feat(core): import progress & perf (#15197)

#### PR Dependency Tree


* **PR #15197** 👈

This tree was auto-generated by
[Charcoal](https://github.com/danerwilliams/charcoal)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Added a new import pipeline with “plan then commit” batch handling for
Markdown, Notion HTML, Obsidian, and Bear backups.
* Enabled native import sessions with progress, cancellation, and
batch-by-batch committing (including assets, folders, icons, and tags).
  * Added web preflight limits for ZIP and multi-file imports.
* **Bug Fixes**
* Improved import error/warning reporting and continued processing when
some items fail.
* Strengthened snapshot-based file/directory picking to preserve paths.
* **Chores**
  * Updated project packaging/configuration for the new import workflow.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

blocksuite/affine/widgets/linked-doc/src/transformers/bear.ts

@@ -7,17 +7,30 @@ import {
 } from '@blocksuite/affine-shared/adapters';
 import { Container } from '@blocksuite/global/di';
 import { sha } from '@blocksuite/global/utils';
-import type { ExtensionType, Schema, Workspace } from '@blocksuite/store';
+import type {
+  DocSnapshot,
+  ExtensionType,
+  Schema,
+  Workspace,
+} from '@blocksuite/store';
 import { extMimeMap, Transformer } from '@blocksuite/store';
 import JSZip from 'jszip';
 
+import {
+  blobsFromAssets,
+  type ImportBatch,
+  type ImportDoc,
+  type ImportFolder,
+  type ImportTag,
+  type ImportWarning,
+} from './import-batch.js';
 import { createCollectionDocCRUD } from './markdown.js';
 
 /** Recursive tree node representing a tag-based folder hierarchy. */
-type FolderHierarchy = {
+export type BearFolderHierarchy = {
   name: string;
   path: string;
-  children: Map<string, FolderHierarchy>;
+  children: Map<string, BearFolderHierarchy>;
   pageId?: string;
   parentPath?: string;
 };
@@ -29,10 +42,11 @@ type BearImportOptions = {
   extensions: ExtensionType[];
 };
 
-type BearImportResult = {
+export type PlanBearBackupResult = {
   docIds: string[];
   tags: Map<string, string[]>;
-  folderHierarchy: FolderHierarchy;
+  folderHierarchy: BearFolderHierarchy;
+  batch: ImportBatch;
 };
 
 type BundleEntry = {
@@ -182,8 +196,8 @@ function deduplicateTags(tags: string[]): string[] {
  */
 function buildFolderHierarchyFromTags(
   tagDocMap: Map<string, string[]>
-): FolderHierarchy {
-  const root: FolderHierarchy = {
+): BearFolderHierarchy {
+  const root: BearFolderHierarchy = {
     name: '',
     path: '',
     children: new Map(),
@@ -226,6 +240,34 @@ function buildFolderHierarchyFromTags(
   return root;
 }
 
+function flattenFolderHierarchy(root: BearFolderHierarchy): ImportFolder[] {
+  const folders: ImportFolder[] = [];
+
+  const visit = (node: BearFolderHierarchy) => {
+    if (node.name) {
+      folders.push({
+        path: node.path,
+        name: node.name,
+        parentPath: node.parentPath,
+        pageId: node.pageId,
+      });
+    }
+    for (const child of node.children.values()) {
+      visit(child);
+    }
+  };
+
+  for (const child of root.children.values()) {
+    visit(child);
+  }
+
+  return folders;
+}
+
+function tagsToBatchTags(tags: Map<string, string[]>): ImportTag[] {
+  return Array.from(tags, ([name, docIds]) => ({ name, docIds }));
+}
+
 const GFM_CALLOUT_MAP: Record<string, string> = {
   IMPORTANT: '\u26A0',
   NOTE: '\uD83D\uDCDD',
@@ -250,9 +292,9 @@ function convertGfmCallouts(markdown: string): string {
     if (!inCodeBlock) {
       lines[i] = lines[i].replace(
         /^(>\s*)\[!(\w+)\]/,
-        (_match, prefix: string, type: string) => {
+        (match, prefix: string, type: string) => {
           const emoji = GFM_CALLOUT_MAP[type.toUpperCase()];
-          return emoji ? `${prefix}[!${emoji}]` : _match;
+          return emoji ? `${prefix}[!${emoji}]` : match;
         }
       );
     }
@@ -306,7 +348,7 @@ function convertHighlights(markdown: string): string {
     if (!inCodeBlock) {
       lines[i] = lines[i].replace(
         /==(\S(?:[^=]|=[^=])*?)==/g,
-        (_match, content: string) => {
+        (...[, content]) => {
           const firstChar = String.fromCodePoint(content.codePointAt(0)!);
           const color = HIGHLIGHT_COLOR_MAP[firstChar];
           if (color) {
@@ -340,16 +382,24 @@ function extractTitle(markdown: string, bundleName: string): string {
   return bundleName.replace(/\.textbundle$/i, '') || 'Untitled';
 }
 
