@refraction-ui/astro 0.4.2 → 0.4.5

package/dist/rich-editor/serialization.ts

@@ -0,0 +1,711 @@
+/**
+ * Serialization — convert documents to/from Markdown, HTML, and plain text.
+ */
+
+import type {
+  Document,
+  Block,
+  InlineContent,
+  TextSegment,
+  Mark,
+  MarkType,
+  BlockType,
+} from './model.js'
+import {
+  createDocument,
+  createBlock,
+  createTextSegment,
+  createMentionSegment,
+  createEmojiSegment,
+  normalizeContent,
+} from './model.js'
+import { EMOJI_MAP } from './emoji.js'
+
+// ---------------------------------------------------------------------------
+// To Markdown
+// ---------------------------------------------------------------------------
+
+function marksToMarkdown(text: string, marks: Mark[]): string {
+  let result = text
+
+  // Sort marks so that wrapping is consistent: link outermost, then bold, italic, etc.
+  const sorted = [...marks].sort((a, b) => {
+    const order: MarkType[] = ['link', 'bold', 'italic', 'strikethrough', 'code', 'underline']
+    return order.indexOf(a.type) - order.indexOf(b.type)
+  })
+
+  for (const mark of sorted) {
+    switch (mark.type) {
+      case 'bold':
+        result = `**${result}**`
+        break
+      case 'italic':
+        result = `*${result}*`
+        break
+      case 'strikethrough':
+        result = `~~${result}~~`
+        break
+      case 'code':
+        result = `\`${result}\``
+        break
+      case 'underline':
+        // No standard markdown for underline; use HTML
+        result = `<u>${result}</u>`
+        break
+      case 'link':
+        result = `[${result}](${mark.attrs?.href ?? ''})`
+        break
+    }
+  }
+
+  return result
+}
+
+function inlineContentToMarkdown(content: InlineContent[]): string {
+  return content
+    .map((seg) => {
+      if (seg.type === 'text') {
+        if (seg.text === '' && seg.marks.length === 0) return ''
+        return marksToMarkdown(seg.text, seg.marks)
+      }
+      if (seg.type === 'mention') {
+        return seg.label
+      }
+      if (seg.type === 'emoji') {
+        return seg.shortcode
+      }
+      return ''
+    })
+    .join('')
+}
+
+function blockToMarkdown(block: Block): string {
+  const content = inlineContentToMarkdown(block.content)
+
+  switch (block.type) {
+    case 'paragraph':
+      return content
+    case 'heading': {
+      const level = (block.meta?.level as number) ?? 1
+      const prefix = '#'.repeat(level)
+      return `${prefix} ${content}`
+    }
+    case 'blockquote':
+      return `> ${content}`
+    case 'code-block': {
+      const lang = (block.meta?.language as string) ?? ''
+      return `\`\`\`${lang}\n${content}\n\`\`\``
+    }
+    case 'list-item': {
+      const listType = block.meta?.listType as string
+      const indent = '  '.repeat(block.indent ?? 0)
+      if (listType === 'ordered') {
+        return `${indent}1. ${content}`
+      }
+      return `${indent}- ${content}`
+    }
+    case 'divider':
+      return '---'
+    case 'image': {
+      const src = (block.meta?.src as string) ?? ''
+      const alt = (block.meta?.alt as string) ?? ''
+      return `![${alt}](${src})`
+    }
+    default:
+      return content
+  }
+}
+
+export function toMarkdown(doc: Document): string {
+  return doc.blocks.map(blockToMarkdown).join('\n\n')
+}
+
+// ---------------------------------------------------------------------------
+// From Markdown
+// ---------------------------------------------------------------------------
+
+function parseInlineMarkdown(text: string): InlineContent[] {
+  const segments: InlineContent[] = []
+  let remaining = text
+
+  // Process inline patterns
+  // We use a priority-based approach: longer/more complex patterns first
+  while (remaining.length > 0) {
+    let matched = false
+
+    // Bold: **text** or __text__
+    const boldMatch = remaining.match(/^\*\*(.+?)\*\*/) || remaining.match(/^__(.+?)__/)
+    if (boldMatch) {
+      segments.push(createTextSegment(boldMatch[1], [{ type: 'bold' }]))
+      remaining = remaining.slice(boldMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // Strikethrough: ~~text~~
+    const strikeMatch = remaining.match(/^~~(.+?)~~/)
+    if (strikeMatch) {
+      segments.push(createTextSegment(strikeMatch[1], [{ type: 'strikethrough' }]))
