feat(export): add metadata file to preserve page icons and ordering on import (#1877)

* feat(export): add metadata file to preserve page icons and ordering on import
- Export includes `docmost-metadata.json`
- Import reads metadata to restore icons and sort siblings by original position

* cleanup

* bonus fixes

* handle unknown prosemirror nodes

* add docmost app  version

apps/server/src/integrations/export/export.service.ts

@@ -20,11 +20,17 @@ import {
   replaceInternalLinks,
   updateAttachmentUrlsToLocalPaths,
 } from './utils';
+import {
+  ExportMetadata,
+  ExportPageMetadata,
+} from '../../common/helpers/types/export-metadata.types';
 import { PageRepo } from '@docmost/db/repos/page/page.repo';
 import { Node } from '@tiptap/pm/model';
 import { EditorState } from '@tiptap/pm/state';
 // eslint-disable-next-line @typescript-eslint/no-require-imports
 import slugify = require('@sindresorhus/slugify');
+// eslint-disable-next-line @typescript-eslint/no-require-imports
+const packageJson = require('../../../package.json');
 import { EnvironmentService } from '../environment/environment.service';
 import {
   getAttachmentIds,
@@ -155,12 +161,15 @@ export class ExportService {
         'pages.id',
         'pages.slugId',
         'pages.title',
+        'pages.icon',
+        'pages.position',
         'pages.content',
         'pages.parentPageId',
         'pages.spaceId',
         'pages.workspaceId',
       ])
       .where('spaceId', '=', spaceId)
+      .where('deletedAt', 'is', null)
       .execute();
 
     const tree = buildTree(pages as Page[]);
@@ -189,10 +198,12 @@ export class ExportService {
     includeAttachments: boolean,
   ): Promise<void> {
     const slugIdToPath: Record<string, string> = {};
+    const pageIdToFilePath: Record<string, string> = {};
+    const pagesMetadata: Record<string, ExportPageMetadata> = {};
 
     computeLocalPath(tree, format, null, '', slugIdToPath);
 
-    const stack: { folder: JSZip; parentPageId: string }[] = [
+    const stack: { folder: JSZip; parentPageId: string | null }[] = [
       { folder: zip, parentPageId: null },
     ];
 
@@ -232,12 +243,33 @@ export class ExportService {
           `${pageTitle}${getExportExtension(format)}`,
           pageExportContent,
         );
+
+        pageIdToFilePath[page.id] = currentPagePath;
+
+        const parentPath = parentPageId ? pageIdToFilePath[parentPageId] : null;
+        pagesMetadata[currentPagePath] = {
+          pageId: page.id,
+          slugId: page.slugId,
+          icon: page.icon ?? null,
+          position: page.position,
+          parentPath,
+        };
+
         if (childPages.length > 0) {
           const pageFolder = folder.folder(pageTitle);
           stack.push({ folder: pageFolder, parentPageId: page.id });
         }
       }
     }
+
+    const metadata: ExportMetadata = {
+      exportedAt: new Date().toISOString(),
+      source: 'docmost',
+      version: packageJson.version,
+      pages: pagesMetadata,
+    };
+
+    zip.file('docmost-metadata.json', JSON.stringify(metadata, null, 2));
   }
 
   async zipAttachments(prosemirrorJson: any, spaceId: string, zip: JSZip) {

apps/server/src/integrations/import/dto/file-task-dto.ts

@@ -15,4 +15,5 @@ export type ImportPageNode = {
   parentPageId: string | null;
   fileExtension: string;
   filePath: string;
+  icon?: string | null;
 };

apps/server/src/integrations/import/services/file-import-task.service.ts

@@ -24,6 +24,8 @@ import { formatImportHtml } from '../utils/import-formatter';
 import {
   buildAttachmentCandidates,
   collectMarkdownAndHtmlFiles,
+  encodeFilePath,
+  readDocmostMetadata,
   stripNotionID,
 } from '../utils/import.utils';
 import { executeTx } from '@docmost/db/utils';
@@ -154,6 +156,7 @@ export class FileImportTaskService {
     const { extractDir, fileTask } = opts;
     const allFiles = await collectMarkdownAndHtmlFiles(extractDir);
     const attachmentCandidates = await buildAttachmentCandidates(extractDir);
+    const docmostMetadata = await readDocmostMetadata(extractDir);
 
     const pagesMap = new Map<string, ImportPageNode>();
 
@@ -164,6 +167,9 @@ export class FileImportTaskService {
         .join('/'); // normalize to forward-slashes
       const ext = path.extname(relPath).toLowerCase();
 
+      const encodedPath = encodeFilePath(relPath);
+      const pageMetadata = docmostMetadata?.pages[encodedPath];
+
       pagesMap.set(relPath, {
         id: v7(),
         slugId: generateSlugId(),
@@ -172,6 +178,7 @@ export class FileImportTaskService {
         parentPageId: null,
         fileExtension: ext,
         filePath: relPath,
+        icon: pageMetadata?.icon ?? null,
       });
     }
 