+function applySnapshotTitle(snapshot: DocSnapshot, title: string) {
+  snapshot.meta.title = title;
+  snapshot.blocks.props.title = {
+    '$blocksuite:internal:text$': true,
+    delta: [{ insert: title }],
+  };
+}
+
 /**
  * Import a Bear .bear2bk backup file.
  * Uses JSZip for lazy/streaming decompression to handle large backups.
  */
-async function importBearBackup({
+async function planBearBackup({
   collection,
   schema,
   imported,
   extensions,
-}: BearImportOptions): Promise<BearImportResult> {
+}: BearImportOptions): Promise<PlanBearBackupResult> {
   const provider = getProvider(extensions);
 
   // JSZip reads the zip directory without decompressing all entries
@@ -358,7 +408,7 @@ async function importBearBackup({
   // Scan entries and group by textbundle
   const bundleMap = new Map<string, BundleEntry>();
 
-  zip.forEach((path, _entry) => {
+  zip.forEach(path => {
     if (path.includes('__MACOSX') || path.includes('.DS_Store')) return;
 
     const tbMatch = path.match(/^(.+?\.textbundle)\/(.*)/i);
@@ -421,6 +471,12 @@ async function importBearBackup({
   }
 
   const docIds: string[] = [];
+  const docs: ImportDoc[] = [];
+  const importBlobs = new Map<
+    string,
+    Awaited<ReturnType<typeof blobsFromAssets>>[0]
+  >();
+  const warnings: ImportWarning[] = [];
   const tagDocMap = new Map<string, string[]>();
 
   // Process bundles sequentially to limit memory.
@@ -497,44 +553,70 @@ async function importBearBackup({
       }
 
       const mdAdapter = new MarkdownAdapter(job, provider);
-      const doc = await mdAdapter.toDoc({
+      const snapshot = await mdAdapter.toDocSnapshot({
         file: markdown,
         assets: job.assetsManager,
       });
 
-      if (doc) {
-        docIds.push(doc.id);
-
-        const metaPatch: Record<string, unknown> = {};
-        if (bearMeta?.creationDate) {
-          const ts = Date.parse(String(bearMeta.creationDate));
-          if (!isNaN(ts)) metaPatch.createDate = ts;
-        }
-        if (bearMeta?.modificationDate) {
-          const ts = Date.parse(String(bearMeta.modificationDate));
-          if (!isNaN(ts)) metaPatch.updatedDate = ts;
-        }
-        if (Object.keys(metaPatch).length) {
-          collection.meta.setDocMeta(doc.id, metaPatch);
-        }
-
-        for (const tag of tags) {
-          if (!tagDocMap.has(tag)) {
-            tagDocMap.set(tag, []);
-          }
-          tagDocMap.get(tag)!.push(doc.id);
-        }
+      const docId = collection.idGenerator();
+      snapshot.meta.id = docId;
+      applySnapshotTitle(snapshot, title);
+      const meta: ImportDoc['meta'] = { title };
+      if (bearMeta?.creationDate) {
+        const ts = Date.parse(String(bearMeta.creationDate));
+        if (!isNaN(ts)) meta.createDate = ts;
       }
-    } catch (err) {
-      console.warn(`Failed to import bundle: ${entry.bundlePath}`, err);
+      if (bearMeta?.modificationDate) {
+        const ts = Date.parse(String(bearMeta.modificationDate));
+        if (!isNaN(ts)) meta.updatedDate = ts;
+      }
+
+      docs.push({
+        id: docId,
+        snapshot,
+        meta,
+      });
+      docIds.push(docId);
+
+      for (const tag of tags) {
+        if (!tagDocMap.has(tag)) {
+          tagDocMap.set(tag, []);
+        }
+        tagDocMap.get(tag)!.push(docId);
+      }
+
+      for (const blob of await blobsFromAssets(
+        pendingAssets,
+        pendingPathBlobIdMap
+      )) {
+        importBlobs.set(blob.blobId, blob);
+      }
+    } catch {
+      warnings.push({
+        code: 'bear-bundle-import-failed',
+        message: `Failed to import bundle: ${entry.bundlePath}`,
+        sourcePath: entry.bundlePath,
+      });
     }
   }
 
   const folderHierarchy = buildFolderHierarchyFromTags(tagDocMap);
-  return { docIds, tags: tagDocMap, folderHierarchy };
+  return {
+    docIds,
+    tags: tagDocMap,
+    folderHierarchy,
+    batch: {
+      docs,
+      blobs: Array.from(importBlobs.values()),
+      folders: flattenFolderHierarchy(folderHierarchy),
+      tags: tagsToBatchTags(tagDocMap),
+      warnings,
+      done: true,
+    },
+  };
 }
 
