feat(editor): support markdown adapter preprocessed with latex delimiters (#11431)

To close [BS-2870](https://linear.app/affine-design/issue/BS-2870/支持识别-和-[-包裹内容为公式)

## Add Markdown Preprocessor Extension and Enhanced LaTeX Support

### Markdown Preprocessor Extension
This PR introduces a new preprocessor extension for Markdown adapters that allows preprocessing of content before conversion:

Adds a new PreprocessorManager for handling text transformations
Introduces extensible preprocessor interface that supports different processing levels (block/slice/doc)

Integrates preprocessor extension into the existing Markdown adapter workflow

### LaTeX Support Enhancement
Extends LaTeX support to handle both traditional and alternative syntax:
Adds support for backslash LaTeX syntax:

Block math: ```\[...\] ``` alongside existing ```$$...$$```
Inline math: ```\(...\) ``` alongside existing ```$...$```

Implements LaTeX preprocessor to standardize syntax before conversion

Updates tests to cover both syntax variants

blocksuite/affine/all/src/__tests__/adapters/markdown.unit.spec.ts

@@ -3680,88 +3680,95 @@ bbb
     expect(nanoidReplacement(rawBlockSnapshot)).toEqual(blockSnapshot);
   });
 
-  test('inline latex', async () => {
-    const markdown = 'inline $E=mc^2$ latex\n';
-    const blockSnapshot: BlockSnapshot = {
-      type: 'block',
-      id: 'matchesReplaceMap[0]',
-      flavour: 'affine:note',
-      props: {
-        xywh: '[0,0,800,95]',
-        background: DefaultTheme.noteBackgrounColor,
-        index: 'a0',
-        hidden: false,
-        displayMode: NoteDisplayMode.DocAndEdgeless,
-      },
-      children: [
-        {
-          type: 'block',
-          id: 'matchesReplaceMap[1]',
-          flavour: 'affine:paragraph',
-          props: {
-            type: 'text',
-            text: {
-              '$blocksuite:internal:text$': true,
-              delta: [
-                {
-                  insert: 'inline ',
-                },
-                {
-                  insert: ' ',
-                  attributes: {
-                    latex: 'E=mc^2',
-                  },
-                },
-                {
-                  insert: ' latex',
-                },
-              ],
-            },
-          },
-          children: [],
+  describe('inline latex', () => {
+    test.each([
+      ['dollar sign syntax', 'inline $E=mc^2$ latex\n'],
+      ['backslash syntax', 'inline \\(E=mc^2\\) latex\n'],
+    ])('should convert %s correctly', async (_, markdown) => {
+      const blockSnapshot: BlockSnapshot = {
+        type: 'block',
+        id: 'matchesReplaceMap[0]',
+        flavour: 'affine:note',
+        props: {
+          xywh: '[0,0,800,95]',
+          background: DefaultTheme.noteBackgrounColor,
+          index: 'a0',
+          hidden: false,
+          displayMode: NoteDisplayMode.DocAndEdgeless,
         },
-      ],
-    };
+        children: [
+          {
+            type: 'block',
+            id: 'matchesReplaceMap[1]',
+            flavour: 'affine:paragraph',
+            props: {
+              type: 'text',
+              text: {
+                '$blocksuite:internal:text$': true,
+                delta: [
+                  {
+                    insert: 'inline ',
+                  },
+                  {
+                    insert: ' ',
+                    attributes: {
+                      latex: 'E=mc^2',
+                    },
+                  },
+                  {
+                    insert: ' latex',
+                  },
+                ],
+              },
+            },
+            children: [],
+          },
+        ],
+      };
 
-    const mdAdapter = new MarkdownAdapter(createJob(), provider);
-    const rawBlockSnapshot = await mdAdapter.toBlockSnapshot({
-      file: markdown,
+      const mdAdapter = new MarkdownAdapter(createJob(), provider);
+      const rawBlockSnapshot = await mdAdapter.toBlockSnapshot({
+        file: markdown,
+      });
+      expect(nanoidReplacement(rawBlockSnapshot)).toEqual(blockSnapshot);
     });
-    expect(nanoidReplacement(rawBlockSnapshot)).toEqual(blockSnapshot);
   });
 