+      remaining = remaining.slice(strikeMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // Italic: *text* or _text_
+    const italicMatch = remaining.match(/^\*(.+?)\*/) || remaining.match(/^_(.+?)_/)
+    if (italicMatch) {
+      segments.push(createTextSegment(italicMatch[1], [{ type: 'italic' }]))
+      remaining = remaining.slice(italicMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // Inline code: `text`
+    const codeMatch = remaining.match(/^`(.+?)`/)
+    if (codeMatch) {
+      segments.push(createTextSegment(codeMatch[1], [{ type: 'code' }]))
+      remaining = remaining.slice(codeMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // Link: [text](url)
+    const linkMatch = remaining.match(/^\[([^\]]+)\]\(([^)]+)\)/)
+    if (linkMatch) {
+      segments.push(
+        createTextSegment(linkMatch[1], [{ type: 'link', attrs: { href: linkMatch[2] } }]),
+      )
+      remaining = remaining.slice(linkMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // Mention: @username (word characters after @)
+    const mentionMatch = remaining.match(/^@(\w+)/)
+    if (mentionMatch) {
+      segments.push(createMentionSegment(mentionMatch[1], `@${mentionMatch[1]}`))
+      remaining = remaining.slice(mentionMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // Emoji shortcode: :name:
+    const emojiMatch = remaining.match(/^:([a-z0-9_]+):/)
+    if (emojiMatch) {
+      const shortcode = `:${emojiMatch[1]}:`
+      const unicode = EMOJI_MAP[shortcode]
+      if (unicode) {
+        segments.push(createEmojiSegment(shortcode, unicode))
+        remaining = remaining.slice(emojiMatch[0].length)
+        matched = true
+        continue
+      }
+    }
+
+    // Underline (HTML): <u>text</u>
+    const underlineMatch = remaining.match(/^<u>(.+?)<\/u>/)
+    if (underlineMatch) {
+      segments.push(createTextSegment(underlineMatch[1], [{ type: 'underline' }]))
+      remaining = remaining.slice(underlineMatch[0].length)
+      matched = true
+      continue
+    }
+
+    if (!matched) {
+      // Consume one character as plain text
+      // Aggregate consecutive plain characters
+      let endIdx = 1
+      while (endIdx < remaining.length) {
+        const ch = remaining[endIdx]
+        if (ch === '*' || ch === '~' || ch === '`' || ch === '[' || ch === '@' || ch === ':' || ch === '_' || ch === '<') break
+        endIdx++
+      }
+      segments.push(createTextSegment(remaining.slice(0, endIdx)))
+      remaining = remaining.slice(endIdx)
+    }
+  }
+
+  return segments.length > 0 ? normalizeContent(segments) : [createTextSegment('')]
+}
+
+function parseMarkdownLine(line: string, isInCodeBlock: boolean): Block | null {
+  if (isInCodeBlock) return null // handled by caller
+
+  // Heading
+  const headingMatch = line.match(/^(#{1,3}) (.*)/)
+  if (headingMatch) {
+    const level = headingMatch[1].length as 1 | 2 | 3
+    return createBlock('heading', parseInlineMarkdown(headingMatch[2]), { level })
+  }
+
+  // Blockquote
+  const quoteMatch = line.match(/^> (.*)/)
+  if (quoteMatch) {
+    return createBlock('blockquote', parseInlineMarkdown(quoteMatch[1]))
+  }
+
+  // Divider
+  if (/^---+$/.test(line.trim())) {
+    return createBlock('divider')
+  }
+
+  // Ordered list
+  const orderedMatch = line.match(/^(\s*)(\d+)\. (.*)/)
+  if (orderedMatch) {
+    const indent = Math.floor(orderedMatch[1].length / 2)
+    const block = createBlock('list-item', parseInlineMarkdown(orderedMatch[3]), { listType: 'ordered' })
+    if (indent > 0) block.indent = indent
+    return block
+  }
+
+  // Unordered list
+  const unorderedMatch = line.match(/^(\s*)[-*] (.*)/)
+  if (unorderedMatch) {
+    const indent = Math.floor(unorderedMatch[1].length / 2)
+    const block = createBlock('list-item', parseInlineMarkdown(unorderedMatch[2]), { listType: 'bullet' })
+    if (indent > 0) block.indent = indent
+    return block
+  }
+
+  // Image
+  const imageMatch = line.match(/^!\[([^\]]*)\]\(([^)]+)\)/)
+  if (imageMatch) {
+    return createBlock('image', [createTextSegment('')], { alt: imageMatch[1], src: imageMatch[2] })
+  }
+
+  // Empty line
+  if (line.trim() === '') {
+    return null
+  }
+
+  // Paragraph (default)
+  return createBlock('paragraph', parseInlineMarkdown(line))
+}
+
