From 4442847b37e6532742e246fedb43fd3d60384545 Mon Sep 17 00:00:00 2001
From: Jacky Zhao <j.zhao2k19@gmail.com>
Date: Thu, 14 Dec 2023 00:07:44 +0000
Subject: [PATCH] fix: internal link selector specificity

---
 quartz/plugins/transformers/ofm.ts |  199 +++++++++++++++++++++++++++++++++++++------------
 1 files changed, 150 insertions(+), 49 deletions(-)

diff --git a/quartz/plugins/transformers/ofm.ts b/quartz/plugins/transformers/ofm.ts
index a9beda4..4c6a6db 100644
--- a/quartz/plugins/transformers/ofm.ts
+++ b/quartz/plugins/transformers/ofm.ts
@@ -1,18 +1,20 @@
 import { PluggableList } from "unified"
 import { QuartzTransformerPlugin } from "../types"
-import { Root, HTML, BlockContent, DefinitionContent, Code } from "mdast"
+import { Root, HTML, BlockContent, DefinitionContent, Code, Paragraph } from "mdast"
+import { Element, Literal, Root as HtmlRoot } from "hast"
 import { Replace, findAndReplace as mdastFindReplace } from "mdast-util-find-and-replace"
 import { slug as slugAnchor } from "github-slugger"
 import rehypeRaw from "rehype-raw"
 import { visit } from "unist-util-visit"
 import path from "path"
-import { JSResource } from "../../resources"
+import { JSResource } from "../../util/resources"
 // @ts-ignore
 import calloutScript from "../../components/scripts/callout.inline.ts"
-import { FilePath, canonicalizeServer, pathToRoot, slugTag, slugifyFilePath } from "../../path"
+import { FilePath, pathToRoot, slugTag, slugifyFilePath } from "../../util/path"
 import { toHast } from "mdast-util-to-hast"
 import { toHtml } from "hast-util-to-html"
 import { PhrasingContent } from "mdast-util-find-and-replace/lib"
+import { capitalize } from "../../util/lang"
 
 export interface Options {
   comments: boolean
@@ -21,6 +23,7 @@
   callouts: boolean
   mermaid: boolean
   parseTags: boolean
+  parseBlockReferences: boolean
   enableInHtmlEmbed: boolean
 }
 
@@ -31,6 +34,7 @@
   callouts: true,
   mermaid: true,
   parseTags: true,
+  parseBlockReferences: true,
   enableInHtmlEmbed: false,
 }
 
@@ -69,6 +73,8 @@
 const calloutMapping: Record<string, keyof typeof callouts> = {
   note: "note",
   abstract: "abstract",
+  summary: "abstract",
+  tldr: "abstract",
   info: "info",
   todo: "todo",
   tip: "tip",
@@ -96,11 +102,7 @@
 
 function canonicalizeCallout(calloutName: string): keyof typeof callouts {
   let callout = calloutName.toLowerCase() as keyof typeof calloutMapping
-  return calloutMapping[callout]
-}
-
-const capitalize = (s: string): string => {
-  return s.substring(0, 1).toUpperCase() + s.substring(1)
+  return calloutMapping[callout] ?? "note"
 }
 