-  test('latex block', async () => {
-    const markdown = '$$\nE=mc^2\n$$\n';
-
-    const blockSnapshot: BlockSnapshot = {
-      type: 'block',
-      id: 'matchesReplaceMap[0]',
-      flavour: 'affine:note',
-      props: {
-        xywh: '[0,0,800,95]',
-        background: DefaultTheme.noteBackgrounColor,
-        index: 'a0',
-        hidden: false,
-        displayMode: NoteDisplayMode.DocAndEdgeless,
-      },
-      children: [
-        {
-          type: 'block',
-          id: 'matchesReplaceMap[1]',
-          flavour: 'affine:latex',
-          props: {
-            latex: 'E=mc^2',
-          },
-          children: [],
+  describe('latex block', () => {
+    test.each([
+      ['dollar sign syntax', '$$\nE=mc^2\n$$\n'],
+      ['backslash syntax', '\\[\nE=mc^2\n\\]\n'],
+    ])('should convert %s correctly', async (_, markdown) => {
+      const blockSnapshot: BlockSnapshot = {
+        type: 'block',
+        id: 'matchesReplaceMap[0]',
+        flavour: 'affine:note',
+        props: {
+          xywh: '[0,0,800,95]',
+          background: DefaultTheme.noteBackgrounColor,
+          index: 'a0',
+          hidden: false,
+          displayMode: NoteDisplayMode.DocAndEdgeless,
         },
-      ],
-    };
+        children: [
+          {
+            type: 'block',
+            id: 'matchesReplaceMap[1]',
+            flavour: 'affine:latex',
+            props: {
+              latex: 'E=mc^2',
+            },
+            children: [],
+          },
+        ],
+      };
 
-    const mdAdapter = new MarkdownAdapter(createJob(), provider);
-    const rawBlockSnapshot = await mdAdapter.toBlockSnapshot({
-      file: markdown,
+      const mdAdapter = new MarkdownAdapter(createJob(), provider);
+      const rawBlockSnapshot = await mdAdapter.toBlockSnapshot({
+        file: markdown,
+      });
+      expect(nanoidReplacement(rawBlockSnapshot)).toEqual(blockSnapshot);
     });
-    expect(nanoidReplacement(rawBlockSnapshot)).toEqual(blockSnapshot);
   });
 
   test('reference', async () => {

blocksuite/affine/all/src/adapters/extension.ts

@@ -20,6 +20,7 @@ import type { ExtensionType } from '@blocksuite/store';
 
 import { defaultBlockHtmlAdapterMatchers } from './html/block-matcher';
 import { defaultBlockMarkdownAdapterMatchers } from './markdown/block-matcher';
+import { defaultMarkdownPreprocessors } from './markdown/preprocessor';
 import { defaultBlockNotionHtmlAdapterMatchers } from './notion-html/block-matcher';
 import { defaultBlockPlainTextAdapterMatchers } from './plain-text/block-matcher';
 
@@ -44,6 +45,7 @@ export const MarkdownAdapterExtension: ExtensionType[] = [
   ...MarkdownInlineToDeltaAdapterExtensions,
   ...defaultBlockMarkdownAdapterMatchers,
   ...InlineDeltaToMarkdownAdapterExtensions,
+  ...defaultMarkdownPreprocessors,
 ];
 
 export const NotionHtmlAdapterExtension: ExtensionType[] = [

blocksuite/affine/all/src/adapters/index.ts

@@ -1,5 +1,6 @@
 export * from './extension.js';
 export * from './html/block-matcher.js';
 export * from './markdown/block-matcher.js';
+export * from './markdown/preprocessor.js';
 export * from './notion-html/block-matcher.js';
 export * from './plain-text/block-matcher.js';

blocksuite/affine/all/src/adapters/markdown/preprocessor.ts

@@ -0,0 +1,7 @@
+import { CodeMarkdownPreprocessorExtension } from '@blocksuite/affine-block-code';
+import { LatexMarkdownPreprocessorExtension } from '@blocksuite/affine-block-latex';
+
+export const defaultMarkdownPreprocessors = [
+  LatexMarkdownPreprocessorExtension,
+  CodeMarkdownPreprocessorExtension,
+];

blocksuite/affine/blocks/block-code/src/adapters/extension.ts

@@ -1,13 +1,13 @@
 import type { ExtensionType } from '@blocksuite/store';
 
 import { CodeBlockHtmlAdapterExtension } from './html.js';
-import { CodeBlockMarkdownAdapterExtension } from './markdown.js';
+import { CodeBlockMarkdownAdapterExtensions } from './markdown/index.js';
 import { CodeBlockNotionHtmlAdapterExtension } from './notion-html.js';
 import { CodeBlockPlainTextAdapterExtension } from './plain-text.js';
 
 export const CodeBlockAdapterExtensions: ExtensionType[] = [
   CodeBlockHtmlAdapterExtension,
-  CodeBlockMarkdownAdapterExtension,
+  CodeBlockMarkdownAdapterExtensions,
   CodeBlockPlainTextAdapterExtension,
   CodeBlockNotionHtmlAdapterExtension,
-];
+].flat();

blocksuite/affine/blocks/block-code/src/adapters/index.ts

@@ -1,4 +1,4 @@
 export * from './html.js';
-export * from './markdown.js';
+export * from './markdown/index.js';
 export * from './notion-html.js';
 export * from './plain-text.js';