+export function fromMarkdown(markdown: string): Document {
+  const lines = markdown.split('\n')
+  const blocks: Block[] = []
+  let i = 0
+
+  while (i < lines.length) {
+    const line = lines[i]
+
+    // Code block
+    if (line.startsWith('```')) {
+      const lang = line.slice(3).trim()
+      const codeLines: string[] = []
+      i++
+      while (i < lines.length && !lines[i].startsWith('```')) {
+        codeLines.push(lines[i])
+        i++
+      }
+      i++ // skip closing ```
+      blocks.push(
+        createBlock('code-block', [createTextSegment(codeLines.join('\n'))], lang ? { language: lang } : undefined),
+      )
+      continue
+    }
+
+    const block = parseMarkdownLine(line, false)
+    if (block) {
+      blocks.push(block)
+    }
+    i++
+  }
+
+  return createDocument(blocks.length > 0 ? blocks : undefined)
+}
+
+// ---------------------------------------------------------------------------
+// To HTML
+// ---------------------------------------------------------------------------
+
+function escapeHtml(text: string): string {
+  return text
+    .replace(/&/g, '&amp;')
+    .replace(/</g, '&lt;')
+    .replace(/>/g, '&gt;')
+    .replace(/"/g, '&quot;')
+}
+
+function marksToHtml(text: string, marks: Mark[]): string {
+  let result = escapeHtml(text)
+
+  for (const mark of marks) {
+    switch (mark.type) {
+      case 'bold':
+        result = `<strong>${result}</strong>`
+        break
+      case 'italic':
+        result = `<em>${result}</em>`
+        break
+      case 'strikethrough':
+        result = `<del>${result}</del>`
+        break
+      case 'code':
+        result = `<code>${result}</code>`
+        break
+      case 'underline':
+        result = `<u>${result}</u>`
+        break
+      case 'link':
+        result = `<a href="${escapeHtml(mark.attrs?.href ?? '')}">${result}</a>`
+        break
+    }
+  }
+
+  return result
+}
+
+function inlineContentToHtml(content: InlineContent[]): string {
+  return content
+    .map((seg) => {
+      if (seg.type === 'text') {
+        return marksToHtml(seg.text, seg.marks)
+      }
+      if (seg.type === 'mention') {
+        return `<span class="mention" data-mention-id="${escapeHtml(seg.id)}">${escapeHtml(seg.label)}</span>`
+      }
+      if (seg.type === 'emoji') {
+        return `<span class="emoji" data-shortcode="${escapeHtml(seg.shortcode)}">${seg.unicode}</span>`
+      }
+      return ''
+    })
+    .join('')
+}
+
+function blockToHtml(block: Block): string {
+  const content = inlineContentToHtml(block.content)
+
+  switch (block.type) {
+    case 'paragraph':
+      return `<p>${content}</p>`
+    case 'heading': {
+      const level = (block.meta?.level as number) ?? 1
+      return `<h${level}>${content}</h${level}>`
+    }
+    case 'blockquote':
+      return `<blockquote>${content}</blockquote>`
+    case 'code-block': {
+      const lang = block.meta?.language as string
+      const langAttr = lang ? ` class="language-${escapeHtml(lang)}"` : ''
+      return `<pre><code${langAttr}>${content}</code></pre>`
+    }
+    case 'list-item':
+      return `<li>${content}</li>`
+    case 'divider':
+      return '<hr>'
+    case 'image': {
+      const src = (block.meta?.src as string) ?? ''
+      const alt = (block.meta?.alt as string) ?? ''
+      return `<img src="${escapeHtml(src)}" alt="${escapeHtml(alt)}">`
+    }
+    default:
+      return `<p>${content}</p>`
+  }
+}
+
+export function toHTML(doc: Document): string {
+  // Group list items into <ul> or <ol>
+  const parts: string[] = []
+  let i = 0
+
+  while (i < doc.blocks.length) {
+    const block = doc.blocks[i]
+
+    if (block.type === 'list-item') {
+      const listType = block.meta?.listType === 'ordered' ? 'ol' : 'ul'
+      const items: string[] = []
+      while (i < doc.blocks.length && doc.blocks[i].type === 'list-item') {
+        items.push(blockToHtml(doc.blocks[i]))
+        i++
+      }
+      parts.push(`<${listType}>${items.join('')}</${listType}>`)
+    } else {
+      parts.push(blockToHtml(block))
+      i++
+    }
+  }
+
+  return parts.join('')
+}
+
+// ---------------------------------------------------------------------------
+// From HTML (basic)
+// ---------------------------------------------------------------------------
+