 /** Public API for importing Bear .bear2bk backup archives. */
 export const BearTransformer = {
-  importBearBackup,
+  planBearBackup,
 };

blocksuite/affine/widgets/linked-doc/src/transformers/import-batch.ts

@@ -0,0 +1,157 @@
+import {
+  type DocMeta,
+  type DocSnapshot,
+  type Schema,
+  Transformer,
+  type Workspace,
+} from '@blocksuite/store';
+
+export type ImportBlob = {
+  blobId: string;
+  sourcePath: string;
+  fileName: string;
+  mime: string;
+  bytes: Uint8Array;
+};
+
+export type ImportIconData =
+  | {
+      type: 'emoji';
+      unicode: string;
+    }
+  | {
+      type: 'image';
+      content: string;
+    };
+
+export type ImportFolder = {
+  path: string;
+  name: string;
+  parentPath?: string;
+  pageId?: string;
+  icon?: ImportIconData;
+};
+
+export type ImportDoc = {
+  id: string;
+  sourcePath?: string;
+  snapshot: DocSnapshot;
+  meta?: Partial<
+    Pick<
+      DocMeta,
+      'title' | 'createDate' | 'updatedDate' | 'tags' | 'favorite' | 'trash'
+    >
+  >;
+};
+
+export type ImportTag = {
+  name: string;
+  docIds: string[];
+};
+
+export type ImportIcon = {
+  docId: string;
+  icon: ImportIconData;
+};
+
+export type ImportWarning = {
+  code: string;
+  message: string;
+  sourcePath?: string;
+};
+
+export type ImportBatch = {
+  docs: ImportDoc[];
+  blobs: ImportBlob[];
+  folders?: ImportFolder[];
+  tags?: ImportTag[];
+  icons?: ImportIcon[];
+  warnings?: ImportWarning[];
+  progress?: {
+    completed: number;
+    total: number;
+  };
+  entryId?: string;
+  isWorkspaceFile?: boolean;
+  done: boolean;
+};
+
+export type ImportCommitResult = {
+  docIds: string[];
+  entryId?: string;
+  isWorkspaceFile?: boolean;
+  rootFolderId?: string;
+  warnings: ImportWarning[];
+};
+
+export async function blobsFromAssets(
+  assets: ReadonlyMap<string, File>,
+  pathBlobIdMap: ReadonlyMap<string, string> = new Map()
+) {
+  const sourcePathByBlobId = new Map<string, string>();
+  for (const [path, blobId] of pathBlobIdMap) {
+    sourcePathByBlobId.set(blobId, path);
+  }
+
+  return Promise.all(
+    Array.from(
+      assets,
+      async ([blobId, file]): Promise<ImportBlob> => ({
+        blobId,
+        sourcePath: sourcePathByBlobId.get(blobId) ?? file.name,
+        fileName: file.name,
+        mime: file.type,
+        bytes: new Uint8Array(await file.arrayBuffer()),
+      })
+    )
+  );
+}
+
+function copyToArrayBuffer(bytes: Uint8Array) {
+  const copy = new Uint8Array(bytes.byteLength);
+  copy.set(bytes);
+  return copy.buffer;
+}
+
+export async function commitImportBatchToWorkspace(
+  collection: Workspace,
+  schema: Schema,
+  batch: ImportBatch
+): Promise<ImportCommitResult> {
+  for (const blob of batch.blobs) {
+    await collection.blobSync.set(
+      blob.blobId,
+      new File([copyToArrayBuffer(blob.bytes)], blob.fileName, {
+        type: blob.mime,
+      })
+    );
+  }
+
+  const transformer = new Transformer({
+    schema,
+    blobCRUD: collection.blobSync,
+    docCRUD: {
+      create: (id: string) => collection.createDoc(id).getStore({ id }),
+      get: (id: string) => collection.getDoc(id)?.getStore({ id }) ?? null,
+      delete: (id: string) => collection.removeDoc(id),
+    },
+    middlewares: [],
+  });
+
+  const docIds: string[] = [];
+  for (const doc of batch.docs) {
+    const store = await transformer.snapshotToDoc(doc.snapshot);
+    if (!store) continue;
+    docIds.push(store.id);
+    if (doc.meta && Object.keys(doc.meta).length) {
+      collection.meta.setDocMeta(store.id, doc.meta);
+    }
+  }
+
+  return {
+    docIds,
+    entryId: batch.entryId,
+    isWorkspaceFile: batch.isWorkspaceFile,
+    warnings: batch.warnings ?? [],
+  };
+}