blocksuite/affine/blocks/block-code/src/adapters/markdown/index.ts

@@ -0,0 +1,12 @@
+import type { ExtensionType } from '@blocksuite/store';
+
+import { CodeBlockMarkdownAdapterExtension } from './markdown.js';
+import { CodeMarkdownPreprocessorExtension } from './preprocessor.js';
+
+export * from './markdown.js';
+export * from './preprocessor.js';
+
+export const CodeBlockMarkdownAdapterExtensions: ExtensionType[] = [
+  CodeMarkdownPreprocessorExtension,
+  CodeBlockMarkdownAdapterExtension,
+];

blocksuite/affine/blocks/block-code/src/adapters/markdown/preprocessor.ts

@@ -0,0 +1,76 @@
+import {
+  type MarkdownAdapterPreprocessor,
+  MarkdownPreprocessorExtension,
+} from '@blocksuite/affine-shared/adapters';
+
+const codePreprocessor: MarkdownAdapterPreprocessor = {
+  name: 'code',
+  levels: ['slice'],
+  preprocess: content => {
+    let codeFence = '';
+    const lines = content
+      .split('\n')
+      .map(line => {
+        if (line.trimStart().startsWith('-')) {
+          return line;
+        }
+        let trimmedLine = line.trimStart();
+        if (!codeFence && trimmedLine.startsWith('```')) {
+          codeFence = trimmedLine.substring(
+            0,
+            trimmedLine.lastIndexOf('```') + 3
+          );
+          if (codeFence.split('').every(c => c === '`')) {
+            return line;
+          }
+          codeFence = '';
+        }
+        if (!codeFence && trimmedLine.startsWith('~~~')) {
+          codeFence = trimmedLine.substring(
+            0,
+            trimmedLine.lastIndexOf('~~~') + 3
+          );
+          if (codeFence.split('').every(c => c === '~')) {
+            return line;
+          }
+          codeFence = '';
+        }
+        if (
+          !!codeFence &&
+          trimmedLine.startsWith(codeFence) &&
+          trimmedLine.lastIndexOf(codeFence) === 0
+        ) {
+          codeFence = '';
+        }
+        if (codeFence) {
+          return line;
+        }
+
+        trimmedLine = trimmedLine.trimEnd();
+        if (!trimmedLine.startsWith('<') && !trimmedLine.endsWith('>')) {
+          // check if it is a url link and wrap it with the angle brackets
+          // sometimes the url includes emphasis `_` that will break URL parsing
+          //
+          // eg. /MuawcBMT1Mzvoar09-_66?mode=page&blockIds=rL2_GXbtLU2SsJVfCSmh_
+          // https://www.markdownguide.org/basic-syntax/#urls-and-email-addresses
+          try {
+            const valid =
+              URL.canParse?.(trimmedLine) ?? Boolean(new URL(trimmedLine));
+            if (valid) {
+              return `<${trimmedLine}>`;
+            }
+          } catch (err) {
+            console.log(err);
+          }
+        }
+
+        return line.replace(/^ /, '&#x20;');
+      })
+      .join('\n');
+
+    return lines;
+  },
+};
+
+export const CodeMarkdownPreprocessorExtension =
+  MarkdownPreprocessorExtension(codePreprocessor);

blocksuite/affine/blocks/block-latex/src/adapters/extension.ts

@@ -1,11 +1,11 @@
 import type { ExtensionType } from '@blocksuite/store';
 
-import { LatexBlockMarkdownAdapterExtension } from './markdown.js';
+import { LatexMarkdownAdapterExtensions } from './markdown/index.js';
 import { LatexBlockNotionHtmlAdapterExtension } from './notion-html.js';
 import { LatexBlockPlainTextAdapterExtension } from './plain-text.js';
 
 export const LatexBlockAdapterExtensions: ExtensionType[] = [
-  LatexBlockMarkdownAdapterExtension,
+  LatexMarkdownAdapterExtensions,
   LatexBlockNotionHtmlAdapterExtension,
   LatexBlockPlainTextAdapterExtension,
-];
+].flat();

blocksuite/affine/blocks/block-latex/src/adapters/index.ts

@@ -1,3 +1,3 @@
-export * from './markdown.js';
+export * from './markdown/index.js';
 export * from './notion-html.js';
 export * from './plain-text.js';

blocksuite/affine/blocks/block-latex/src/adapters/markdown/index.ts

@@ -0,0 +1,12 @@
+import type { ExtensionType } from '@blocksuite/store';
+
+import { LatexBlockMarkdownAdapterExtension } from './markdown.js';
+import { LatexMarkdownPreprocessorExtension } from './preprocessor.js';
+
+export * from './markdown.js';
+export * from './preprocessor.js';
+
+export const LatexMarkdownAdapterExtensions: ExtensionType[] = [
+  LatexMarkdownPreprocessorExtension,
+  LatexBlockMarkdownAdapterExtension,
+];

