@peaceroad/markdown-it-figure-with-p-caption 0.15.2 → 0.16.1

package/README.md

@@ -27,6 +27,7 @@ Optionally, you can auto-number image and table caption paragraphs starting from
   - `f-img-multiple` for mixed layouts.
 - Automatic detection inspects only the first image in the paragraph. If it yields a caption, the entire figure reuses that caption while later images keep their own `alt`/`title`.
 - Paragraphs that contain only images also convert when they appear inside loose lists (leave blank lines between items), blockquotes, or description lists.
+- Caption detection intentionally skips paragraphs that are the first block inside a list item (`list_item_open` immediately before the paragraph). In practice, `- Figure. ...` followed by an image in the same item is treated as plain text unless you insert another block first.
 
 ### Table
 
@@ -50,6 +51,7 @@ Optionally, you can auto-number image and table caption paragraphs starting from
 ### Embedded content by iframe
 
 - Inline HTML `<iframe>` elements become `<figure class="f-video">` when they point to known video hosts (YouTube `youtube.com` / `youtube-nocookie.com`, Vimeo `player.vimeo.com`).
+- `<div>` wrappers are treated as iframe-type embeds only when the same HTML block contains an `<iframe ...>` tag (for example common video wrapper markup).
 - Blockquote-based social embeds (Twitter/X `twitter-tweet`, Mastodon `mastodon-embed`, Bluesky `bluesky-embed`, Instagram `instagram-media`, Tumblr `text-post-media`) are treated like iframe-type embeds when their `class` matches those providers. By default they become `<figure class="f-img">` so the caption label behaves like an image label (Labels can also use quote labels). You can override that figure class with `figureClassThatWrapsIframeTypeBlockquote` or the global `allIframeTypeFigureClassName`.
 - `p7d-markdown-it-p-captions` ships with a `Slide.` label. When you use it (for example with Speaker Deck or SlideShare iframes), the `<figure>` wrapper automatically switches to `f-slide` (or whatever you set via `figureClassThatWrapsSlides`) so slides can get their own layout. If `allIframeTypeFigureClassName` is also configured, that class takes precedence even for slides, so you get a uniform embed wrapper without touching the slide option.
 - All other iframes fall back to `<figure class="f-iframe">` unless you override the class via `allIframeTypeFigureClassName`.
@@ -57,7 +59,9 @@ Optionally, you can auto-number image and table caption paragraphs starting from
 ### label span class name
 
 - The label inside the figcaption (the `span` element used for the label) is generated by `p7d-markdown-it-p-captions`, not by this plugin. By default the class name is formed by combining `classPrefix` with the mark name, producing names such as `f-img-label`, `f-video-label`, `f-blockquote-label`, and `f-slide-label`. 
-- With `markdown-it-attrs`, any attribute block (`{ .foo #bar }`) attached to the caption paragraph is moved to the generated `<figure>` by default (`styleProcess: true`). This keeps per-figure classes/IDs on the wrapper instead of the original paragraph; disable the option only if you explicitly want the attributes to stay on the paragraph.
+- With `markdown-it-attrs`, attributes attached to image-only paragraphs (for example `![...](...) {.foo #bar}`) are forwarded to the generated `<figure>`.
+- `styleProcess` controls parsing of a trailing `{...}` block from the last text token of an image-only paragraph in this plugin's own scanner. It is a narrow fallback parser, not full `markdown-it-attrs` parity, and attributes already attached to paragraph tokens by `markdown-it-attrs` are still forwarded.
+- Attributes attached to caption paragraphs stay on the converted `<figcaption>` token after paragraph-to-figcaption conversion.
 
 ## Behavior Customization
 
@@ -67,6 +71,7 @@ Optionally, you can auto-number image and table caption paragraphs starting from
 - `figureClassThatWrapsIframeTypeBlockquote`: override the class used when blockquote-based embeds (Twitter, Mastodon, Bluesky) are wrapped.
 - `figureClassThatWrapsSlides`: override the class assigned when a caption paragraph uses the `Slide.` label.
 - `classPrefix` (default `f`) controls the CSS namespace for every figure (`f-img`, `f-table`, etc.) so you can align with existing styles.
+- Wrapper/class-prefix options are trimmed during setup; whitespace-only values fall back to the default class for that option.
 
 ### Wrapping without captions
 
@@ -85,7 +90,7 @@ Every option below is forwarded verbatim to `p7d-markdown-it-p-captions`, which
 - `wrapCaptionBody`: wrap the non-label caption text in a span element.
 - `hasNumClass`: add a class attribute to label span element if it has a label number.
 - `labelClassFollowsFigure`: mirror the resolved `<figure>` class onto the `figcaption` spans (`f-embed-label`, `f-embed-label-joint`, `f-embed-body`, etc.) when you want captions styled alongside the wrapper.
-- `figureToLabelClassMap`: extend `labelClassFollowsFigure` by mapping specific figure classes (e.g., `f-embed`) to custom caption label classes such as `caption-embed caption-social` for fine-grained control.
+- `figureToLabelClassMap`: extend `labelClassFollowsFigure` by mapping specific figure classes (e.g., `f-embed`) to custom caption label classes such as `caption-embed caption-social` for fine-grained control. When this map is provided and `labelClassFollowsFigure` is not set explicitly, figure-following mode is enabled automatically.
 - `labelPrefixMarker`: allow a leading marker before labels (string or array, e.g., `*Figure. ...`). Arrays are limited to two markers; extras are ignored.
 
