Remove extra blank lines between list items in markdown export

[bgreenlee]

Fixes #1881

[vercel]

@bgreenlee is attempting to deploy a commit to the TypeCell Team on Vercel.

A member of the Team first needs to authorize it.

[pkg-pr-new]

Open in StackBlitz

@blocknote/ariakit

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/ariakit@1885

@blocknote/code-block

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/code-block@1885

@blocknote/core

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/core@1885

@blocknote/mantine

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/mantine@1885

@blocknote/react

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/react@1885

@blocknote/server-util

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/server-util@1885

@blocknote/shadcn

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/shadcn@1885

@blocknote/xl-ai

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/xl-ai@1885

@blocknote/xl-docx-exporter

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/xl-docx-exporter@1885

@blocknote/xl-email-exporter

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/xl-email-exporter@1885

@blocknote/xl-multi-column

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/xl-multi-column@1885

@blocknote/xl-odt-exporter

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/xl-odt-exporter@1885

@blocknote/xl-pdf-exporter

npm i https://pkg.pr.new/TypeCellOS/BlockNote/@blocknote/xl-pdf-exporter@1885

commit: 99215eb

[nperez0111]

I don't think that this solution is sufficient enough, you'll see that in your test cases it also strips the new line between lists (e.g. in markdown/lists/basic.md the bullet list to numbered list boundary).

I think there are two solutions here:

1. Make the markdown parser smarter (we have a somewhat unusual structure of <li><p>Text content which I'm sure is why the parser is tripping up)
2. Simplify the the HTML output so that the markdown parser doesn't have to be so smart (by removing that inner paragraph element & unpacking it's children into the <li> element)

I took a stab at the second approach and got something that worked better:

diff --git i/packages/core/src/api/exporters/html/util/serializeBlocksExternalHTML.ts w/packages/core/src/api/exporters/html/util/serializeBlocksExternalHTML.ts
index f74757c8d..9dc25da9a 100644
--- i/packages/core/src/api/exporters/html/util/serializeBlocksExternalHTML.ts
+++ w/packages/core/src/api/exporters/html/util/serializeBlocksExternalHTML.ts
@@ -172,6 +172,14 @@ function serializeBlock<
       fragment.append(list);
     }
     const li = doc.createElement("li");
+    //unpack p nodes into their children
+    const childNodes = Array.from(elementFragment.childNodes);
+    for (const child of childNodes) {
+      if (child.nodeName === "P") {
+        const children = Array.from(child.childNodes);
+        child.replaceWith(...children);
+      }
+    }
     li.append(elementFragment);
     fragment.lastChild!.appendChild(li);
   } else {
diff --git i/packages/core/src/api/exporters/markdown/markdownExporter.ts w/packages/core/src/api/exporters/markdown/markdownExporter.ts
index 812f20ffb..ce82f56f9 100644
--- i/packages/core/src/api/exporters/markdown/markdownExporter.ts
+++ w/packages/core/src/api/exporters/markdown/markdownExporter.ts
@@ -25,6 +25,8 @@ export function cleanHTMLToMarkdown(cleanHTMLString: string) {
     );
   }
 
+  console.log("cleanHTMLString", cleanHTMLString);
+
   const markdownString = deps.unified
     .unified()
     .use(deps.rehypeParse.default, { fragment: true })
@@ -37,10 +39,7 @@ export function cleanHTMLToMarkdown(cleanHTMLString: string) {
     })
     .processSync(cleanHTMLString);
 
-  let result = markdownString.value as string;
-
-  // Remove extra blank lines between list items
-  result = result.replace(/\n\n(?=\d+\.|-|\*)/g, "\n");
+  const result = markdownString.value as string;
 
   return result;
 }

Before going with this solution though, we'd need to double check that things like background colors still work for the HTML output, I don't remember at this moment whether they are set on the paragraph or higher. But, if we can be sure that the output isn't affected by this, then unpacking paragraphs may end up being the better solution since it is more HTML-like & easier to parse

[bgreenlee]

Make sense. Let me know if I can help.