blocksuite/affine/blocks/block-latex/src/adapters/markdown/preprocessor.ts

@@ -0,0 +1,25 @@
+import {
+  type MarkdownAdapterPreprocessor,
+  MarkdownPreprocessorExtension,
+} from '@blocksuite/affine-shared/adapters';
+
+const latexPreprocessor: MarkdownAdapterPreprocessor = {
+  name: 'latex',
+  levels: ['block', 'slice', 'doc'],
+  preprocess: content => {
+    // Replace block-level LaTeX delimiters \[ \] with $$ $$
+    const blockProcessedContent = content.replace(
+      /\\\[(.*?)\\\]/gs,
+      (_, equation) => `$$${equation}$$`
+    );
+    // Replace inline LaTeX delimiters \( \) with $ $
+    const inlineProcessedContent = blockProcessedContent.replace(
+      /\\\((.*?)\\\)/gs,
+      (_, equation) => `$${equation}$`
+    );
+    return inlineProcessedContent;
+  },
+};
+
+export const LatexMarkdownPreprocessorExtension =
+  MarkdownPreprocessorExtension(latexPreprocessor);

blocksuite/affine/shared/src/adapters/html/delta-converter.ts

@@ -11,7 +11,7 @@ import {
   DeltaASTConverter,
   type DeltaASTConverterOptions,
   type InlineDeltaMatcher,
-} from '../types/adapter.js';
+} from '../types/delta-converter.js';
 import type { HtmlAST, InlineHtmlAST } from '../types/hast.js';
 import { AdapterTextUtils } from '../utils/text.js';
 

blocksuite/affine/shared/src/adapters/index.ts

@@ -30,11 +30,14 @@ export {
   MarkdownAdapter,
   MarkdownAdapterFactoryExtension,
   MarkdownAdapterFactoryIdentifier,
+  type MarkdownAdapterPreprocessor,
   type MarkdownAST,
   MarkdownASTToDeltaExtension,
   type MarkdownASTToDeltaMatcher,
   MarkdownASTToDeltaMatcherIdentifier,
   MarkdownDeltaConverter,
+  MarkdownPreprocessorExtension,
+  MarkdownPreprocessorManager,
 } from './markdown';
 export * from './middlewares';
 export * from './mix-text';

blocksuite/affine/shared/src/adapters/markdown/delta-converter.ts

@@ -10,7 +10,7 @@ import {
   type ASTToDeltaMatcher,
   DeltaASTConverter,
   type InlineDeltaMatcher,
-} from '../types/adapter.js';
+} from '../types/delta-converter.js';
 import type { MarkdownAST } from './type.js';
 
 export type InlineDeltaToMarkdownAdapterMatcher =

blocksuite/affine/shared/src/adapters/markdown/index.ts

@@ -1,4 +1,5 @@
 export * from './block-adapter.js';
 export * from './delta-converter.js';
 export * from './markdown.js';
+export * from './preprocessor.js';
 export * from './type.js';

blocksuite/affine/shared/src/adapters/markdown/markdown.ts

@@ -37,6 +37,7 @@ import {
   MarkdownDeltaConverter,
 } from './delta-converter';
 import { remarkGfm } from './gfm';
+import { MarkdownPreprocessorManager } from './preprocessor';
 import type { Markdown, MarkdownAST } from './type';
 
 type MarkdownToSliceSnapshotPayload = {
@@ -167,6 +168,7 @@ export class MarkdownAdapter extends BaseAdapter<Markdown> {
   };
 
   deltaConverter: MarkdownDeltaConverter;
+  preprocessorManager: MarkdownPreprocessorManager;
 
   readonly blockMatchers: BlockMarkdownAdapterMatcher[];
 
@@ -187,6 +189,7 @@ export class MarkdownAdapter extends BaseAdapter<Markdown> {
       inlineDeltaToMarkdownAdapterMatchers,
       markdownInlineToDeltaMatchers
     );
+    this.preprocessorManager = new MarkdownPreprocessorManager(provider);
   }
 