 ### Automatic numbering
@@ -433,13 +438,13 @@ A paragraph.
 
 ### Styles
 
-This example uses `classPrefix: 'custom'` and leaves `styleProcess: true` so a trailing `{.notice}` block moves onto the `<figure>` wrapper.
+This example uses `classPrefix: 'custom'` and leaves `styleProcess: true` so a trailing `{.notice}` block moves onto the `<figure>` wrapper. This fallback only handles the final trailing attrs block on an image-only paragraph; for broader attrs syntax support, keep using `markdown-it-attrs`.
 
 ```
 [Markdown]
-Figure. Highlighted cat. {.notice}
+Figure. Highlighted cat.
 
-![Highlighted cat](cat.jpg)
+![Highlighted cat](cat.jpg) {.notice}
 [HTML]
 <figure class="custom-img notice">
 <figcaption><span class="custom-img-label">Figure<span class="custom-img-label-joint">.</span></span> Highlighted cat.</figcaption>
@@ -577,7 +582,7 @@ Video. Custom embed.
 </figure>
 ```
 
-Need matching caption classes too? Combine this option with `labelClassFollowsFigure` (and optionally `figureToLabelClassMap`) so the `figcaption` spans inherit the embed class you just applied (e.g., `f-embed-label`, `f-embed-label-joint`).
+Need matching caption classes too? Use `labelClassFollowsFigure` (and optionally `figureToLabelClassMap`) so the `figcaption` spans inherit the embed class you just applied (e.g., `f-embed-label`, `f-embed-label-joint`). If `figureToLabelClassMap` is provided, figure-following mode is enabled automatically unless `labelClassFollowsFigure` is set explicitly.
 
 
 ### Caption markers

package/index.js

@@ -1,7 +1,9 @@
-import { setCaptionParagraph, markReg } from 'p7d-markdown-it-p-captions'
+import {
+  setCaptionParagraph,
+  getMarkRegStateForLanguages,
+} from 'p7d-markdown-it-p-captions'
 
 const htmlRegCache = new Map()
-const cleanCaptionRegCache = new Map()
 const blueskyEmbedReg = /^<blockquote class="bluesky-embed"[^]*?>[\s\S]*?$/
 const videoIframeReg = /^<[^>]*? src="https:\/\/(?:www.youtube-nocookie.com|player.vimeo.com)\//i
 const classNameReg = /^<[^>]*? class="(twitter-tweet|instagram-media|text-post-media|bluesky-embed|mastodon-embed)"/
@@ -11,22 +13,58 @@ const idAttrReg = /^#/
 const attrParseReg = /^(.*?)="?(.*)"?$/
 const sampLangReg = /^ *(?:samp|shell|console)(?:(?= )|$)/
 const endBlockquoteScriptReg = /<\/blockquote> *<script[^>]*?><\/script>$/
-const imgCaptionMarkReg = markReg && markReg.img ? markReg.img : null
+const iframeTagReg = /<iframe(?=[\s>])/i
 const asciiLabelReg = /^[A-Za-z]/
-const trailingDigitsReg = /(\d+)\s*$/
 const CHECK_TYPE_TOKEN_MAP = {
   table_open: 'table',
   pre_open: 'pre',
   blockquote_open: 'blockquote',
 }
-const HTML_TAG_CANDIDATES = ['video', 'audio', 'iframe', 'blockquote', 'div']
-const fallbackLabelDefaults = {
-  img: { en: 'Figure', ja: '図' },
-  table: { en: 'Table', ja: '表' },
-}
+const HTML_TAG_DETECTORS = [
+  { candidate: 'video', lookupTag: 'video', hintKey: 'hasVideoHint' },
+  { candidate: 'audio', lookupTag: 'audio', hintKey: 'hasAudioHint' },
+  { candidate: 'iframe', lookupTag: 'iframe', hintKey: 'hasIframeHint' },
+  { candidate: 'blockquote', lookupTag: 'blockquote', hintKey: 'hasBlockquoteHint' },
+  {
+    candidate: 'div',
+    lookupTag: 'div',
+    hintKey: 'hasDivHint',
+    requiresIframeTag: true,
+    matchedTag: 'iframe',
+    setVideoIframe: true,
+  },
+]
+const fallbackLabelDefaults = { en: 'Figure', ja: '図' }
 
 const escapeRegExp = (value) => value.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')
-const buildClassPrefix = (value) => (value ? value + '-' : '')
+const normalizeOptionalClassName = (value) => {
+  if (value === null || value === undefined) return ''
+  const normalized = String(value).trim()
+  return normalized || ''
+}
+const buildClassPrefix = (value) => {
+  const normalized = normalizeOptionalClassName(value)
+  return normalized ? normalized + '-' : ''
+}
+const normalizeClassOptionWithFallback = (value, fallbackValue) => {
+  const normalized = normalizeOptionalClassName(value)
+  return normalized || fallbackValue
+}
+const normalizeLanguages = (value) => {
+  if (!Array.isArray(value)) return ['en', 'ja']
+  const normalized = []
+  const seen = new Set()
+  for (let i = 0; i < value.length; i++) {
+    const lang = value[i]
+    if (typeof lang !== 'string') continue
+    const trimmed = lang.trim()
+    if (!trimmed || seen.has(trimmed)) continue
+    seen.add(trimmed)
+    normalized.push(trimmed)
+  }
+  if (normalized.length === 0) return ['en', 'ja']
+  return normalized
+}
 const normalizeLabelPrefixMarkers = (value) => {
   if (typeof value === 'string') {
     return value ? [value] : []
@@ -228,31 +266,28 @@ const isSentenceBoundaryChar = (char) => {
   return char === '.' || char === '!' || char === '?' || char === '。' || char === '！' || char === '？'
 }
 
-const getAutoFallbackLabel = (text, captionType) => {
-  const type = captionType === 'table' ? 'table' : 'img'
+const getAutoFallbackLabel = (text) => {
   const lang = detectCaptionLanguage(text)
-  const defaults = fallbackLabelDefaults[type] || fallbackLabelDefaults.img
-  if (lang === 'ja') return defaults.ja || defaults.en || ''
-  return defaults.en || defaults.ja || ''
+  if (lang === 'ja') return fallbackLabelDefaults.ja || fallbackLabelDefaults.en || ''
+  return fallbackLabelDefaults.en || fallbackLabelDefaults.ja || ''
 }
 
-const getPersistedFallbackLabel = (text, captionType, fallbackState) => {
-  const type = captionType === 'table' ? 'table' : 'img'
-  if (!fallbackState) return getAutoFallbackLabel(text, type)
-  if (fallbackState[type]) return fallbackState[type]
-  const resolved = getAutoFallbackLabel(text, type)
-  fallbackState[type] = resolved
+const getPersistedFallbackLabel = (text, fallbackState) => {
+  if (!fallbackState) return getAutoFallbackLabel(text)
+  if (fallbackState.img) return fallbackState.img
+  const resolved = getAutoFallbackLabel(text)
+  fallbackState.img = resolved
   return resolved
 }
 