blocksuite/affine/widgets/linked-doc/src/transformers/index.ts

@@ -1,9 +1,22 @@
 export { BearTransformer } from './bear.js';
 export { DocxTransformer } from './docx.js';
 export { HtmlTransformer } from './html.js';
+export {
+  blobsFromAssets,
+  commitImportBatchToWorkspace,
+  type ImportBatch,
+  type ImportBlob,
+  type ImportCommitResult,
+  type ImportDoc,
+  type ImportFolder,
+  type ImportIcon,
+  type ImportIconData,
+  type ImportTag,
+  type ImportWarning,
+} from './import-batch.js';
 export { MarkdownTransformer } from './markdown.js';
 export { NotionHtmlTransformer } from './notion-html.js';
 export { ObsidianTransformer } from './obsidian.js';
 export { PdfTransformer } from './pdf.js';
-export { createAssetsArchive, download } from './utils.js';
+export { createAssetsArchive, download, Unzip } from './utils.js';
 export { ZipTransformer } from './zip.js';

blocksuite/affine/widgets/linked-doc/src/transformers/markdown.ts

@@ -16,6 +16,7 @@ import { BlockSuiteError, ErrorCode } from '@blocksuite/global/exceptions';
 import { sha } from '@blocksuite/global/utils';
 import type {
   DocMeta,
+  DocSnapshot,
   ExtensionType,
   Schema,
   Store,
@@ -23,6 +24,12 @@ import type {
 } from '@blocksuite/store';
 import { extMimeMap, Transformer } from '@blocksuite/store';
 
+import {
+  blobsFromAssets,
+  type ImportBatch,
+  type ImportDoc,
+  type ImportFolder,
+} from './import-batch.js';
 import type { AssetMap, ImportedFileEntry, PathBlobIdMap } from './type.js';
 import { createAssetsArchive, download, parseMatter, Unzip } from './utils.js';
 
@@ -438,6 +445,14 @@ function prepareNotionMarkdownFile({
   };
 }
 
+function applySnapshotTitle(snapshot: DocSnapshot, title: string) {
+  snapshot.meta.title = title;
+  snapshot.blocks.props.title = {
+    '$blocksuite:internal:text$': true,
+    delta: [{ insert: title }],
+  };
+}
+
 /**
  * Filters hidden/system entries that should never participate in imports.
  */
@@ -685,7 +700,7 @@ async function importMarkdownToDoc({
  * @param options.imported The zip file as a Blob
  * @returns A Promise that resolves to an array of IDs of the newly created docs
  */
-type FolderHierarchy = {
+export type FolderHierarchy = {
   name: string;
   path: string;
   children: Map<string, FolderHierarchy>;
@@ -693,18 +708,19 @@ type FolderHierarchy = {
   parentPath?: string;
 };
 
-export type ImportMarkdownZipResult = {
+export type PlanMarkdownZipResult = {
   docIds: string[];
   folderHierarchy?: FolderHierarchy;
+  batch: ImportBatch;
 };
 
-async function importMarkdownZip({
+async function planMarkdownZip({
   collection,
   schema,
   imported,
   extensions,
-}: ImportMarkdownZipOptions): Promise<ImportMarkdownZipResult> {
-  return importMarkdownZipInternal({
+}: ImportMarkdownZipOptions): Promise<PlanMarkdownZipResult> {
+  return planMarkdownZipInternal({
     collection,
     schema,
     imported,
@@ -712,13 +728,13 @@ async function importMarkdownZip({
   });
 }
 
-async function importNotionMarkdownZip({
+async function planNotionMarkdownZip({
   collection,
   schema,
   imported,
   extensions,
-}: ImportMarkdownZipOptions): Promise<ImportMarkdownZipResult> {
-  return importMarkdownZipInternal({
+}: ImportMarkdownZipOptions): Promise<PlanMarkdownZipResult> {
+  return planMarkdownZipInternal({
     collection,
     schema,
     imported,
@@ -731,7 +747,7 @@ async function importNotionMarkdownZip({
   });
 }
 