   private _astToMarkdown(ast: Root) {
@@ -273,7 +276,11 @@ export class MarkdownAdapter extends BaseAdapter<Markdown> {
   async toBlockSnapshot(
     payload: ToBlockSnapshotPayload<Markdown>
   ): Promise<BlockSnapshot> {
-    const markdownAst = this._markdownToAst(payload.file);
+    const markdownFile = this.preprocessorManager.process(
+      'block',
+      payload.file
+    );
+    const markdownAst = this._markdownToAst(markdownFile);
     const blockSnapshotRoot = {
       type: 'block',
       id: nanoid(),
@@ -297,7 +304,8 @@ export class MarkdownAdapter extends BaseAdapter<Markdown> {
   async toDocSnapshot(
     payload: ToDocSnapshotPayload<Markdown>
   ): Promise<DocSnapshot> {
-    const markdownAst = this._markdownToAst(payload.file);
+    const markdownFile = this.preprocessorManager.process('doc', payload.file);
+    const markdownAst = this._markdownToAst(markdownFile);
     const blockSnapshotRoot = {
       type: 'block',
       id: nanoid(),
@@ -356,67 +364,11 @@ export class MarkdownAdapter extends BaseAdapter<Markdown> {
   async toSliceSnapshot(
     payload: MarkdownToSliceSnapshotPayload
   ): Promise<SliceSnapshot | null> {
-    let codeFence = '';
-    payload.file = payload.file
-      .split('\n')
-      .map(line => {
-        if (line.trimStart().startsWith('-')) {
-          return line;
-        }
-        let trimmedLine = line.trimStart();
-        if (!codeFence && trimmedLine.startsWith('```')) {
-          codeFence = trimmedLine.substring(
-            0,
-            trimmedLine.lastIndexOf('```') + 3
-          );
-          if (codeFence.split('').every(c => c === '`')) {
-            return line;
-          }
-          codeFence = '';
-        }
-        if (!codeFence && trimmedLine.startsWith('~~~')) {
-          codeFence = trimmedLine.substring(
-            0,
-            trimmedLine.lastIndexOf('~~~') + 3
-          );
-          if (codeFence.split('').every(c => c === '~')) {
-            return line;
-          }
-          codeFence = '';
-        }
-        if (
-          !!codeFence &&
-          trimmedLine.startsWith(codeFence) &&
-          trimmedLine.lastIndexOf(codeFence) === 0
-        ) {
-          codeFence = '';
-        }
-        if (codeFence) {
-          return line;
-        }
-
-        trimmedLine = trimmedLine.trimEnd();
-        if (!trimmedLine.startsWith('<') && !trimmedLine.endsWith('>')) {
-          // check if it is a url link and wrap it with the angle brackets
-          // sometimes the url includes emphasis `_` that will break URL parsing
-          //
-          // eg. /MuawcBMT1Mzvoar09-_66?mode=page&blockIds=rL2_GXbtLU2SsJVfCSmh_
-          // https://www.markdownguide.org/basic-syntax/#urls-and-email-addresses
-          try {
-            const valid =
-              URL.canParse?.(trimmedLine) ?? Boolean(new URL(trimmedLine));
-            if (valid) {
-              return `<${trimmedLine}>`;
-            }
-          } catch (err) {
-            console.log(err);
-          }
-        }
-
-        return line.replace(/^ /, '&#x20;');
-      })
-      .join('\n');
-    const markdownAst = this._markdownToAst(payload.file);
+    const markdownFile = this.preprocessorManager.process(
+      'slice',
+      payload.file
+    );
+    const markdownAst = this._markdownToAst(markdownFile);
     const blockSnapshotRoot = {
       type: 'block',
       id: nanoid(),

blocksuite/affine/shared/src/adapters/markdown/preprocessor.ts

@@ -0,0 +1,40 @@
+import {
+  createIdentifier,
+  type ServiceIdentifier,
+  type ServiceProvider,
+} from '@blocksuite/global/di';
+import type { ExtensionType } from '@blocksuite/store';
+
+import {
+  type AdapterPreprocessor,
+  PreprocessorManager,
+} from '../types/preprocessor';
+import type { Markdown } from './type';
+
+export type MarkdownAdapterPreprocessor = AdapterPreprocessor<Markdown>;
+
+const MarkdownPreprocessorIdentifier =
+  createIdentifier<MarkdownAdapterPreprocessor>('MarkdownPreprocessor');
+
+export const MarkdownPreprocessorExtension = (
+  preprocessor: MarkdownAdapterPreprocessor
+): ExtensionType & {
+  identifier: ServiceIdentifier<MarkdownAdapterPreprocessor>;
+} => {
+  const identifier = MarkdownPreprocessorIdentifier(preprocessor.name);
+  return {
+    setup: di => {
+      di.addImpl(identifier, () => preprocessor);
+    },
+    identifier,
+  };
+};
+
+export class MarkdownPreprocessorManager extends PreprocessorManager<
+  Markdown,
+  MarkdownAdapterPreprocessor
+> {
+  constructor(provider: ServiceProvider) {
+    super(provider, MarkdownPreprocessorIdentifier);
+  }
+}

blocksuite/affine/shared/src/adapters/notion-html/delta-converter.ts

@@ -11,7 +11,7 @@ import {
   DeltaASTConverter,
   type DeltaASTConverterOptions,
   type InlineDeltaMatcher,
-} from '../types/adapter.js';
+} from '../types/delta-converter.js';
 import type { HtmlAST, InlineHtmlAST } from '../types/hast.js';
 
 export type InlineDeltaToNotionHtmlAdapterMatcher =

blocksuite/affine/shared/src/adapters/plain-text/delta-converter.ts

@@ -5,12 +5,12 @@ import {
 import type { DeltaInsert, ExtensionType } from '@blocksuite/store';
 
 import type { AffineTextAttributes } from '../../types/index.js';
+import type { TextBuffer } from '../types/adapter.js';
 import {
   type ASTToDeltaMatcher,
   DeltaASTConverter,
   type InlineDeltaMatcher,
-  type TextBuffer,
-} from '../types/adapter.js';
+} from '../types/delta-converter.js';
 
 export type InlineDeltaToPlainTextAdapterMatcher =
   InlineDeltaMatcher<TextBuffer>;

blocksuite/affine/shared/src/adapters/types/adapter.ts

@@ -4,15 +4,13 @@ import {
   type ASTWalker,
   type ASTWalkerContext,
   type BaseAdapter,
-  type BaseTextAttributes,
   type BlockSnapshot,
   BlockSnapshotSchema,
-  type DeltaInsert,
   type NodeProps,
   type Transformer,
 } from '@blocksuite/store';
 