-const buildCaptionWithFallback = (text, fallbackOption, captionType = 'img', fallbackState) => {
+const buildCaptionWithFallback = (text, fallbackOption, fallbackState) => {
   const trimmedText = (text || '').trim()
   if (!fallbackOption) return ''
   let label = ''
   if (typeof fallbackOption === 'string') {
     label = fallbackOption.trim()
   } else if (fallbackOption === true) {
-    label = getPersistedFallbackLabel(trimmedText, captionType, fallbackState)
+    label = getPersistedFallbackLabel(trimmedText, fallbackState)
   }
   if (!label) return trimmedText
   const isAsciiLabel = asciiLabelReg.test(label)
@@ -331,6 +366,28 @@ const updateInlineTokenContent = (inlineToken, originalText, newText) => {
     inlineToken.content.slice(index + originalText.length)
 }
 
+const parseTrailingPositiveInteger = (text) => {
+  if (typeof text !== 'string' || text.length === 0) return null
+  let end = text.length - 1
+  while (end >= 0 && text.charCodeAt(end) === 0x20) end--
+  if (end < 0) return null
+  const lastCode = text.charCodeAt(end)
+  if (lastCode < 0x30 || lastCode > 0x39) return null
+  let start = end
+  while (start >= 0) {
+    const code = text.charCodeAt(start)
+    if (code < 0x30 || code > 0x39) break
+    start--
+  }
+  let value = 0
+  let digitBase = 1
+  for (let i = end; i > start; i--) {
+    value += (text.charCodeAt(i) - 0x30) * digitBase
+    digitBase *= 10
+  }
+  return value
+}
+
 const ensureAutoFigureNumbering = (tokens, range, caption, figureNumberState, opt) => {
   const captionType = caption.name === 'img' ? 'img' : (caption.name === 'table' ? 'table' : '')
   if (!captionType) return
@@ -345,10 +402,9 @@ const ensureAutoFigureNumbering = (tokens, range, caption, figureNumberState, op
   if (end >= 0) {
     const code = originalText.charCodeAt(end)
     if (code >= 0x30 && code <= 0x39) {
-      const existingMatch = originalText.match(trailingDigitsReg)
-      if (existingMatch && existingMatch[1]) {
-        const explicitValue = parseInt(existingMatch[1], 10)
-        if (!Number.isNaN(explicitValue) && explicitValue > (figureNumberState[captionType] || 0)) {
+      const explicitValue = parseTrailingPositiveInteger(originalText)
+      if (explicitValue !== null) {
+        if (explicitValue > (figureNumberState[captionType] || 0)) {
           figureNumberState[captionType] = explicitValue
         }
         return
@@ -364,27 +420,27 @@ const ensureAutoFigureNumbering = (tokens, range, caption, figureNumberState, op
   updateInlineTokenContent(inlineToken, originalText, newLabelText)
 }
 
+const matchAutoCaptionText = (text, reg) => {
+  if (!text || !reg) return ''
+  const trimmed = text.trim()
+  if (trimmed && reg.test(trimmed)) return trimmed
+  return ''
+}
+
 const getAutoCaptionFromImage = (imageToken, opt, fallbackLabelState) => {
+  const imgCaptionMarkReg = opt && opt.imgCaptionMarkReg ? opt.imgCaptionMarkReg : null
   if (!opt.autoCaptionDetection) return ''
   if (!imgCaptionMarkReg && !opt.autoAltCaption && !opt.autoTitleCaption) return ''
-  const tryMatch = (text) => {
-    if (!text) return ''
-    const trimmed = text.trim()
-    if (trimmed && imgCaptionMarkReg && imgCaptionMarkReg.test(trimmed)) {
-      return trimmed
-    }
-    return ''
-  }
 
   const altText = getImageAltText(imageToken)
-  let caption = tryMatch(altText)
+  let caption = matchAutoCaptionText(altText, imgCaptionMarkReg)
   if (caption) {
     clearImageAltAttr(imageToken)
     return caption
   }
   if (!caption && opt.autoAltCaption) {
     const altForFallback = altText || ''
-    caption = buildCaptionWithFallback(altForFallback, opt.autoAltCaption, 'img', fallbackLabelState)
+    caption = buildCaptionWithFallback(altForFallback, opt.autoAltCaption, fallbackLabelState)
     if (imageToken) {
       clearImageAltAttr(imageToken)
     }
@@ -392,14 +448,14 @@ const getAutoCaptionFromImage = (imageToken, opt, fallbackLabelState) => {
   if (caption) return caption
 
   const titleText = getImageTitleText(imageToken)
-  caption = tryMatch(titleText)
+  caption = matchAutoCaptionText(titleText, imgCaptionMarkReg)
   if (caption) {
     clearImageTitleAttr(imageToken)
     return caption
   }
   if (!caption && opt.autoTitleCaption) {
     const titleForFallback = titleText || ''
-    caption = buildCaptionWithFallback(titleForFallback, opt.autoTitleCaption, 'img', fallbackLabelState)
+    caption = buildCaptionWithFallback(titleForFallback, opt.autoTitleCaption, fallbackLabelState)
     if (imageToken) {
       clearImageTitleAttr(imageToken)
     }
@@ -408,21 +464,118 @@ const getAutoCaptionFromImage = (imageToken, opt, fallbackLabelState) => {
 }
 
 const getHtmlReg = (tag) => {
-  if (htmlRegCache.has(tag)) return htmlRegCache.get(tag)
+  const cached = htmlRegCache.get(tag)
+  if (cached) return cached
   const regexStr = `^<${tag} ?[^>]*?>[\\s\\S]*?<\\/${tag}>(\\n| *?)(<script [^>]*?>(?:<\\/script>)?)? *(\\n|$)`
   const reg = new RegExp(regexStr)
   htmlRegCache.set(tag, reg)
   return reg
 }
 