-async function importMarkdownZipInternal({
+async function planMarkdownZipInternal({
   collection,
   schema,
   imported,
@@ -741,12 +757,13 @@ async function importMarkdownZipInternal({
   prepareMarkdownFile = prepareDefaultMarkdownFile,
   preserveCommonRoot = false,
   recursiveZip = false,
-}: ImportMarkdownZipInternalOptions): Promise<ImportMarkdownZipResult> {
+}: ImportMarkdownZipInternalOptions): Promise<PlanMarkdownZipResult> {
   const provider = getProvider([
     markdownZipDocLinkToDeltaMatcher,
     ...extensions,
   ]);
   const docIds: string[] = [];
+  const docs: ImportDoc[] = [];
   const pendingAssets: AssetMap = new Map();
   const pendingPathBlobIdMap: PathBlobIdMap = new Map();
   const docPathMap: Array<{ fullPath: string; docId: string }> = [];
@@ -812,16 +829,20 @@ async function importMarkdownZipInternal({
         assets: job.assetsManager,
       });
       snapshot.meta.id = pageId;
-      const doc = await job.snapshotToDoc(snapshot);
-      if (doc) {
-        applyMetaPatch(collection, doc.id, meta);
-        docIds.push(doc.id);
-        docPathMap.push({ fullPath, docId: doc.id });
-      }
+      applySnapshotTitle(snapshot, preferredTitle);
+      docs.push({
+        id: pageId,
+        snapshot,
+        meta: {
+          ...meta,
+          title: preferredTitle,
+        },
+      });
+      docIds.push(pageId);
+      docPathMap.push({ fullPath, docId: pageId });
     })
   );
 
-  // Build folder hierarchy from zip paths
   const folderHierarchy = buildMarkdownZipFolderHierarchy(
     docPathMap,
     normalizeFolderName,
@@ -829,7 +850,18 @@ async function importMarkdownZipInternal({
     createRootFolderForTopLevelDocs
   );
 
-  return { docIds, folderHierarchy };
+  return {
+    docIds,
+    folderHierarchy,
+    batch: {
+      docs,
+      blobs: await blobsFromAssets(pendingAssets, pendingPathBlobIdMap),
+      folders: folderHierarchy
+        ? flattenFolderHierarchy(folderHierarchy)
+        : undefined,
+      done: true,
+    },
+  };
 }
 
 /**
@@ -921,10 +953,34 @@ function buildMarkdownZipFolderHierarchy(
   return root.children.size > 0 ? root : undefined;
 }
 
+function flattenFolderHierarchy(root: FolderHierarchy): ImportFolder[] {
+  const folders: ImportFolder[] = [];
+
+  const visit = (node: FolderHierarchy) => {
+    if (node.name) {
+      folders.push({
+        path: node.path,
+        name: node.name,
+        parentPath: node.parentPath,
+        pageId: node.pageId,
+      });
+    }
+    for (const child of node.children.values()) {
+      visit(child);
+    }
+  };
+
+  for (const child of root.children.values()) {
+    visit(child);
+  }
+
+  return folders;
+}
+
 export const MarkdownTransformer = {
   exportDoc,
   importMarkdownToBlock,
   importMarkdownToDoc,
-  importMarkdownZip,
-  importNotionMarkdownZip,
+  planMarkdownZip,
+  planNotionMarkdownZip,
 };

blocksuite/affine/widgets/linked-doc/src/transformers/notion-html.ts

@@ -12,6 +12,14 @@ import {
   type Workspace,
 } from '@blocksuite/store';
 
+import {
+  blobsFromAssets,
+  type ImportBatch,
+  type ImportDoc,
+  type ImportFolder,
+  type ImportIconData,
+  type ImportWarning,
+} from './import-batch.js';
 import { Unzip } from './utils.js';
 
 type ImportNotionZipOptions = {
@@ -21,12 +29,12 @@ type ImportNotionZipOptions = {
   extensions: ExtensionType[];
 };
 
-type PageIcon = {
+export type PageIcon = {
   type: 'emoji' | 'image';
-  content: string; // emoji unicode or image URL/data
+  content: string;
 };
 
-type FolderHierarchy = {
+export type FolderHierarchy = {
   name: string;
   path: string;
   children: Map<string, FolderHierarchy>;
@@ -35,12 +43,13 @@ type FolderHierarchy = {
   icon?: PageIcon;
 };
 
-type ImportNotionZipResult = {
+export type PlanNotionHtmlZipResult = {
   entryId: string | undefined;
   pageIds: string[];
   isWorkspaceFile: boolean;
   hasMarkdown: boolean;
   folderHierarchy?: FolderHierarchy;
+  batch: ImportBatch;
 };
 
 function getProvider(extensions: ExtensionType[]) {
@@ -61,22 +70,9 @@ function parseFolderPath(filePath: string): {
 }
 
 function extractPageIcon(doc: Document): PageIcon | undefined {
-  // Look for Notion page icon in the HTML
-  // Notion export format: <div class="page-header-icon undefined"><span class="icon">✅</span></div>
-
-  console.log('=== Extracting page icon ===');
-
-  // Check if there's a head section with title for debugging
-  const headTitle = doc.querySelector('head title');
-  if (headTitle) {
-    console.log('Page title from head:', headTitle.textContent);
-  }
-
-  // Look for the exact Notion export structure: .page-header-icon .icon
   const notionIconSpan = doc.querySelector('.page-header-icon .icon');
   if (notionIconSpan && notionIconSpan.textContent) {
     const iconContent = notionIconSpan.textContent.trim();
-    console.log('Found Notion icon (.page-header-icon .icon):', iconContent);
     if (/\p{Emoji}/u.test(iconContent)) {
       return {
         type: 'emoji',
@@ -85,50 +81,29 @@ function extractPageIcon(doc: Document): PageIcon | undefined {
     }
   }
 