@@ -224,6 +231,8 @@ export class FileImportTaskService {
 
       if (!pagesMap.has(mdPath) && !pagesMap.has(htmlPath)) {
         const folderName = path.basename(folderPath);
+        const encodedMdPath = encodeFilePath(mdPath);
+        const placeholderMetadata = docmostMetadata?.pages[encodedMdPath];
         pagesMap.set(mdPath, {
           id: v7(),
           slugId: generateSlugId(),
@@ -232,6 +241,7 @@ export class FileImportTaskService {
           parentPageId: null,
           fileExtension: '.md',
           filePath: mdPath,
+          icon: placeholderMetadata?.icon ?? null,
         });
       }
     });
@@ -266,11 +276,39 @@ export class FileImportTaskService {
       siblingsMap.set(page.parentPageId, group);
     });
 
+    const encodedPathsMap = new Map<string, string>();
+    if (docmostMetadata) {
+      pagesMap.forEach((_, filePath) => {
+        encodedPathsMap.set(filePath, encodeFilePath(filePath));
+      });
+    }
+
+    // Sort siblings by metadata position if available, otherwise alphabetically
+    const sortSiblings = (siblings: ImportPageNode[]) => {
+      if (docmostMetadata) {
+        siblings.sort((a, b) => {
+          const posA =
+            docmostMetadata.pages[encodedPathsMap.get(a.filePath)]?.position;
+          const posB =
+            docmostMetadata.pages[encodedPathsMap.get(b.filePath)]?.position;
+          if (posA && posB) {
+            // Use direct comparison to match PostgreSQL collation 'C' (byte order)
+            if (posA < posB) return -1;
+            if (posA > posB) return 1;
+            return 0;
+          }
+          return a.name.localeCompare(b.name);
+        });
+      } else {
+        siblings.sort((a, b) => a.name.localeCompare(b.name));
+      }
+    };
+
     // get root pages
     const rootSibs = siblingsMap.get(null);
 
     if (rootSibs?.length) {
-      rootSibs.sort((a, b) => a.name.localeCompare(b.name));
+      sortSiblings(rootSibs);
 
       // get first position key from the server
       const nextPosition = await this.pageService.nextPagePosition(
@@ -292,7 +330,7 @@ export class FileImportTaskService {
     siblingsMap.forEach((sibs, parentId) => {
       if (parentId === null) return; // root already done
 
-      sibs.sort((a, b) => a.name.localeCompare(b.name));
+      sortSiblings(sibs);
 
       let prevPos: string | null = null;
       for (const page of sibs) {
@@ -426,7 +464,7 @@ export class FileImportTaskService {
               id: page.id,
               slugId: page.slugId,
               title: title || page.name,
-              icon: pageIcon || null,
+              icon: page.icon || pageIcon || null,
               content: prosemirrorJson,
               textContent: jsonToText(prosemirrorJson),
               ydoc: await this.importService.createYdoc(prosemirrorJson),

apps/server/src/integrations/import/utils/import.utils.ts

@@ -1,6 +1,7 @@
 import { Logger } from '@nestjs/common';
 import { promises as fs } from 'fs';
 import * as path from 'path';
+import { ExportMetadata } from '../../../common/helpers/types/export-metadata.types';
 
 export async function buildAttachmentCandidates(
   extractDir: string,
@@ -35,9 +36,15 @@ export function resolveRelativeAttachmentPath(
   try {
     mainRel = decodeURIComponent(mainRel);
   } catch (err) {
-    Logger.warn(`URI malformed for attachment path: ${mainRel}. Falling back to raw path.`, 'ImportUtils');
+    Logger.warn(
+      `URI malformed for attachment path: ${mainRel}. Falling back to raw path.`,
+      'ImportUtils',
+    );
   }
-  const fallback = path.normalize(path.join(pageDir, mainRel)).split(path.sep).join('/');
+  const fallback = path
+    .normalize(path.join(pageDir, mainRel))
+    .split(path.sep)
+    .join('/');
 
   if (attachmentCandidates.has(mainRel)) {
     return mainRel;
@@ -76,3 +83,26 @@ export function stripNotionID(fileName: string): string {
   const notionIdPattern = /[ -]?[a-z0-9]{32}$/i;
   return fileName.replace(notionIdPattern, '').trim();
 }
+
+export function encodeFilePath(filePath: string): string {
+  return filePath
+    .split('/')
+    .map((segment) => encodeURIComponent(segment))
+    .join('/');
+}
+
+export async function readDocmostMetadata(
+  extractDir: string,
+): Promise<ExportMetadata | null> {
+  const metadataPath = path.join(extractDir, 'docmost-metadata.json');
+  try {
+    const content = await fs.readFile(metadataPath, 'utf-8');
+    const metadata = JSON.parse(content) as ExportMetadata;
+    if (metadata.source === 'docmost' && metadata.pages) {
+      return metadata;
+    }
+    return null;
+  } catch {
+    return null;
+  }
+}