-const checkPrevCaption = (tokens, n, caption, fNum, sp, opt, captionState) => {
+const getHtmlDetectionHints = (content) => {
+  const hasBlueskyHint = content.indexOf('bluesky-embed') !== -1
+  const hasVideoHint = content.indexOf('<video') !== -1
+  const hasAudioHint = content.indexOf('<audio') !== -1
+  const hasIframeHint = content.indexOf('<iframe') !== -1
+  const hasBlockquoteHint = content.indexOf('<blockquote') !== -1
+  const hasDivHint = content.indexOf('<div') !== -1
+  return {
+    hasBlueskyHint,
+    hasVideoHint,
+    hasAudioHint,
+    hasIframeHint,
+    hasBlockquoteHint,
+    hasDivHint,
+    hasIframeTag: hasIframeHint || (hasDivHint && iframeTagReg.test(content)),
+    hasBlueskyEmbed: hasBlueskyHint && blueskyEmbedReg.test(content),
+  }
+}
+
+const hasAnyHtmlDetectionHint = (hints) => {
+  return !!(
+    hints.hasBlueskyHint ||
+    hints.hasVideoHint ||
+    hints.hasAudioHint ||
+    hints.hasIframeHint ||
+    hints.hasBlockquoteHint ||
+    hints.hasDivHint
+  )
+}
+
+const appendHtmlBlockNewlineIfNeeded = (token, hasTag) => {
+  if ((hasTag[2] && hasTag[3] !== '\n') || (hasTag[1] !== '\n' && hasTag[2] === undefined)) {
+    token.content += '\n'
+  }
+}
+
+const consumeBlockquoteEmbedScript = (tokens, token, startIndex) => {
+  let addedCont = ''
+  let j = startIndex + 1
+  while (j < tokens.length) {
+    const nextToken = tokens[j]
+    if (nextToken.type === 'inline' && endBlockquoteScriptReg.test(nextToken.content)) {
+      addedCont += nextToken.content + '\n'
+      if (tokens[j + 1] && tokens[j + 1].type === 'paragraph_close') {
+        tokens.splice(j + 1, 1)
+      }
+      nextToken.content = ''
+      if (nextToken.children) {
+        for (let k = 0; k < nextToken.children.length; k++) {
+          nextToken.children[k].content = ''
+        }
+      }
+      break
+    }
+    if (nextToken.type === 'paragraph_open') {
+      addedCont += '\n'
+      tokens.splice(j, 1)
+      continue
+    }
+    j++
+  }
+  token.content += addedCont
+}
+
+const detectHtmlTagCandidate = (tokens, token, startIndex, detector, hints, sp) => {
+  if (detector.requiresIframeTag && !hints.hasIframeTag) return ''
+  const hasTagHint = !!(detector.hintKey && hints[detector.hintKey])
+  const allowBlueskyFallback = detector.candidate === 'blockquote' && hints.hasBlueskyEmbed
+  if (!hasTagHint && !allowBlueskyFallback) return ''
+  const hasTag = hasTagHint ? token.content.match(getHtmlReg(detector.lookupTag)) : null
+  const isBlueskyBlockquote = detector.candidate === 'blockquote' && !hasTag && hints.hasBlueskyEmbed
+  if (!hasTag && !isBlueskyBlockquote) return ''
+  if (hasTag) {
+    appendHtmlBlockNewlineIfNeeded(token, hasTag)
+    if (detector.setVideoIframe) {
+      sp.isVideoIframe = true
+    }
+    return detector.matchedTag || detector.candidate
+  }
+  consumeBlockquoteEmbedScript(tokens, token, startIndex)
+  return 'blockquote'
+}
+
+const isIframeTypeEmbedBlockquote = (content) => {
+  return content.indexOf('class="') !== -1 && classNameReg.test(content)
+}
+
+const resolveHtmlWrapWithoutCaption = (matchedTag, sp, opt) => {
+  const htmlWrapWithoutCaption = opt.htmlWrapWithoutCaption
+  if (!htmlWrapWithoutCaption) return false
+  if (matchedTag === 'blockquote') {
+    return !!(sp.isIframeTypeBlockquote && htmlWrapWithoutCaption.iframeTypeBlockquote)
+  }
+  return !!htmlWrapWithoutCaption[matchedTag]
+}
+
+const checkPrevCaption = (tokens, n, caption, sp, opt, captionState) => {
   if(n < 3) return caption
   const captionStartToken = tokens[n-3]
   const captionInlineToken = tokens[n-2]
   const captionEndToken = tokens[n-1]
   if (captionStartToken === undefined || captionEndToken === undefined) return
   if (captionStartToken.type !== 'paragraph_open' || captionEndToken.type !== 'paragraph_close') return
-  setCaptionParagraph(n-3, captionState, caption, fNum, sp, opt)
+  setCaptionParagraph(n-3, captionState, caption, null, sp, opt)
   const captionName = sp && sp.captionDecision ? sp.captionDecision.mark : ''
   if(!captionName) {
     if (opt.labelPrefixMarkerWithoutLabelPrevReg) {
@@ -439,14 +592,14 @@ const checkPrevCaption = (tokens, n, caption, fNum, sp, opt, captionState) => {
   return
 }
 