-  // Look for page header area for debugging
-  const pageHeader = doc.querySelector('.page-header-icon');
-  if (pageHeader) {
-    console.log(
-      'Found .page-header-icon:',
-      pageHeader.outerHTML.substring(0, 300) + '...'
-    );
-  }
-
-  // Fallback: try to find emoji icons with older selectors
   const emojiIcon = doc.querySelector('.page-header-icon .notion-emoji');
   if (emojiIcon && emojiIcon.textContent) {
-    console.log(
-      'Found emoji icon (.page-header-icon .notion-emoji):',
-      emojiIcon.textContent
-    );
     return {
       type: 'emoji',
       content: emojiIcon.textContent.trim(),
     };
   }
 
-  // Try alternative emoji selectors
   const altEmojiIcon = doc.querySelector('[role="img"][aria-label]');
   if (
     altEmojiIcon &&
     altEmojiIcon.textContent &&
     /\p{Emoji}/u.test(altEmojiIcon.textContent)
   ) {
-    console.log(
-      'Found emoji icon ([role="img"][aria-label]):',
-      altEmojiIcon.textContent
-    );
     return {
       type: 'emoji',
       content: altEmojiIcon.textContent.trim(),
     };
   }
 
-  // Look for image icons in the page header
   const imageIcon = doc.querySelector('.page-header-icon img');
   if (imageIcon) {
     const src = imageIcon.getAttribute('src');
-    console.log('Found image icon (.page-header-icon img):', src);
     if (src) {
       return {
         type: 'image',
@@ -137,27 +112,15 @@ function extractPageIcon(doc: Document): PageIcon | undefined {
     }
   }
 
-  // Fallback: Look for any span with emoji class "icon" in page header area
   const iconSpans = doc.querySelectorAll('span.icon');
   for (const span of iconSpans) {
     if (span.textContent && /\p{Emoji}/u.test(span.textContent.trim())) {
       const parent = span.parentElement;
-      console.log(
-        'Found emoji in span.icon:',
-        span.textContent,
-        'parent classes:',
-        parent?.className
-      );
-      // Check if this is in a page header context
       if (
         parent &&
         (parent.classList.contains('page-header-icon') ||
           parent.closest('.page-header-icon'))
       ) {
-        console.log(
-          'Using emoji from span.icon in page header:',
-          span.textContent
-        );
         return {
           type: 'emoji',
           content: span.textContent.trim(),
@@ -166,15 +129,11 @@ function extractPageIcon(doc: Document): PageIcon | undefined {
     }
   }
 
-  // Fallback: Try to find icons in the page title area that might contain emoji
   const pageTitle = doc.querySelector('.page-title, h1');
   if (pageTitle && pageTitle.textContent) {
-    console.log('Page title element found:', pageTitle.textContent);
     const text = pageTitle.textContent.trim();
-    // Check if the title starts with an emoji
     const emojiMatch = text.match(/^(\p{Emoji}+)/u);
     if (emojiMatch) {
-      console.log('Found emoji in title:', emojiMatch[1]);
       return {
         type: 'emoji',
         content: emojiMatch[1],
@@ -182,7 +141,6 @@ function extractPageIcon(doc: Document): PageIcon | undefined {
     }
   }
 
-  console.log('No page icon found');
   return undefined;
 }
 
@@ -200,7 +158,6 @@ function buildFolderHierarchy(
     let current = root;
     let currentPath = '';
 
-    // Navigate/create folder structure
     for (const folderName of folderParts) {
       const parentPath = currentPath;
       currentPath = currentPath ? `${currentPath}/${folderName}` : folderName;
@@ -216,7 +173,6 @@ function buildFolderHierarchy(
       current = current.children.get(folderName)!;
     }
 
-    // If this is a page file, associate it with the current folder
     if (fileName.endsWith('.html') && !fileName.startsWith('index.html')) {
       const pageName = fileName.replace(/\.html$/, '');
       if (!current.children.has(pageName)) {
@@ -229,7 +185,6 @@ function buildFolderHierarchy(
           icon: icon,
         });
       } else {
-        // Update existing entry with pageId and icon
         const existingPage = current.children.get(pageName)!;
         existingPage.pageId = pageId;
         if (icon) {
@@ -242,29 +197,59 @@ function buildFolderHierarchy(
   return root;
 }
 