 // !?               -> optional embedding
@@ -108,28 +110,35 @@
 // ([^\[\]\|\#]+)   -> one or more non-special characters ([,],|, or #) (name)
 // (#[^\[\]\|\#]+)? -> # then one or more non-special characters (heading link)
 // (|[^\[\]\|\#]+)? -> | then one or more non-special characters (alias)
-const wikilinkRegex = new RegExp(/!?\[\[([^\[\]\|\#]+)(#[^\[\]\|\#]+)?(\|[^\[\]\|\#]+)?\]\]/, "g")
-const highlightRegex = new RegExp(/==(.+)==/, "g")
+export const wikilinkRegex = new RegExp(
+  /!?\[\[([^\[\]\|\#]+)?(#+[^\[\]\|\#]+)?(\|[^\[\]\|\#]+)?\]\]/,
+  "g",
+)
+const highlightRegex = new RegExp(/==([^=]+)==/, "g")
 const commentRegex = new RegExp(/%%(.+)%%/, "g")
 // from https://github.com/escwxyz/remark-obsidian-callout/blob/main/src/index.ts
 const calloutRegex = new RegExp(/^\[\!(\w+)\]([+-]?)/)
-// (?:^| )   -> non-capturing group, tag should start be separated by a space or be the start of the line
-// #(\w+)    -> tag itself is # followed by a string of alpha-numeric characters
-const tagRegex = new RegExp(/(?:^| )#(\w+)/, "g")
+const calloutLineRegex = new RegExp(/^> *\[\!\w+\][+-]?.*$/, "gm")
+// (?:^| )              -> non-capturing group, tag should start be separated by a space or be the start of the line
+// #(...)               -> capturing group, tag itself must start with #
+// (?:[-_\p{L}])+       -> non-capturing group, non-empty string of (Unicode-aware) alpha-numeric characters, hyphens and/or underscores
+// (?:\/[-_\p{L}]+)*)   -> non-capturing group, matches an arbitrary number of tag strings separated by "/"
+const tagRegex = new RegExp(/(?:^| )#((?:[-_\p{L}\d])+(?:\/[-_\p{L}\d]+)*)/, "gu")
+const blockReferenceRegex = new RegExp(/\^([A-Za-z0-9]+)$/, "g")
 
 export const ObsidianFlavoredMarkdown: QuartzTransformerPlugin<Partial<Options> | undefined> = (
   userOpts,
 ) => {
   const opts = { ...defaultOptions, ...userOpts }
 
+  const mdastToHtml = (ast: PhrasingContent | Paragraph) => {
+    const hast = toHast(ast, { allowDangerousHtml: true })!
+    return toHtml(hast, { allowDangerousHtml: true })
+  }
+
   const findAndReplace = opts.enableInHtmlEmbed
     ? (tree: Root, regex: RegExp, replace?: Replace | null | undefined) => {
         if (replace) {
-          const mdastToHtml = (ast: PhrasingContent) => {
-            const hast = toHast(ast, { allowDangerousHtml: true })!
-            return toHtml(hast, { allowDangerousHtml: true })
-          }
-
           visit(tree, "html", (node: HTML) => {
             if (typeof replace === "string") {
               node.value = node.value.replace(regex, replace)
@@ -157,18 +166,30 @@
   return {
     name: "ObsidianFlavoredMarkdown",
     textTransform(_ctx, src) {
+      // pre-transform blockquotes
+      if (opts.callouts) {
+        src = src.toString()
+        src = src.replaceAll(calloutLineRegex, (value) => {
+          // force newline after title of callout
+          return value + "\n> "
+        })
+      }
+
       // pre-transform wikilinks (fix anchors to things that may contain illegal syntax e.g. codeblocks, latex)
       if (opts.wikilinks) {
         src = src.toString()
-        return src.replaceAll(wikilinkRegex, (value, ...capture) => {
-          const [fp, rawHeader, rawAlias] = capture
-          const anchor = rawHeader?.trim().slice(1)
-          const displayAnchor = anchor ? `#${slugAnchor(anchor)}` : ""
-          const displayAlias = rawAlias ?? ""
+        src = src.replaceAll(wikilinkRegex, (value, ...capture) => {
+          const [rawFp, rawHeader, rawAlias] = capture
+          const fp = rawFp ?? ""
+          const anchor = rawHeader?.trim().replace(/^#+/, "")
+          const blockRef = Boolean(anchor?.startsWith("^")) ? "^" : ""
+          const displayAnchor = anchor ? `#${blockRef}${slugAnchor(anchor)}` : ""
+          const displayAlias = rawAlias ?? rawHeader?.replace("#", "|") ?? ""
           const embedDisplay = value.startsWith("!") ? "!" : ""
           return `${embedDisplay}[[${fp}${displayAnchor}${displayAlias}]]`
         })
       }
+
       return src
     },
     markdownPlugins() {
@@ -177,15 +198,15 @@
         plugins.push(() => {
           return (tree: Root, _file) => {
             findAndReplace(tree, wikilinkRegex, (value: string, ...capture: string[]) => {
-              let [fp, rawHeader, rawAlias] = capture
-              fp = fp.trim()
+              let [rawFp, rawHeader, rawAlias] = capture
+              const fp = rawFp?.trim() ?? ""
               const anchor = rawHeader?.trim() ?? ""
               const alias = rawAlias?.slice(1).trim()
 