-const checkNextCaption = (tokens, en, caption, fNum, sp, opt, captionState) => {
+const checkNextCaption = (tokens, en, caption, sp, opt, captionState) => {
   if (en + 2 > tokens.length) return
   const captionStartToken = tokens[en+1]
   const captionInlineToken = tokens[en+2]
   const captionEndToken = tokens[en+3]
   if (captionStartToken === undefined || captionEndToken === undefined) return
   if (captionStartToken.type !== 'paragraph_open' || captionEndToken.type !== 'paragraph_close') return
-  setCaptionParagraph(en+1, captionState, caption, fNum, sp, opt)
+  setCaptionParagraph(en+1, captionState, caption, null, sp, opt)
   const captionName = sp && sp.captionDecision ? sp.captionDecision.mark : ''
   if(!captionName) {
     if (opt.labelPrefixMarkerWithoutLabelNextReg) {
@@ -468,10 +621,13 @@ const cleanCaptionTokenAttrs = (token, captionName, opt) => {
   const prefix = opt.captionClassPrefix || ''
   const targetClass = prefix + captionName
   if (!targetClass) return
-  let reg = cleanCaptionRegCache.get(targetClass)
+  const cleanCaptionRegCache = opt.cleanCaptionRegCache
+  let reg = cleanCaptionRegCache && cleanCaptionRegCache.get(targetClass)
   if (!reg) {
     reg = new RegExp('(?:^|\\s)' + escapeRegExp(targetClass) + '(?=\\s|$)', 'g')
-    cleanCaptionRegCache.set(targetClass, reg)
+    if (cleanCaptionRegCache) {
+      cleanCaptionRegCache.set(targetClass, reg)
+    }
   }
   for (let i = token.attrs.length - 1; i >= 0; i--) {
     if (token.attrs[i][0] === 'class') {
@@ -568,28 +724,31 @@ const wrapWithFigure = (tokens, range, checkTokenTagName, caption, replaceInstea
   if (rangeEndMap) {
     figureEndToken.map = [rangeEndMap[0], rangeEndMap[1]]
   }
-  const breakToken = new TokenConstructor('text', '', 0)
-  breakToken.content = '\n'
-  if (opt.styleProcess && caption.isNext && sp.attrs.length > 0) {
-    for (let i = 0; i < sp.attrs.length; i++) {
-      const attr = sp.attrs[i]
-      figureStartToken.attrJoin(attr[0], attr[1])
-    }
+  const createBreakToken = () => {
+    const breakToken = new TokenConstructor('text', '', 0)
+    breakToken.content = '\n'
+    return breakToken
   }
-  // For vsce
-  if (caption.name === 'img' && tokens[n].attrs) {
-    for (let i = 0; i < tokens[n].attrs.length; i++) {
-      const attr = tokens[n].attrs[i]
-      figureStartToken.attrJoin(attr[0], attr[1])
+  if (caption.name === 'img') {
+    const joinAttrs = (attrs) => {
+      if (!attrs || attrs.length === 0) return
+      for (let i = 0; i < attrs.length; i++) {
+        const attr = attrs[i]
+        figureStartToken.attrJoin(attr[0], attr[1])
+      }
     }
+    // `styleProcess` should keep working even when markdown-it-attrs is absent.
+    if (opt.styleProcess) joinAttrs(sp.attrs)
+    // Forward attrs already materialized by markdown-it-attrs on the image paragraph.
+    joinAttrs(tokens[n].attrs)
   }
   if (replaceInsteadOfWrap) {
-    tokens.splice(en, 1, breakToken, figureEndToken, breakToken)
-    tokens.splice(n, 1, figureStartToken, breakToken)
+    tokens.splice(en, 1, createBreakToken(), figureEndToken, createBreakToken())
+    tokens.splice(n, 1, figureStartToken, createBreakToken())
     en = en + 2
   } else {
-    tokens.splice(en+1, 0, figureEndToken, breakToken)
-    tokens.splice(n, 0, figureStartToken, breakToken)
+    tokens.splice(en+1, 0, figureEndToken, createBreakToken())
+    tokens.splice(n, 0, figureStartToken, createBreakToken())
     en = en + 3
   }
   range.start = n
@@ -597,10 +756,10 @@ const wrapWithFigure = (tokens, range, checkTokenTagName, caption, replaceInstea
   return
 }
 
-const checkCaption = (tokens, n, en, caption, fNum, sp, opt, captionState) => {
-  checkPrevCaption(tokens, n, caption, fNum, sp, opt, captionState)
+const checkCaption = (tokens, n, en, caption, sp, opt, captionState) => {
+  checkPrevCaption(tokens, n, caption, sp, opt, captionState)
   if (caption.isPrev) return
-  checkNextCaption(tokens, en, caption, fNum, sp, opt, captionState)
+  checkNextCaption(tokens, en, caption, sp, opt, captionState)
   return
 }
 
@@ -695,60 +854,16 @@ const detectFenceToken = (token, n, caption) => {
 
 const detectHtmlBlockToken = (tokens, token, n, caption, sp, opt) => {
   if (!token || token.type !== 'html_block') return null
-  const content = token.content
-  const hasBlueskyHint = content.indexOf('bluesky-embed') !== -1
-  const hasBlueskyEmbed = hasBlueskyHint && blueskyEmbedReg.test(content)
+  const hints = getHtmlDetectionHints(token.content)
+  if (!hasAnyHtmlDetectionHint(hints)) return null
   let matchedTag = ''
-  for (let i = 0; i < HTML_TAG_CANDIDATES.length; i++) {
-    const candidate = HTML_TAG_CANDIDATES[i]
-    const treatDivAsIframe = candidate === 'div'
-    const lookupTag = treatDivAsIframe ? 'div' : candidate
-    const hasTagHint = content.indexOf('<' + lookupTag) !== -1
-    if (!hasTagHint && !(candidate === 'blockquote' && hasBlueskyEmbed)) continue
-    const hasTag = hasTagHint ? content.match(getHtmlReg(lookupTag)) : null
-    const isBlueskyBlockquote = !hasTag && hasBlueskyEmbed && candidate === 'blockquote'
-    if (!(hasTag || isBlueskyBlockquote)) continue
-    if (hasTag) {
-      if ((hasTag[2] && hasTag[3] !== '\n') || (hasTag[1] !== '\n' && hasTag[2] === undefined)) {
-        token.content += '\n'
-      }
-      matchedTag = treatDivAsIframe ? 'iframe' : candidate
-      if (treatDivAsIframe) {
-        sp.isVideoIframe = true
-      }
-    } else {
-      let addedCont = ''
-      let j = n + 1
-      while (j < tokens.length) {
-        const nextToken = tokens[j]
-        if (nextToken.type === 'inline' && endBlockquoteScriptReg.test(nextToken.content)) {
-          addedCont += nextToken.content + '\n'
-          if (tokens[j + 1] && tokens[j + 1].type === 'paragraph_close') {
-            tokens.splice(j + 1, 1)
-          }
-          nextToken.content = ''
-          if (nextToken.children) {
-            for (let k = 0; k < nextToken.children.length; k++) {
-              nextToken.children[k].content = ''
-            }
-          }
-          break
-        }
-        if (nextToken.type === 'paragraph_open') {
-          addedCont += '\n'
-          tokens.splice(j, 1)
-          continue
-        }
-        j++
-      }
-      token.content += addedCont
-      matchedTag = 'blockquote'
-    }
-    break
+  for (let i = 0; i < HTML_TAG_DETECTORS.length; i++) {
+    matchedTag = detectHtmlTagCandidate(tokens, token, n, HTML_TAG_DETECTORS[i], hints, sp)
+    if (matchedTag) break
   }
   if (!matchedTag) return null
   if (matchedTag === 'blockquote') {
-    if (classNameReg.test(token.content)) {
+    if (isIframeTypeEmbedBlockquote(token.content)) {
       sp.isIframeTypeBlockquote = true
     } else {
       return null
@@ -758,16 +873,7 @@ const detectHtmlBlockToken = (tokens, token, n, caption, sp, opt) => {
     sp.isVideoIframe = true
   }
   caption.name = matchedTag
-  let wrapWithoutCaption = false
-  if (matchedTag === 'iframe' && opt.iframeWithoutCaption) {
-    wrapWithoutCaption = true
-  } else if (matchedTag === 'video' && opt.videoWithoutCaption) {
-    wrapWithoutCaption = true
-  } else if (matchedTag === 'audio' && opt.audioWithoutCaption) {
-    wrapWithoutCaption = true
-  } else if (matchedTag === 'blockquote' && sp.isIframeTypeBlockquote && opt.iframeTypeBlockquoteWithoutCaption) {
-    wrapWithoutCaption = true
-  }
+  const wrapWithoutCaption = resolveHtmlWrapWithoutCaption(matchedTag, sp, opt)
   return {
     type: 'html',
     tagName: matchedTag,
@@ -778,23 +884,54 @@ const detectHtmlBlockToken = (tokens, token, n, caption, sp, opt) => {
   }
 }
 