+function parseHtmlInlineContent(html: string): InlineContent[] {
+  const segments: InlineContent[] = []
+  let remaining = html
+
+  while (remaining.length > 0) {
+    let matched = false
+
+    // <strong>text</strong>
+    const boldMatch = remaining.match(/^<strong>(.*?)<\/strong>/)
+    if (boldMatch) {
+      segments.push(createTextSegment(unescapeHtml(boldMatch[1]), [{ type: 'bold' }]))
+      remaining = remaining.slice(boldMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <em>text</em>
+    const emMatch = remaining.match(/^<em>(.*?)<\/em>/)
+    if (emMatch) {
+      segments.push(createTextSegment(unescapeHtml(emMatch[1]), [{ type: 'italic' }]))
+      remaining = remaining.slice(emMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <del>text</del>
+    const delMatch = remaining.match(/^<del>(.*?)<\/del>/)
+    if (delMatch) {
+      segments.push(createTextSegment(unescapeHtml(delMatch[1]), [{ type: 'strikethrough' }]))
+      remaining = remaining.slice(delMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <code>text</code>
+    const codeMatch = remaining.match(/^<code>(.*?)<\/code>/)
+    if (codeMatch) {
+      segments.push(createTextSegment(unescapeHtml(codeMatch[1]), [{ type: 'code' }]))
+      remaining = remaining.slice(codeMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <u>text</u>
+    const uMatch = remaining.match(/^<u>(.*?)<\/u>/)
+    if (uMatch) {
+      segments.push(createTextSegment(unescapeHtml(uMatch[1]), [{ type: 'underline' }]))
+      remaining = remaining.slice(uMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <a href="url">text</a>
+    const linkMatch = remaining.match(/^<a href="([^"]*)">(.*?)<\/a>/)
+    if (linkMatch) {
+      segments.push(
+        createTextSegment(unescapeHtml(linkMatch[2]), [{ type: 'link', attrs: { href: unescapeHtml(linkMatch[1]) } }]),
+      )
+      remaining = remaining.slice(linkMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <span class="mention" data-mention-id="id">label</span>
+    const mentionMatch = remaining.match(
+      /^<span class="mention" data-mention-id="([^"]*)">(.*?)<\/span>/,
+    )
+    if (mentionMatch) {
+      segments.push(createMentionSegment(unescapeHtml(mentionMatch[1]), unescapeHtml(mentionMatch[2])))
+      remaining = remaining.slice(mentionMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <span class="emoji" data-shortcode=":name:">unicode</span>
+    const emojiMatch = remaining.match(
+      /^<span class="emoji" data-shortcode="([^"]*)">(.*?)<\/span>/,
+    )
+    if (emojiMatch) {
+      segments.push(createEmojiSegment(unescapeHtml(emojiMatch[1]), emojiMatch[2]))
+      remaining = remaining.slice(emojiMatch[0].length)
+      matched = true
+      continue
+    }
+
+    if (!matched) {
+      // Plain text until next tag
+      const nextTagIdx = remaining.indexOf('<', 1)
+      if (nextTagIdx === -1) {
+        segments.push(createTextSegment(unescapeHtml(remaining)))
+        break
+      }
+      // Check if the first char is '<' (unknown tag)
+      if (remaining[0] === '<') {
+        // Skip unknown tag
+        const closeIdx = remaining.indexOf('>')
+        if (closeIdx !== -1) {
+          remaining = remaining.slice(closeIdx + 1)
+        } else {
+          remaining = remaining.slice(1)
+        }
+      } else {
+        segments.push(createTextSegment(unescapeHtml(remaining.slice(0, nextTagIdx))))
+        remaining = remaining.slice(nextTagIdx)
+      }
+    }
+  }
+
+  return segments.length > 0 ? normalizeContent(segments) : [createTextSegment('')]
+}
+
+function unescapeHtml(text: string): string {
+  return text
+    .replace(/&amp;/g, '&')
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&quot;/g, '"')
+}
+
+export function fromHTML(html: string): Document {
+  const blocks: Block[] = []
+  let remaining = html
+
+  while (remaining.length > 0) {
+    let matched = false
+
+    // <h1>...</h1>, <h2>...</h2>, <h3>...</h3>
+    const headingMatch = remaining.match(/^<h([123])>(.*?)<\/h\1>/)
+    if (headingMatch) {
+      const level = parseInt(headingMatch[1]) as 1 | 2 | 3
+      blocks.push(createBlock('heading', parseHtmlInlineContent(headingMatch[2]), { level }))
+      remaining = remaining.slice(headingMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <p>...</p>