               // embed cases
               if (value.startsWith("!")) {
-                const ext: string | undefined = path.extname(fp).toLowerCase()
-                const url = slugifyFilePath(fp as FilePath) + ext
+                const ext: string = path.extname(fp).toLowerCase()
+                const url = slugifyFilePath(fp as FilePath)
                 if ([".png", ".jpg", ".jpeg", ".gif", ".bmp", ".svg"].includes(ext)) {
                   const dims = alias ?? ""
                   let [width, height] = dims.split("x", 2)
@@ -218,14 +239,21 @@
                     type: "html",
                     value: `<iframe src="${url}"></iframe>`,
                   }
-                } else {
-                  // TODO: this is the node embed case
+                } else if (ext === "") {
+                  const block = anchor
+                  return {
+                    type: "html",
+                    data: { hProperties: { transclude: true } },
+                    value: `<blockquote class="transclude" data-url="${url}" data-block="${block}"><a href="${
+                      url + anchor
+                    }" class="transclude-inner">Transclude of ${url}${block}</a></blockquote>`,
+                  }
                 }
+
                 // otherwise, fall through to regular link
               }
 
               // internal link
-              // const url = transformInternalLink(fp + anchor)
               const url = fp + anchor
               return {
                 type: "link",
@@ -284,7 +312,7 @@
               }
 
               const text = firstChild.children[0].value
-              const restChildren = firstChild.children.splice(1)
+              const restChildren = firstChild.children.slice(1)
               const [firstLine, ...remainingLines] = text.split("\n")
               const remainingText = remainingLines.join("\n")
 
@@ -296,16 +324,21 @@
                 )
                 const collapse = collapseChar === "+" || collapseChar === "-"
                 const defaultState = collapseChar === "-" ? "collapsed" : "expanded"
-                const title =
+                const titleContent =
                   match.input.slice(calloutDirective.length).trim() || capitalize(calloutType)
+                const titleNode: Paragraph = {
+                  type: "paragraph",
+                  children: [{ type: "text", value: titleContent + " " }, ...restChildren],
+                }
+                const title = mdastToHtml(titleNode)
 
                 const toggleIcon = `<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="fold">
                   <polyline points="6 9 12 15 18 9"></polyline>
                 </svg>`
 
-                const titleNode: HTML = {
+                const titleHtml: HTML = {
                   type: "html",
-                  value: `<div 
+                  value: `<div
                   class="callout-title"
                 >
                   <div class="callout-icon">${callouts[calloutType]}</div>
@@ -314,7 +347,7 @@
                 </div>`,
                 }
 
-                const blockquoteContent: (BlockContent | DefinitionContent)[] = [titleNode]
+                const blockquoteContent: (BlockContent | DefinitionContent)[] = [titleHtml]
                 if (remainingText.length > 0) {
                   blockquoteContent.push({
                     type: "paragraph",
@@ -323,7 +356,6 @@
                         type: "text",
                         value: remainingText,
                       },
-                      ...restChildren,
                     ],
                   })
                 }
@@ -367,16 +399,20 @@
       if (opts.parseTags) {
         plugins.push(() => {
           return (tree: Root, file) => {
-            const slug = canonicalizeServer(file.data.slug!)
-            const base = pathToRoot(slug)
-            findAndReplace(tree, tagRegex, (value: string, tag: string) => {
-              if (file.data.frontmatter) {
+            const base = pathToRoot(file.data.slug!)
+            findAndReplace(tree, tagRegex, (_value: string, tag: string) => {
+              // Check if the tag only includes numbers
+              if (/^\d+$/.test(tag)) {
+                return false
+              }
+              tag = slugTag(tag)
+              if (file.data.frontmatter && !file.data.frontmatter.tags.includes(tag)) {
                 file.data.frontmatter.tags.push(tag)
               }
 