-const detectImageParagraph = (tokens, token, nextToken, n, caption, sp, opt) => {
-  if (!token || token.type !== 'paragraph_open') return null
-  if (!nextToken || nextToken.type !== 'inline' || !nextToken.children || nextToken.children.length === 0) return null
-  if (nextToken.children[0].type !== 'image') return null
+const hasLeadingImageChild = (token) => {
+  return !!(token &&
+    token.type === 'inline' &&
+    token.children &&
+    token.children.length > 0 &&
+    token.children[0] &&
+    token.children[0].type === 'image')
+}
 
+const detectImageParagraph = (nextToken, n, caption, sp, opt) => {
+  const multipleImagesEnabled = !!opt.multipleImages
+  const styleProcessEnabled = !!opt.styleProcess
+  const allowSingleImageWithoutCaption = !!opt.oneImageWithoutCaption
+  const children = nextToken.children
+  const imageToken = children[0]
+  const childrenLength = children.length
   let imageNum = 1
   let isMultipleImagesHorizontal = true
   let isMultipleImagesVertical = true
   let isValid = true
   caption.name = 'img'
-  const children = nextToken.children
-  const childrenLength = children.length
+  if (childrenLength === 1) {
+    return {
+      type: 'image',
+      tagName: 'img',
+      en: n + 2,
+      replaceInsteadOfWrap: true,
+      wrapWithoutCaption: allowSingleImageWithoutCaption,
+      canWrap: true,
+      imageToken,
+    }
+  }
+  if (!multipleImagesEnabled && childrenLength > 2) {
+    return {
+      type: 'image',
+      tagName: 'img',
+      en: n + 2,
+      replaceInsteadOfWrap: true,
+      wrapWithoutCaption: false,
+      canWrap: false,
+      imageToken,
+    }
+  }
   for (let childIndex = 1; childIndex < childrenLength; childIndex++) {
     const child = children[childIndex]
     if (childIndex === childrenLength - 1 && child.type === 'text') {
       const rawContent = child.content
-      if (opt.styleProcess && rawContent && rawContent.indexOf('{') !== -1) {
+      if (styleProcessEnabled && rawContent && rawContent.indexOf('{') !== -1 && rawContent.indexOf('}') !== -1) {
         const imageAttrs = rawContent.match(imageAttrsReg)
         if (imageAttrs) {
           const parsedAttrs = parseImageAttrs(imageAttrs[1])
@@ -802,6 +939,7 @@ const detectImageParagraph = (tokens, token, nextToken, n, caption, sp, opt) =>
             for (let i = 0; i < parsedAttrs.length; i++) {
               sp.attrs.push(parsedAttrs[i])
             }
+            child.content = ''
           }
           break
         }
@@ -812,7 +950,7 @@ const detectImageParagraph = (tokens, token, nextToken, n, caption, sp, opt) =>
       break
     }
 
-    if (!opt.multipleImages) {
+    if (!multipleImagesEnabled) {
       isValid = false
       break
     }
@@ -831,7 +969,7 @@ const detectImageParagraph = (tokens, token, nextToken, n, caption, sp, opt) =>
     isValid = false
     break
   }
-  if (isValid && imageNum > 1 && opt.multipleImages) {
+  if (isValid && imageNum > 1 && multipleImagesEnabled) {
     if (isMultipleImagesHorizontal) {
       caption.nameSuffix = '-horizontal'
     } else if (isMultipleImagesVertical) {
@@ -854,18 +992,13 @@ const detectImageParagraph = (tokens, token, nextToken, n, caption, sp, opt) =>
     tagName,
     en,
     replaceInsteadOfWrap: true,
-    wrapWithoutCaption: isValid && !!opt.oneImageWithoutCaption,
+    wrapWithoutCaption: isValid && allowSingleImageWithoutCaption,
     canWrap: isValid,
-    imageToken: children[0]
+    imageToken,
   }
 }
 
 const figureWithCaption = (state, opt) => {
-  let fNum = {
-    img: 0,
-    table: 0,
-  }
-
   const figureNumberState = {
     img: 0,
     table: 0,
@@ -873,14 +1006,13 @@ const figureWithCaption = (state, opt) => {
 
   const fallbackLabelState = {
     img: null,
-    table: null,
   }
 
   const captionState = { tokens: state.tokens, Token: state.Token }
-  figureWithCaptionCore(state.tokens, opt, fNum, figureNumberState, fallbackLabelState, state.Token, captionState, null, 0)
+  figureWithCaptionCore(state.tokens, opt, figureNumberState, fallbackLabelState, state.Token, captionState, null, 0)
 }
 