-import type { AffineTextAttributes } from '../../types/index.js';
+import type { DeltaASTConverter } from './delta-converter.js';
 
 export const isBlockSnapshotNode = (node: unknown): node is BlockSnapshot =>
   BlockSnapshotSchema.safeParse(node).success;
@@ -21,13 +19,6 @@ export type TextBuffer = {
   content: string;
 };
 
-export type DeltaASTConverterOptions = {
-  trim?: boolean;
-  pre?: boolean;
-  pageMap?: Map<string, string>;
-  removeLastBr?: boolean;
-};
-
 export type AdapterContext<
   ONode extends object,
   TNode extends object = never,
@@ -124,56 +115,6 @@ export type BlockAdapterMatcher<
   };
 };
 
-export abstract class DeltaASTConverter<
-  TextAttributes extends BaseTextAttributes = BaseTextAttributes,
-  AST = unknown,
-> {
-  /**
-   * Convert AST format to delta format
-   */
-  abstract astToDelta(
-    ast: AST,
-    options?: unknown
-  ): DeltaInsert<TextAttributes>[];
-
-  /**
-   * Convert delta format to AST format
-   */
-  abstract deltaToAST(
-    deltas: DeltaInsert<TextAttributes>[],
-    options?: unknown
-  ): AST[];
-}
-
-export type InlineDeltaMatcher<TNode extends object = never> = {
-  name: keyof AffineTextAttributes | string;
-  match: (delta: DeltaInsert<AffineTextAttributes>) => boolean;
-  toAST: (
-    delta: DeltaInsert<AffineTextAttributes>,
-    context: {
-      configs: Map<string, string>;
-      current: TNode;
-    },
-    provider?: ServiceProvider
-  ) => TNode;
-};
-
-export type ASTToDeltaMatcher<AST> = {
-  name: string;
-  match: (ast: AST) => boolean;
-  toDelta: (
-    ast: AST,
-    context: {
-      configs: Map<string, string>;
-      options: DeltaASTConverterOptions;
-      toDelta: (
-        ast: AST,
-        options?: DeltaASTConverterOptions
-      ) => DeltaInsert<AffineTextAttributes>[];
-    }
-  ) => DeltaInsert<AffineTextAttributes>[];
-};
-
 export type AdapterFactory = {
   // TODO(@chen): Make it return the specific adapter type
   get: (job: Transformer) => BaseAdapter;

blocksuite/affine/shared/src/adapters/types/delta-converter.ts

@@ -0,0 +1,61 @@
+import type { ServiceProvider } from '@blocksuite/global/di';
+import type { BaseTextAttributes, DeltaInsert } from '@blocksuite/store';
+
+import type { AffineTextAttributes } from '../../types';
+
+export type DeltaASTConverterOptions = {
+  trim?: boolean;
+  pre?: boolean;
+  pageMap?: Map<string, string>;
+  removeLastBr?: boolean;
+};
+
+export abstract class DeltaASTConverter<
+  TextAttributes extends BaseTextAttributes = BaseTextAttributes,
+  AST = unknown,
+> {
+  /**
+   * Convert AST format to delta format
+   */
+  abstract astToDelta(
+    ast: AST,
+    options?: unknown
+  ): DeltaInsert<TextAttributes>[];
+
+  /**
+   * Convert delta format to AST format
+   */
+  abstract deltaToAST(
+    deltas: DeltaInsert<TextAttributes>[],
+    options?: unknown
+  ): AST[];
+}
+
+export type InlineDeltaMatcher<TNode extends object = never> = {
+  name: keyof AffineTextAttributes | string;
+  match: (delta: DeltaInsert<AffineTextAttributes>) => boolean;
+  toAST: (
+    delta: DeltaInsert<AffineTextAttributes>,
+    context: {
+      configs: Map<string, string>;
+      current: TNode;
+    },
+    provider?: ServiceProvider
+  ) => TNode;
+};
+
+export type ASTToDeltaMatcher<AST> = {
+  name: string;
+  match: (ast: AST) => boolean;
+  toDelta: (
+    ast: AST,
+    context: {
+      configs: Map<string, string>;
+      options: DeltaASTConverterOptions;
+      toDelta: (
+        ast: AST,
+        options?: DeltaASTConverterOptions
+      ) => DeltaInsert<AffineTextAttributes>[];
+    }
+  ) => DeltaInsert<AffineTextAttributes>[];
+};