               return {
                 type: "link",
-                url: base + `/tags/${slugTag(tag)}`,
+                url: base + `/tags/${tag}`,
                 data: {
                   hProperties: {
                     className: ["tag-link"],
@@ -385,7 +421,7 @@
                 children: [
                   {
                     type: "text",
-                    value,
+                    value: `#${tag}`,
                   },
                 ],
               }
@@ -393,11 +429,65 @@
           }
         })
       }
-
       return plugins
     },
     htmlPlugins() {
-      return [rehypeRaw]
+      const plugins = [rehypeRaw]
+
+      if (opts.parseBlockReferences) {
+        plugins.push(() => {
+          const inlineTagTypes = new Set(["p", "li"])
+          const blockTagTypes = new Set(["blockquote"])
+          return (tree, file) => {
+            file.data.blocks = {}
+
+            visit(tree, "element", (node, index, parent) => {
+              if (blockTagTypes.has(node.tagName)) {
+                const nextChild = parent?.children.at(index! + 2) as Element
+                if (nextChild && nextChild.tagName === "p") {
+                  const text = nextChild.children.at(0) as Literal
+                  if (text && text.value && text.type === "text") {
+                    const matches = text.value.match(blockReferenceRegex)
+                    if (matches && matches.length >= 1) {
+                      parent!.children.splice(index! + 2, 1)
+                      const block = matches[0].slice(1)
+
+                      if (!Object.keys(file.data.blocks!).includes(block)) {
+                        node.properties = {
+                          ...node.properties,
+                          id: block,
+                        }
+                        file.data.blocks![block] = node
+                      }
+                    }
+                  }
+                }
+              } else if (inlineTagTypes.has(node.tagName)) {
+                const last = node.children.at(-1) as Literal
+                if (last && last.value && typeof last.value === "string") {
+                  const matches = last.value.match(blockReferenceRegex)
+                  if (matches && matches.length >= 1) {
+                    last.value = last.value.slice(0, -matches[0].length)
+                    const block = matches[0].slice(1)
+
+                    if (!Object.keys(file.data.blocks!).includes(block)) {
+                      node.properties = {
+                        ...node.properties,
+                        id: block,
+                      }
+                      file.data.blocks![block] = node
+                    }
+                  }
+                }
+              }
+            })
+
+            file.data.htmlAst = tree
+          }
+        })
+      }
+
+      return plugins
     },
     externalResources() {
       const js: JSResource[] = []
@@ -414,12 +504,16 @@
         js.push({
           script: `
           import mermaid from 'https://cdn.jsdelivr.net/npm/mermaid/dist/mermaid.esm.min.mjs';
+          const darkMode = document.documentElement.getAttribute('saved-theme') === 'dark'
+          mermaid.initialize({
+            startOnLoad: false,
+            securityLevel: 'loose',
+            theme: darkMode ? 'dark' : 'default'
+          });
           document.addEventListener('nav', async () => {
-            const darkMode = document.documentElement.getAttribute('saved-theme') === 'dark'
-            mermaid.initialize({ 
-              securityLevel: 'loose',
-              theme: darkMode ? 'dark' : 'default'
-            });
+            await mermaid.run({
+              querySelector: '.mermaid'
+            })
           });
           `,
           loadTime: "afterDOMReady",
@@ -432,3 +526,10 @@
     },
   }
 }
+
+declare module "vfile" {
+  interface DataMap {
+    blocks: Record<string, Element>
+    htmlAst: HtmlRoot
+  }
+}

--
Gitblit v1.10.0