-const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLabelState, TokenConstructor, captionState, parentType = null, startIndex = 0) => {
+const figureWithCaptionCore = (tokens, opt, figureNumberState, fallbackLabelState, TokenConstructor, captionState, parentType = null, startIndex = 0) => {
   const rRange = { start: startIndex, end: startIndex }
   const rCaption = {
     name: '', nameSuffix: '', isPrev: false, isNext: false
@@ -898,7 +1030,7 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
     const containerType = getNestedContainerType(token)
 
     if (containerType && containerType !== 'blockquote') {
-      const closeIndex = figureWithCaptionCore(tokens, opt, fNum, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, n + 1)
+      const closeIndex = figureWithCaptionCore(tokens, opt, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, n + 1)
       n = (typeof closeIndex === 'number' ? closeIndex : n) + 1
       continue
     }
@@ -912,11 +1044,13 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
     const tokenType = token.type
     const blockType = CHECK_TYPE_TOKEN_MAP[tokenType]
     if (tokenType === 'paragraph_open') {
-      resetRangeState(rRange, n)
-      resetCaptionState(rCaption)
-      resetSpecialState(rSp)
       const nextToken = tokens[n + 1]
-      detection = detectImageParagraph(tokens, token, nextToken, n, rCaption, rSp, opt)
+      if (hasLeadingImageChild(nextToken)) {
+        resetRangeState(rRange, n)
+        resetCaptionState(rCaption)
+        resetSpecialState(rSp)
+        detection = detectImageParagraph(nextToken, n, rCaption, rSp, opt)
+      }
     } else if (tokenType === 'html_block') {
       resetRangeState(rRange, n)
       resetCaptionState(rCaption)
@@ -936,7 +1070,7 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
 
     if (!detection) {
       if (containerType === 'blockquote') {
-        const closeIndex = figureWithCaptionCore(tokens, opt, fNum, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, n + 1)
+        const closeIndex = figureWithCaptionCore(tokens, opt, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, n + 1)
         n = (typeof closeIndex === 'number' ? closeIndex : n) + 1
       } else {
         n++
@@ -947,7 +1081,7 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
     rRange.end = detection.en
 
     rSp.figureClassName = resolveFigureClassName(detection.tagName, rSp, opt)
-    checkCaption(tokens, rRange.start, rRange.end, rCaption, fNum, rSp, opt, captionState)
+    checkCaption(tokens, rRange.start, rRange.end, rCaption, rSp, opt, captionState)
     applyCaptionDrivenFigureClass(rCaption, rSp, opt)
 
     let hasCaption = rCaption.isPrev || rCaption.isNext
@@ -962,7 +1096,7 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
     if (detection.canWrap === false) {
       let nextIndex = rRange.end + 1
       if (containerType === 'blockquote') {
-        const closeIndex = figureWithCaptionCore(tokens, opt, fNum, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, rRange.start + 1)
+        const closeIndex = figureWithCaptionCore(tokens, opt, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, rRange.start + 1)
         nextIndex = Math.max(nextIndex, (typeof closeIndex === 'number' ? closeIndex : rRange.end) + 1)
       }
       n = nextIndex
@@ -989,7 +1123,7 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
         rRange.start += insertedLength
         rRange.end += insertedLength
         n += insertedLength
-        checkCaption(tokens, rRange.start, rRange.end, rCaption, fNum, rSp, opt, captionState)
+        checkCaption(tokens, rRange.start, rRange.end, rCaption, rSp, opt, captionState)
         applyCaptionDrivenFigureClass(rCaption, rSp, opt)
       }
       ensureAutoFigureNumbering(tokens, rRange, rCaption, figureNumberState, opt)
@@ -1017,7 +1151,7 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
     }
 
     if (containerType === 'blockquote') {
-      const closeIndex = figureWithCaptionCore(tokens, opt, fNum, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, rRange.start + 1)
+      const closeIndex = figureWithCaptionCore(tokens, opt, figureNumberState, fallbackLabelState, TokenConstructor, captionState, containerType, rRange.start + 1)
       const fallbackIndex = rCaption.name ? rRange.end : n
       nextIndex = Math.max(nextIndex, (typeof closeIndex === 'number' ? closeIndex : fallbackIndex) + 1)
     }
@@ -1029,11 +1163,14 @@ const figureWithCaptionCore = (tokens, opt, fNum, figureNumberState, fallbackLab
 
 const mditFigureWithPCaption = (md, option) => {
   let opt = {
+    // Caption languages delegated to p-captions.
+    languages: ['en', 'ja'],
+
     // --- figure-wrapper behavior ---
     classPrefix: 'f',
     figureClassThatWrapsIframeTypeBlockquote: null,
     figureClassThatWrapsSlides: null,
-    styleProcess : true,
+    styleProcess: true,
     oneImageWithoutCaption: false,
     iframeWithoutCaption: false,
     videoWithoutCaption: false,
@@ -1047,7 +1184,7 @@ const mditFigureWithPCaption = (md, option) => {
     // Applies only to the first image within an image-only paragraph (even when multipleImages is true).
     // Priority: caption paragraphs (before/after) > alt text > title attribute; auto detection only runs when no paragraph caption exists.
     autoCaptionDetection: true,
-    autoAltCaption: false, // allow alt text (when matching markReg.img or fallback) to build captions automatically
+    autoAltCaption: false, // allow alt text (when matching markReg.img) to build captions automatically
     autoTitleCaption: false, // same as above but reads from the title attribute when alt isn't usable
 