+    const pMatch = remaining.match(/^<p>(.*?)<\/p>/)
+    if (pMatch) {
+      blocks.push(createBlock('paragraph', parseHtmlInlineContent(pMatch[1])))
+      remaining = remaining.slice(pMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <blockquote>...</blockquote>
+    const bqMatch = remaining.match(/^<blockquote>(.*?)<\/blockquote>/)
+    if (bqMatch) {
+      blocks.push(createBlock('blockquote', parseHtmlInlineContent(bqMatch[1])))
+      remaining = remaining.slice(bqMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <pre><code...>...</code></pre>
+    const preMatch = remaining.match(/^<pre><code(?: class="language-([^"]*)")?>(.*?)<\/code><\/pre>/)
+    if (preMatch) {
+      const lang = preMatch[1]
+      blocks.push(
+        createBlock('code-block', [createTextSegment(unescapeHtml(preMatch[2]))], lang ? { language: lang } : undefined),
+      )
+      remaining = remaining.slice(preMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <ul>...</ul>
+    const ulMatch = remaining.match(/^<ul>(.*?)<\/ul>/)
+    if (ulMatch) {
+      const liMatches = ulMatch[1].matchAll(/<li>(.*?)<\/li>/g)
+      for (const li of liMatches) {
+        blocks.push(createBlock('list-item', parseHtmlInlineContent(li[1]), { listType: 'bullet' }))
+      }
+      remaining = remaining.slice(ulMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <ol>...</ol>
+    const olMatch = remaining.match(/^<ol>(.*?)<\/ol>/)
+    if (olMatch) {
+      const liMatches = olMatch[1].matchAll(/<li>(.*?)<\/li>/g)
+      for (const li of liMatches) {
+        blocks.push(createBlock('list-item', parseHtmlInlineContent(li[1]), { listType: 'ordered' }))
+      }
+      remaining = remaining.slice(olMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <hr> or <hr/>
+    const hrMatch = remaining.match(/^<hr\s*\/?>/)
+    if (hrMatch) {
+      blocks.push(createBlock('divider'))
+      remaining = remaining.slice(hrMatch[0].length)
+      matched = true
+      continue
+    }
+
+    // <img src="..." alt="...">
+    const imgMatch = remaining.match(/^<img src="([^"]*)" alt="([^"]*)"\s*\/?>/)
+    if (imgMatch) {
+      blocks.push(createBlock('image', [createTextSegment('')], { src: unescapeHtml(imgMatch[1]), alt: unescapeHtml(imgMatch[2]) }))
+      remaining = remaining.slice(imgMatch[0].length)
+      matched = true
+      continue
+    }
+
+    if (!matched) {
+      // Skip whitespace or unknown content
+      const nextTagIdx = remaining.indexOf('<', 1)
+      if (nextTagIdx === -1) {
+        // Remaining text as paragraph
+        const trimmed = remaining.trim()
+        if (trimmed) {
+          blocks.push(createBlock('paragraph', [createTextSegment(trimmed)]))
+        }
+        break
+      }
+      if (remaining[0] !== '<') {
+        const text = remaining.slice(0, nextTagIdx).trim()
+        if (text) {
+          blocks.push(createBlock('paragraph', [createTextSegment(text)]))
+        }
+        remaining = remaining.slice(nextTagIdx)
+      } else {
+        // Unknown tag, skip
+        const closeIdx = remaining.indexOf('>')
+        if (closeIdx !== -1) {
+          remaining = remaining.slice(closeIdx + 1)
+        } else {
+          break
+        }
+      }
+    }
+  }
+
+  return createDocument(blocks.length > 0 ? blocks : undefined)
+}
+
+// ---------------------------------------------------------------------------
+// To Plain Text
+// ---------------------------------------------------------------------------
+
+function blockToPlainText(block: Block): string {
+  const content = block.content
+    .map((seg) => {
+      if (seg.type === 'text') return seg.text
+      if (seg.type === 'mention') return seg.label
+      if (seg.type === 'emoji') return seg.unicode
+      return ''
+    })
+    .join('')
+
+  switch (block.type) {
+    case 'heading':
+      return content
+    case 'blockquote':
+      return content
+    case 'list-item': {
+      const indent = '  '.repeat(block.indent ?? 0)
+      const listType = block.meta?.listType
+      const prefix = listType === 'ordered' ? '1. ' : '- '
+      return `${indent}${prefix}${content}`
+    }
+    case 'divider':
+      return '---'
+    default:
+      return content
+  }
+}
+
+export function toPlainText(doc: Document): string {
+  return doc.blocks.map(blockToPlainText).join('\n')
+}