blocksuite/affine/shared/src/adapters/types/index.ts

@@ -1,2 +1,4 @@
 export * from './adapter.js';
+export * from './delta-converter.js';
 export * from './hast.js';
+export * from './preprocessor.js';

blocksuite/affine/shared/src/adapters/types/preprocessor.ts

@@ -0,0 +1,88 @@
+import type { ServiceIdentifier, ServiceProvider } from '@blocksuite/global/di';
+
+/**
+ * Level of preprocessing
+ * - doc: Process at to doc snapshot level
+ * - slice: Process at to slice snapshot level
+ * - block: Process at to block snapshot level
+ */
+export type PreprocessLevel = 'doc' | 'slice' | 'block';
+
+/**
+ * Interface for adapter preprocessor
+ * @template T Type of content to process, defaults to string
+ */
+export type AdapterPreprocessor<T = string> = {
+  /**
+   * Unique name of the preprocessor
+   */
+  name: string;
+
+  /**
+   * Levels this preprocessor supports
+   */
+  levels: PreprocessLevel[];
+
+  /**
+   * Process the content
+   * @param content Content to process
+   * @returns Processed content
+   */
+  preprocess: (content: T) => T;
+};
+
+/**
+ * Manager class for handling preprocessors
+ * @template T Type of content to process
+ * @template P Type of preprocessor
+ */
+export abstract class PreprocessorManager<T, P extends AdapterPreprocessor<T>> {
+  protected readonly preprocessors: Map<PreprocessLevel, Set<P>>;
+
+  constructor(
+    protected readonly provider: ServiceProvider,
+    protected readonly identifier: ServiceIdentifier<P>
+  ) {
+    this.preprocessors = new Map();
+
+    // Initialize Sets for each level
+    this.preprocessors.set('doc', new Set());
+    this.preprocessors.set('slice', new Set());
+    this.preprocessors.set('block', new Set());
+
+    // Register all preprocessors from provider
+    this.initializePreprocessors();
+  }
+
+  /**
+   * Initialize preprocessors from provider
+   */
+  private initializePreprocessors(): void {
+    const preprocessors = Array.from(
+      this.provider.getAll(this.identifier).values()
+    );
+
+    for (const preprocessor of preprocessors) {
+      for (const level of preprocessor.levels) {
+        const levelSet = this.preprocessors.get(level);
+        if (levelSet) {
+          levelSet.add(preprocessor);
+        }
+      }
+    }
+  }
+
+  /**
+   * Pre process content at specified level
+   * @param level Level to process at
+   * @param content Content to process
+   * @returns Processed content
+   */
+  process(level: PreprocessLevel, content: T): T {
+    const processors = this.preprocessors.get(level) ?? new Set();
+    return Array.from(processors).reduce(
+      (result, preprocessor) => preprocessor.preprocess(result),
+      content
+    );
+  }
+}

packages/common/nbstore/src/sync/indexer/crawler.ts

@@ -1,9 +1,5 @@
-import { defaultBlockMarkdownAdapterMatchers } from '@blocksuite/affine/adapters';
+import { MarkdownAdapterExtension } from '@blocksuite/affine/adapters';
 import { Container } from '@blocksuite/affine/global/di';
-import {
-  InlineDeltaToMarkdownAdapterExtensions,
-  MarkdownInlineToDeltaAdapterExtensions,
-} from '@blocksuite/affine/inlines/preset';
 import type {
   AttachmentBlockModel,
   BookmarkBlockModel,
@@ -123,11 +119,7 @@ function generateMarkdownPreviewBuilder(
   };
 
   const container = new Container();
-  [
-    ...MarkdownInlineToDeltaAdapterExtensions,
-    ...defaultBlockMarkdownAdapterMatchers,
-    ...InlineDeltaToMarkdownAdapterExtensions,
-  ].forEach(ext => {
+  [...MarkdownAdapterExtension].forEach(ext => {
     ext.setup(container);
   });
 

packages/frontend/apps/android/src/app.tsx

@@ -26,12 +26,8 @@ import { configureBrowserWorkspaceFlavours } from '@affine/core/modules/workspac
 import { getWorkerUrl } from '@affine/env/worker';
 import { I18n } from '@affine/i18n';
 import { StoreManagerClient } from '@affine/nbstore/worker/client';
-import { defaultBlockMarkdownAdapterMatchers } from '@blocksuite/affine/adapters';
+import { MarkdownAdapterExtension } from '@blocksuite/affine/adapters';
 import { Container } from '@blocksuite/affine/global/di';
-import {
-  InlineDeltaToMarkdownAdapterExtensions,
-  MarkdownInlineToDeltaAdapterExtensions,
-} from '@blocksuite/affine/inlines/preset';
 import {
   docLinkBaseURLMiddleware,
   MarkdownAdapter,
@@ -203,11 +199,7 @@ framework.impl(AIButtonProvider, {
     const snapshot = transformer.docToSnapshot(blockSuiteDoc);
 