     // --- label prefix marker helpers ---
@@ -1069,11 +1206,30 @@ const mditFigureWithPCaption = (md, option) => {
     removeUnnumberedLabelExceptMarks: [],
     removeMarkNameInCaptionClass: false,
     wrapCaptionBody: false,
+    labelClassFollowsFigure: false,
+    figureToLabelClassMap: null,
   }
   const hasExplicitAutoLabelNumberSets = option && Object.prototype.hasOwnProperty.call(option, 'autoLabelNumberSets')
   const hasExplicitFigureClassThatWrapsIframeTypeBlockquote = option && Object.prototype.hasOwnProperty.call(option, 'figureClassThatWrapsIframeTypeBlockquote')
   const hasExplicitFigureClassThatWrapsSlides = option && Object.prototype.hasOwnProperty.call(option, 'figureClassThatWrapsSlides')
+  const hasExplicitLabelClassFollowsFigure = option && Object.prototype.hasOwnProperty.call(option, 'labelClassFollowsFigure')
   if (option) Object.assign(opt, option)
+  if (!hasExplicitLabelClassFollowsFigure && opt.figureToLabelClassMap) {
+    opt.labelClassFollowsFigure = true
+  }
+  opt.classPrefix = normalizeOptionalClassName(opt.classPrefix)
+  opt.allIframeTypeFigureClassName = normalizeOptionalClassName(opt.allIframeTypeFigureClassName)
+  opt.languages = normalizeLanguages(opt.languages)
+  opt.markRegState = getMarkRegStateForLanguages(opt.languages)
+  opt.imgCaptionMarkReg = opt.markRegState && opt.markRegState.markReg
+    ? opt.markRegState.markReg.img
+    : null
+  opt.htmlWrapWithoutCaption = {
+    iframe: !!opt.iframeWithoutCaption,
+    video: !!opt.videoWithoutCaption,
+    audio: !!opt.audioWithoutCaption,
+    iframeTypeBlockquote: !!opt.iframeTypeBlockquoteWithoutCaption,
+  }
   // Normalize option shorthands now so downstream logic works with a consistent { img, table } shape.
   opt.autoLabelNumberSets = normalizeAutoLabelNumberSets(opt.autoLabelNumberSets)
   if (opt.autoLabelNumber && !hasExplicitAutoLabelNumberSets) {
@@ -1083,16 +1239,29 @@ const mditFigureWithPCaption = (md, option) => {
   const classPrefix = buildClassPrefix(opt.classPrefix)
   opt.figureClassPrefix = classPrefix
   opt.captionClassPrefix = classPrefix
+  const defaultIframeTypeBlockquoteClass = classPrefix + 'img'
+  const defaultSlideFigureClass = classPrefix + 'slide'
   if (!hasExplicitFigureClassThatWrapsIframeTypeBlockquote) {
-    opt.figureClassThatWrapsIframeTypeBlockquote = classPrefix + 'img'
+    opt.figureClassThatWrapsIframeTypeBlockquote = defaultIframeTypeBlockquoteClass
+  } else {
+    opt.figureClassThatWrapsIframeTypeBlockquote = normalizeClassOptionWithFallback(
+      opt.figureClassThatWrapsIframeTypeBlockquote,
+      defaultIframeTypeBlockquoteClass,
+    )
   }
   if (!hasExplicitFigureClassThatWrapsSlides) {
-    opt.figureClassThatWrapsSlides = classPrefix + 'slide'
+    opt.figureClassThatWrapsSlides = defaultSlideFigureClass
+  } else {
+    opt.figureClassThatWrapsSlides = normalizeClassOptionWithFallback(
+      opt.figureClassThatWrapsSlides,
+      defaultSlideFigureClass,
+    )
   }
   // Precompute label-class permutations so numbering lookup doesn't rebuild arrays per caption.
   opt.labelClassLookup = buildLabelClassLookup(opt)
   const markerList = normalizeLabelPrefixMarkers(opt.labelPrefixMarker)
   opt.labelPrefixMarkerReg = buildLabelPrefixMarkerRegFromList(markerList)
+  opt.cleanCaptionRegCache = new Map()
   if (opt.allowLabelPrefixMarkerWithoutLabel === true) {
     const markerPair = resolveLabelPrefixMarkerPair(markerList)
     opt.labelPrefixMarkerWithoutLabelPrevReg = buildLabelPrefixMarkerRegFromList(markerPair.prev)

package/package.json

@@ -1,11 +1,14 @@
 {
   "name": "@peaceroad/markdown-it-figure-with-p-caption",
-  "version": "0.15.2",
+  "version": "0.16.1",
   "description": "A markdown-it plugin. For a paragraph with only one image, a table or code block or blockquote, and by writing a caption paragraph immediately before or after, they are converted into the figure element with the figcaption element.",
   "main": "index.js",
   "type": "module",
   "scripts": {
-    "test": "node test/test.js"
+    "test": "node test/test.js",
+    "test:p-captions": "node test/test-p-captions.js",
+    "test:all": "npm run test && npm run test:p-captions",
+    "perf": "node test/performance/benchmark.js"
   },
   "repository": {
     "type": "git",
@@ -17,16 +20,16 @@
     "url": "https://github.com/peaceroad/p7d-markdown-it-figure-with-p-caption/issues"
   },
   "devDependencies": {
-    "@peaceroad/markdown-it-cjk-breaks-mod": "^0.1.4",
-    "@peaceroad/markdown-it-renderer-fence": "^0.3.1",
-    "@peaceroad/markdown-it-renderer-image": "^0.7.0",
-    "@peaceroad/markdown-it-strong-ja": "^0.6.2",
+    "@peaceroad/markdown-it-cjk-breaks-mod": "^0.1.10",
+    "@peaceroad/markdown-it-renderer-fence": "^0.6.1",
+    "@peaceroad/markdown-it-renderer-image": "^0.13.0",
+    "@peaceroad/markdown-it-strong-ja": "^0.9.0",
     "highlight.js": "^11.11.1",
     "markdown-it": "^14.1.0",
     "markdown-it-attrs": "^4.3.1"
   },
   "dependencies": {
-    "p7d-markdown-it-p-captions": "^0.20.1"
+    "p7d-markdown-it-p-captions": "^0.21.0"
   },
   "files": [
     "index.js",