-/**
- * Imports a Notion zip file into the BlockSuite collection.
- *
- * @param options - The options for importing.
- * @param options.collection - The BlockSuite document collection.
- * @param options.schema - The schema of the BlockSuite document collection.
- * @param options.imported - The imported zip file as a Blob.
- *
- * @returns A promise that resolves to an object containing:
- *          - entryId: The ID of the entry page (if any).
- *          - pageIds: An array of imported page IDs.
- *          - isWorkspaceFile: Whether the imported file is a workspace file.
- *          - hasMarkdown: Whether the zip contains markdown files.
- *          - folderHierarchy: The parsed folder hierarchy from the Notion export.
- */
-async function importNotionZip({
+function toImportIconData(icon?: PageIcon): ImportIconData | undefined {
+  if (!icon) return undefined;
+  if (icon.type === 'emoji') {
+    return {
+      type: 'emoji',
+      unicode: icon.content,
+    };
+  }
+  return {
+    type: 'image',
+    content: icon.content,
+  };
+}
+
+function flattenFolderHierarchy(root: FolderHierarchy): ImportFolder[] {
+  const folders: ImportFolder[] = [];
+
+  const visit = (node: FolderHierarchy) => {
+    if (node.name) {
+      folders.push({
+        path: node.path,
+        name: node.name,
+        parentPath: node.parentPath,
+        pageId: node.pageId,
+        icon: toImportIconData(node.icon),
+      });
+    }
+    for (const child of node.children.values()) {
+      visit(child);
+    }
+  };
+
+  for (const child of root.children.values()) {
+    visit(child);
+  }
+
+  return folders;
+}
+
+async function planNotionHtmlZip({
   collection,
   schema,
   imported,
   extensions,
-}: ImportNotionZipOptions): Promise<ImportNotionZipResult> {
+}: ImportNotionZipOptions): Promise<PlanNotionHtmlZipResult> {
   const provider = getProvider(extensions);
   const pageIds: string[] = [];
+  const docs: ImportDoc[] = [];
+  const blobs = new Map<
+    string,
+    Awaited<ReturnType<typeof blobsFromAssets>>[0]
+  >();
+  const warnings: ImportWarning[] = [];
   let isWorkspaceFile = false;
   let hasMarkdown = false;
   let entryId: string | undefined;
@@ -280,7 +265,7 @@ async function importNotionZip({
     const pageMap = new Map<string, string>();
     const pagePaths: string[] = [];
     const promises: Promise<void>[] = [];
-    const pendingAssets = new Map<string, Blob>();
+    const pendingAssets = new Map<string, File>();
     const pendingPathBlobIdMap = new Map<string, string>();
     for (const { path, content, index } of unzip) {
       if (path.startsWith('__MACOSX/')) continue;
@@ -305,11 +290,7 @@ async function importNotionZip({
           const text = await content.text();
           const doc = new DOMParser().parseFromString(text, 'text/html');
           const pageBody = doc.querySelector('.page-body');
-          if (pageBody && pageBody.children.length === 0) {
-            // Skip empty pages
-            continue;
-          }
-          // Extract page icon from the HTML
+          if (pageBody && pageBody.children.length === 0) continue;
           pageIcon = extractPageIcon(doc);
         }
 
@@ -327,10 +308,12 @@ async function importNotionZip({
         continue;
       }
       if (index === 0 && fileName.endsWith('.csv')) {
-        window.open(
-          'https://affine.pro/blog/import-your-data-from-notion-into-affine',
-          '_blank'
-        );
+        warnings.push({
+          code: 'notion-csv-export',
+          message:
+            'The imported Notion export appears to be CSV instead of HTML.',
+          sourcePath: path,
+        });
         continue;
       }
       if (fileName.endsWith('.zip')) {
@@ -342,7 +325,7 @@ async function importNotionZip({
       }
       const blob = content;
       const ext = path.split('.').at(-1) ?? '';
-      const mime = extMimeMap.get(ext) ?? '';
+      const mime = extMimeMap.get(ext.toLowerCase()) ?? '';
       const key = await sha(await blob.arrayBuffer());
       const filePathSplit = path.split('/');
       while (filePathSplit.length > 1) {
@@ -375,32 +358,57 @@ async function importNotionZip({
           pathBlobIdMap.set(key, value);
         }
       }
-      const page = await htmlAdapter.toDoc({
+      const snapshot = await htmlAdapter.toDocSnapshot({
         file: await zipFile.get(path)!.text(),
         pageId: pageMap.get(path),
         pageMap,
         assets: job.assetsManager,
       });
-      if (page) {
-        pageIds.push(page.id);
-      }
+      docs.push({
+        id: snapshot.meta.id,
+        snapshot,
+      });
+      pageIds.push(snapshot.meta.id);
     });
     promises.push(...pagePromises);
+    promises.push(
+      blobsFromAssets(pendingAssets, pendingPathBlobIdMap).then(importBlobs => {
+        for (const blob of importBlobs) {
+          blobs.set(blob.blobId, blob);
+        }
+      })
+    );
     return promises;
   };
   const allPromises = await parseZipFile(imported);
   await Promise.all(allPromises.flat());
   entryId = entryId ?? pageIds[0];
 
