require 'rails_helper' require 'html_to_markdown' describe HtmlToMarkdown do def html_to_markdown(html, opts = {}) HtmlToMarkdown.new(html, opts).to_markdown end it "remove whitespaces" do expect(html_to_markdown(<<-HTML
Hello,

    This is the 1st paragraph.   

        This is another paragraph
HTML )).to eq("Hello,\n\nThis is the 1st paragraph.\n\nThis is another paragraph") end it "skips hidden tags" do expect(html_to_markdown(%Q{

Hello cruel World!

})).to eq("Hello World!") end it "converts " do expect(html_to_markdown("Strong")).to eq("**Strong**") expect(html_to_markdown("Str*ng")).to eq("__Str*ng__") end it "converts " do expect(html_to_markdown("Bold")).to eq("**Bold**") expect(html_to_markdown("B*ld")).to eq("__B*ld__") end it "converts " do expect(html_to_markdown("Emphasis")).to eq("*Emphasis*") expect(html_to_markdown("Emph*sis")).to eq("_Emph*sis_") end it "converts " do expect(html_to_markdown("Italic")).to eq("*Italic*") expect(html_to_markdown("It*lic")).to eq("_It*lic_") end it "converts " do expect(html_to_markdown(%Q{Discourse})).to eq("[Discourse](https://www.discourse.org)") end it "removes empty & invalid " do expect(html_to_markdown(%Q{Discourse})).to eq("Discourse") expect(html_to_markdown(%Q{Discourse})).to eq("Discourse") expect(html_to_markdown(%Q{Discourse})).to eq("Discourse") end HTML_WITH_IMG ||= %Q{Discourse Logo} HTML_WITH_CID_IMG ||= %Q{Discourse Logo} it "converts " do expect(html_to_markdown(HTML_WITH_IMG)).to eq("![Discourse Logo](https://www.discourse.org/logo.svg)") end it "keeps with 'keep_img_tags'" do expect(html_to_markdown(HTML_WITH_IMG, keep_img_tags: true)).to eq(HTML_WITH_IMG) end it "removes empty & invalid " do expect(html_to_markdown(%Q{})).to eq("") expect(html_to_markdown(%Q{})).to eq("") expect(html_to_markdown(%Q{})).to eq("") end it "keeps with src='cid:' whith 'keep_cid_imgs'" do expect(html_to_markdown(HTML_WITH_CID_IMG, keep_cid_imgs: true)).to eq("![Discourse Logo](cid:ii_1525434659ddb4cb)") expect(html_to_markdown(HTML_WITH_CID_IMG, keep_img_tags: true, keep_cid_imgs: true)).to eq("\"Discourse") end it "skips hidden " do expect(html_to_markdown(%Q{})).to eq("") expect(html_to_markdown(%Q{})).to eq("") expect(html_to_markdown(%Q{})).to eq("") expect(html_to_markdown(%Q{})).to eq("") end (1..6).each do |n| it "converts " do expect(html_to_markdown("Header #{n}")).to eq("#" * n + " Header #{n}") end end it "converts
" do expect(html_to_markdown("Before
Inside
After")).to eq("Before\nInside\nAfter") end it "converts
" do expect(html_to_markdown("Before
Inside
After")).to eq("Before\n\n---\n\nInside\n\n---\n\nAfter") end it "converts " do expect(html_to_markdown("Teletype")).to eq("`Teletype`") end it "converts " do expect(html_to_markdown("Code")).to eq("`Code`") end it "supports " do expect(html_to_markdown("This is an insertion")).to eq("This is an insertion") end it "supports " do expect(html_to_markdown("This is a deletion")).to eq("This is a deletion") end it "supports " do expect(html_to_markdown("H2O")).to eq("H2O") end it "supports " do expect(html_to_markdown("Super Script!")).to eq("Super Script!") end it "supports " do expect(html_to_markdown("Small")).to eq("Small") end it "supports " do expect(html_to_markdown("CTRL+C")).to eq("CTRL+C") end it "supports " do expect(html_to_markdown(%Q{CDCK})).to eq(%Q{CDCK}) end it "supports " do expect(html_to_markdown("Strike Through")).to eq("Strike Through") end it "supports " do expect(html_to_markdown("Strike Through")).to eq("Strike Through") end it "supports
" do expect(html_to_markdown("
Quote
")).to eq("> Quote") end it "supports
    " do expect(html_to_markdown("
    • 🍏
    • 🍐
    • 🍌
    ")).to eq("- 🍏\n- 🍐\n- 🍌") expect(html_to_markdown("
      \n
    • 🍏
    • \n
    • 🍐
    • \n
    • 🍌
    • \n
    ")).to eq("- 🍏\n- 🍐\n- 🍌") end it "supports
      " do expect(html_to_markdown("
      1. 🍆
      2. 🍅
      3. 🍄
      ")).to eq("1. 🍆\n1. 🍅\n1. 🍄") end it "supports

      inside

    1. " do expect(html_to_markdown("
      • 🍏

      • 🍐

      • 🍌

      ")).to eq("- 🍏\n\n- 🍐\n\n- 🍌") end it "supports
        inside
          " do expect(html_to_markdown(<<-HTML
          • Fruits
            • 🍏
            • 🍐
            • 🍌
          • Vegetables
            • 🍆
            • 🍅
            • 🍄
          HTML )).to eq("- Fruits\n - 🍏\n - 🍐\n - 🍌\n- Vegetables\n - 🍆\n - 🍅\n - 🍄") end it "supports bare
        • " do expect(html_to_markdown("
        • I'm alone
        • ")).to eq("- I'm alone") end it "supports
          " do
              expect(html_to_markdown("
          var foo = 'bar';
          ")).to eq("```\nvar foo = 'bar';\n```") expect(html_to_markdown("
          var foo = 'bar';
          ")).to eq("```\nvar foo = 'bar';\n```") expect(html_to_markdown(%Q{
          var foo = 'bar';
          })).to eq("```javascript\nvar foo = 'bar';\n```") end it "works" do expect(html_to_markdown("
          • A list item with a blockquote:

            This is a blockquote
            inside a list item.

          ")).to eq("- A list item with a blockquote:\n\n > This is a **blockquote**\n > inside a list item.") end it "supports html document" do expect(html_to_markdown("Hello
          World
          ")).to eq("Hello\nWorld") end it "handles

          " do expect(html_to_markdown("

          1st paragraph

          2nd paragraph

          ")).to eq("1st paragraph\n\n2nd paragraph") end it "handles
          " do expect(html_to_markdown("
          1st div
          2nd div
          ")).to eq("1st div\n\n2nd div") end it "swallows " do expect(html_to_markdown("Span")).to eq("Span") end it "swallows " do expect(html_to_markdown("Underline")).to eq("Underline") end it "removes ")).to eq("") end it "removes ")).to eq("") end it "handles divs within spans" do html = "
          1st paragraph
          2nd paragraph
          " expect(html_to_markdown(html)).to eq("1st paragraph\n2nd paragraph") end end