feat(core): support Notion markdown zip imports (#14910)

## What changed
- Markdown zip imports now resolve local `.md` links like `./test/2.md`
into AFFiNE linked-page references when the target document exists in
the same archive.
- Added Notion Markdown `.zip` import support in the desktop import
dialog, including nested zip traversal, Notion title extraction,
hash-stripped folder names, attachments, and folder hierarchy
integration.
- Added i18n entries and adapter coverage for standard Markdown zip
links and Notion Markdown zip imports.

## Why
Markdown and Notion exports often contain links between notes using
relative `.md` paths. Keeping those as plain URLs makes imported
workspaces harder to navigate, so the importer now preassigns document
ids and rewrites resolvable archive-local markdown links into linked
pages.

## Notes
The markdown zip folder hierarchy implementation now comes from latest
`origin/canary`, so this PR only layers relative-link resolution and
Notion Markdown zip support on top of that upstream behavior.

## Validation
- `yarn vitest --run
blocksuite/affine/all/src/__tests__/adapters/markdown.unit.spec.ts`
- `yarn tsc -b blocksuite/affine/all/tsconfig.json --verbose`
- `git diff --check`

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

* **New Features**
* Added support for importing Notion Markdown exports in `.zip` format,
including subpages, attachments, and nested folders.
* Internal links inside imported Markdown now resolve correctly between
pages, preserving link text when available.
* The import dialog now includes a dedicated “Notion (Markdown, .zip)”
option.

* **Bug Fixes**
* Improved filename handling so non-Latin characters in ZIP imports are
preserved correctly.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Co-authored-by: DarkSky <25152247+darkskygit@users.noreply.github.com>
Co-authored-by: DarkSky <darksky2048@gmail.com>
This commit is contained in:
Jachin
2026-07-04 16:42:48 -04:00
committed by GitHub
parent 8063af2da8
commit 477015f064
7 changed files with 725 additions and 52 deletions
@@ -68,8 +68,15 @@ export class Unzip {
private fixFileNameEncoding(fileName: string): string {
try {
// check if contains non-ASCII characters
if (fileName.split('').some(char => char.charCodeAt(0) > 127)) {
// `fflate` already returns valid Unicode filenames for UTF-8 zip entries.
// Only retry decoding for legacy byte-like mojibake strings, otherwise
// normal CJK characters can be corrupted by truncating their code points.
if (
fileName.split('').some(char => {
const code = char.charCodeAt(0);
return code >= 0x80 && code <= 0xff;
})
) {
// try different encodings
const fixedName = this.tryDifferentEncodings(fileName);
if (fixedName && fixedName !== fileName) {