     const container = new Container();
-    [
-      ...MarkdownInlineToDeltaAdapterExtensions,
-      ...defaultBlockMarkdownAdapterMatchers,
-      ...InlineDeltaToMarkdownAdapterExtensions,
-    ].forEach(ext => {
+    [...MarkdownAdapterExtension].forEach(ext => {
       ext.setup(container);
     });
     const provider = container.provider();

packages/frontend/apps/ios/src/app.tsx

@@ -38,13 +38,9 @@ import { configureBrowserWorkspaceFlavours } from '@affine/core/modules/workspac
 import { getWorkerUrl } from '@affine/env/worker';
 import { I18n } from '@affine/i18n';
 import { StoreManagerClient } from '@affine/nbstore/worker/client';
-import { defaultBlockMarkdownAdapterMatchers } from '@blocksuite/affine/adapters';
+import { MarkdownAdapterExtension } from '@blocksuite/affine/adapters';
 import { MarkdownTransformer } from '@blocksuite/affine/blocks/root';
 import { Container } from '@blocksuite/affine/global/di';
-import {
-  InlineDeltaToMarkdownAdapterExtensions,
-  MarkdownInlineToDeltaAdapterExtensions,
-} from '@blocksuite/affine/inlines/preset';
 import {
   docLinkBaseURLMiddleware,
   MarkdownAdapter,
@@ -266,11 +262,7 @@ const frameworkProvider = framework.provider();
     const snapshot = transformer.docToSnapshot(blockSuiteDoc);
 
     const container = new Container();
-    [
-      ...MarkdownInlineToDeltaAdapterExtensions,
-      ...defaultBlockMarkdownAdapterMatchers,
-      ...InlineDeltaToMarkdownAdapterExtensions,
-    ].forEach(ext => {
+    [...MarkdownAdapterExtension].forEach(ext => {
       ext.setup(container);
     });
     const provider = container.provider();

packages/frontend/core/src/blocksuite/ai/components/text-renderer.ts

@@ -1,5 +1,5 @@
 import { createReactComponentFromLit } from '@affine/component';
-import { defaultBlockMarkdownAdapterMatchers } from '@blocksuite/affine/adapters';
+import { MarkdownAdapterExtension } from '@blocksuite/affine/adapters';
 import {
   defaultImageProxyMiddleware,
   ImageProxyService,
@@ -7,10 +7,6 @@ import {
 import { PageEditorBlockSpecs } from '@blocksuite/affine/extensions';
 import { Container, type ServiceProvider } from '@blocksuite/affine/global/di';
 import { WithDisposable } from '@blocksuite/affine/global/lit';
-import {
-  InlineDeltaToMarkdownAdapterExtensions,
-  MarkdownInlineToDeltaAdapterExtensions,
-} from '@blocksuite/affine/inlines/preset';
 import { codeBlockWrapMiddleware } from '@blocksuite/affine/shared/adapters';
 import { LinkPreviewerService } from '@blocksuite/affine/shared/services';
 import {
@@ -215,11 +211,7 @@ export class TextRenderer extends WithDisposable(ShadowlessElement) {
         provider = this.host.std.provider;
       } else {
         const container = new Container();
-        [
-          ...MarkdownInlineToDeltaAdapterExtensions,
-          ...defaultBlockMarkdownAdapterMatchers,
-          ...InlineDeltaToMarkdownAdapterExtensions,
-        ].forEach(ext => {
+        [...MarkdownAdapterExtension].forEach(ext => {
           ext.setup(container);
         });
 

packages/frontend/core/src/blocksuite/ai/mini-mindmap/__tests__/mindmap-preview.unit.spec.ts

@@ -1,20 +1,12 @@
-import { defaultBlockMarkdownAdapterMatchers } from '@blocksuite/affine/adapters';
+import { MarkdownAdapterExtension } from '@blocksuite/affine/adapters';
 import { Container } from '@blocksuite/affine/global/di';
-import {
-  InlineDeltaToMarkdownAdapterExtensions,
-  MarkdownInlineToDeltaAdapterExtensions,
-} from '@blocksuite/affine/inlines/preset';
 import { TestWorkspace } from '@blocksuite/affine/store/test';
 import { describe, expect, test } from 'vitest';
 
 import { markdownToMindmap } from '../mindmap-preview.js';
 
 const container = new Container();
-[
-  ...MarkdownInlineToDeltaAdapterExtensions,
-  ...defaultBlockMarkdownAdapterMatchers,
-  ...InlineDeltaToMarkdownAdapterExtensions,
-].forEach(ext => {
+[...MarkdownAdapterExtension].forEach(ext => {
   ext.setup(container);
 });
 const provider = container.provider();