-  // Build folder hierarchy from collected paths
   const folderHierarchy =
     pagePathsWithIds.length > 0
       ? buildFolderHierarchy(pagePathsWithIds)
       : undefined;
 
-  return { entryId, pageIds, isWorkspaceFile, hasMarkdown, folderHierarchy };
+  return {
+    entryId,
+    pageIds,
+    isWorkspaceFile,
+    hasMarkdown,
+    folderHierarchy,
+    batch: {
+      docs,
+      blobs: Array.from(blobs.values()),
+      folders: folderHierarchy
+        ? flattenFolderHierarchy(folderHierarchy)
+        : undefined,
+      warnings,
+      entryId,
+      isWorkspaceFile,
+      done: true,
+    },
+  };
 }
 
 export const NotionHtmlTransformer = {
-  importNotionZip,
+  planNotionHtmlZip,
 };

blocksuite/affine/widgets/linked-doc/src/transformers/obsidian.ts

@@ -20,7 +20,11 @@ import { extMimeMap, nanoid } from '@blocksuite/store';
 import type { Html, Text } from 'mdast';
 
 import {
-  applyMetaPatch,
+  blobsFromAssets,
+  type ImportBatch,
+  type ImportDoc,
+} from './import-batch.js';
+import {
   bindImportedAssetsToJob,
   createMarkdownImportJob,
   getProvider,
@@ -129,8 +133,8 @@ function preprocessTitleHeader(markdown: string): string {
 
 function preprocessObsidianCallouts(markdown: string): string {
   return markdown.replace(
-    /^(> *)\[!([^\]\n]+)\]([+-]?)([^\n]*)/gm,
-    (_, prefix, type, _fold, rest) => {
+    /^(> *)\[!([^\]\n]+)\](?:[+-]?)([^\n]*)/gm,
+    (_, prefix, type, rest) => {
       const calloutToken =
         CALLOUT_TYPE_MAP[type.trim().toLowerCase()] ?? DEFAULT_CALLOUT_EMOJI;
       const title = rest.trim();
@@ -688,12 +692,16 @@ export type ImportObsidianVaultResult = {
   docEmojis: Map<string, string>;
 };
 
-export async function importObsidianVault({
+export type PlanObsidianVaultResult = ImportObsidianVaultResult & {
+  batch: ImportBatch;
+};
+
+export async function planObsidianVault({
   collection,
   schema,
   importedFiles,
   extensions,
-}: ImportObsidianVaultOptions): Promise<ImportObsidianVaultResult> {
+}: ImportObsidianVaultOptions): Promise<PlanObsidianVaultResult> {
   const provider = getProvider([
     obsidianWikilinkToDeltaMatcher,
     obsidianAttachmentEmbedMarkdownAdapterMatcher,
@@ -701,6 +709,7 @@ export async function importObsidianVault({
   ]);
 
   const docIds: string[] = [];
+  const docs: ImportDoc[] = [];
   const docEmojis = new Map<string, string>();
   const pendingAssets: AssetMap = new Map();
   const pendingPathBlobIdMap: PathBlobIdMap = new Map();
@@ -813,22 +822,31 @@ export async function importObsidianVault({
 
       if (snapshot) {
         snapshot.meta.id = predefinedId;
-        const doc = await job.snapshotToDoc(snapshot);
-        if (doc) {
-          applyMetaPatch(collection, doc.id, {
-            ...meta,
-            title: preferredTitle,
-            trash: false,
-          });
-          docIds.push(doc.id);
-        }
+        docs.push({
+          id: predefinedId,
+          snapshot,
+          meta: { ...meta, title: preferredTitle, trash: false },
+        });
+        docIds.push(predefinedId);
       }
     })
   );
 
-  return { docIds, docEmojis };
+  return {
+    docIds,
+    docEmojis,
+    batch: {
+      docs,
+      blobs: await blobsFromAssets(pendingAssets, pendingPathBlobIdMap),
+      icons: Array.from(docEmojis, ([docId, emoji]) => ({
+        docId,
+        icon: { type: 'emoji', unicode: emoji },
+      })),
+      done: true,
+    },
+  };
 }
 
 export const ObsidianTransformer = {
-  importObsidianVault,
+  planObsidianVault,
 };