'with a basic table with header' do
+      let(:html) do
+        <<~HTML
+          <table>
+            <thead>
+              <tr><th>Name</th><th>Age</th></tr>
+            </thead>
+            <tbody>
+              <tr><td>Alice</td><td>30</td></tr>
+            </tbody>
+          </table>
+        HTML
+      end
+
+      it 'extracts exactly one table' do
+        result = described_class.convert_with_tables(html)
+
+        expect(result[:tables].length).to eq(1)
+      end
+
+      it 'extracts cells as rows of columns' do
+        result = described_class.convert_with_tables(html)
+        table = result[:tables][0]
+
+        expect(table[:cells]).to be_an(Array)
+        expect(table[:cells].length).to eq(2)
+        expect(table[:cells][0]).to eq(%w[Name Age])
+        expect(table[:cells][1]).to eq(%w[Alice 30])
+      end
+
+      it 'provides markdown representation' do
+        result = described_class.convert_with_tables(html)
+        table = result[:tables][0]
+
+        expect(table[:markdown]).to be_a(String)
+        expect(table[:markdown]).to include('Name')
+        expect(table[:markdown]).to include('Alice')
+      end
+
+      it 'marks header rows correctly' do
+        result = described_class.convert_with_tables(html)
+        table = result[:tables][0]
+
+        expect(table[:is_header_row]).to be_an(Array)
+        expect(table[:is_header_row].length).to eq(2)
+        expect(table[:is_header_row][0]).to be true
+        expect(table[:is_header_row][1]).to be false
+      end
+
+      it 'includes converted markdown content' do
+        result = described_class.convert_with_tables(html)
+
+        expect(result[:content]).to be_a(String)
+        expect(result[:content]).not_to be_empty
+      end
+    end
+
+    context 'with empty HTML' do
+      it 'returns empty tables array' do
+        result = described_class.convert_with_tables('')
+
+        expect(result[:tables]).to eq([])
+        expect(result[:content]).to be_a(String)
+      end
+    end
+
+    context 'with HTML containing no tables' do
+      it 'returns empty tables array' do
+        html = '<p>No tables here</p>'
+        result = described_class.convert_with_tables(html)
+
+        expect(result[:tables]).to eq([])
+        expect(result[:content]).to include('No tables here')
+      end
+    end
+
+    context 'with multiple tables' do
+      let(:html) do
+        <<~HTML
+          <table>
+            <tr><th>A</th></tr>
+            <tr><td>1</td></tr>
+          </table>
+          <p>Some text between tables</p>
+          <table>
+            <tr><th>B</th><th>C</th></tr>
+            <tr><td>2</td><td>3</td></tr>
+            <tr><td>4</td><td>5</td></tr>
+          </table>
+        HTML
+      end
+
+      it 'extracts all tables' do
+        result = described_class.convert_with_tables(html)
+
+        expect(result[:tables].length).to eq(2)
+      end
+
+      it 'preserves table order' do
+        result = described_class.convert_with_tables(html)
+
+        first_table = result[:tables][0]
+        second_table = result[:tables][1]
+
+        expect(first_table[:cells][0]).to eq(['A'])
+        expect(second_table[:cells][0]).to eq(%w[B C])
+      end
+
+      it 'extracts correct row counts per table' do
+        result = described_class.convert_with_tables(html)
+
+        expect(result[:tables][0][:cells].length).to eq(2)
+        expect(result[:tables][1][:cells].length).to eq(3)
+      end
+    end
+
+    context 'with special characters in cells' do
+      let(:html) do
+        <<~HTML
+          <table>
+            <tr><th>Key</th><th>Value</th></tr>
+            <tr><td>Brackets &lt;&gt;</td><td>Ampersand &amp;</td></tr>
+            <tr><td>Quotes "double"</td><td>Quotes 'single'</td></tr>
+            <tr><td>Unicode: cafe\u0301</td><td>Emoji: test</td></tr>
+          </table>
+        HTML
+      end
+
+      it 'handles HTML entities in cells' do
+        result = described_class.convert_with_tables(html)
+        table = result[:tables][0]
+
+        expect(table[:cells][1][0]).to include('<>')
+        expect(table[:cells][1][1]).to include('&')
+      end
+
+      it 'handles quotes in cells' do
+        result = described_class.convert_with_tables(html)
+        table = result[:tables][0]
+
+        expect(table[:cells][2][0]).to include('"double"')
+        expect(table[:cells][2][1]).to include("'single'")
+      end
+
+      it 'handles unicode in cells' do
+        result = described_class.convert_with_tables(html)
+        table = result[:tables][0]
+
+        expect(table[:cells][3][0]).to be_a(String)
+      end
+    end
+
+    context 'with conversion options' do
+      it 'accepts options hash' do
+        html = '<table><tr><th>Header</th></tr><tr><td>Data</td></tr></table>'
+        result = described_class.convert_with_tables(html, { heading_style: :atx })
+
+        expect(result).to be_a(Hash)
+        expect(result[:tables].length).to eq(1)
+      end
+
+      it 'accepts nil options' do
+        html = '<table><tr><td>Data</td></tr></table>'
+        result = described_class.convert_with_tables(html, nil, nil)
+
+        expect(result).to be_a(Hash)
+        expect(result[:tables].length).to eq(1)
+      end
+    end
+
+    context 'with metadata config' do
+      it 'includes metadata when configured' do
+        html = '<html><head><title>Test</title></head><body><table><tr><td>Data</td></tr></table></body></html>'
+        result = described_class.convert_with_tables(html, nil, { extract_headers: true })
+
+        expect(result[:metadata]).to be_a(Hash).or(be_nil)
+      end
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: html-to-markdown
 version: !ruby/object:Gem::Version
-  version: 2.27.3
+  version: 2.28.1
 platform: ruby
 authors:
 - Na'aman Hirschfeld
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-03-05 00:00:00.000000000 Z
+date: 2026-03-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys
@@ -61,6 +61,7 @@ files:
 - ext/html-to-markdown-rb/native/src/conversion/inline_images.rs
 - ext/html-to-markdown-rb/native/src/conversion/metadata.rs
 - ext/html-to-markdown-rb/native/src/conversion/mod.rs
+- ext/html-to-markdown-rb/native/src/conversion/tables.rs
 - ext/html-to-markdown-rb/native/src/lib.rs
 - ext/html-to-markdown-rb/native/src/options.rs
 - ext/html-to-markdown-rb/native/src/profiling.rs
@@ -1958,6 +1959,7 @@ files:
 - rust-vendor/html-to-markdown-rs/tests/tables_test.rs
 - rust-vendor/html-to-markdown-rs/tests/test_custom_elements.rs
 - rust-vendor/html-to-markdown-rs/tests/test_issue_187.rs
+- rust-vendor/html-to-markdown-rs/tests/test_issue_218.rs
 - rust-vendor/html-to-markdown-rs/tests/test_nested_simple.rs
 - rust-vendor/html-to-markdown-rs/tests/test_script_style_stripping.rs
 - rust-vendor/html-to-markdown-rs/tests/test_spa_bisect.rs
@@ -9737,6 +9739,7 @@ files:
 - sig/open3.rbs
 - spec/cli_proxy_spec.rb
 - spec/convert_spec.rb
+- spec/convert_with_tables_spec.rb
 - spec/metadata_extraction_spec.rb
 - spec/spec_helper.rb
 - spec/visitor_issue_187_spec.rb