Desktop: Support importing .one files from OneNote 2016

[personalizedrefrigerator]

Summary

This pull request adds support for importing OneNote 2016/OneNote desktop .one files from the desktop app.

Background

The desktop app previously only supported importing alternate format notebooks downloaded directly from SharePoint or OneDrive.

These notebooks are imported with a fork of msiemens/onenote.rs, rendered to HTML with msiemens/one2html, then imported into Joplin in HTML format. The msiemens/onenote.rs parser, however, as stated in the one2html readme, onenote.rs lacks support for files created by the OneNote desktop app:

• Due to limitations of the OneNote parser only files downloaded from OneDrive are supported. This means you can't convert files created by the OneNote 2016 desktop application using this tool.
  — https://github.com/msiemens/one2html?tab=readme-ov-file#limitations

This pull request updates the forked onenote.rs to support parsing OneNote 2016-format notebooks.

Import process

Although this pull request makes it possible to import data from OneNote 2016/OneNote desktop, the process is still somewhat complicated. For possible improvements, see the "Follow-up tasks" section.

To import a .one file:

1. Move the OneNote 2016 .one files into a folder.
2. Convert the folder to a .zip file.
3. Select the .zip file from Joplin's File > Import > ZIP dialog.

To import a .onepkg file:

1. Extract the .onepkg file.
   • onepkg files are CAB archives. On Windows, it may be possible to extract the onepkg file by first changing the extension to .cab.
2. Create a new .zip file containing the .one and .onetoc2 files from step 1.
3. Import the .zip file from step 2 using Joplin's file > import menu.

Change summary

• Added a new local_onestore module and moved the previous alternate format parsing logic into fsshttpb_onestore.
• Implemented parsing logic for the OneNote desktop format in local_onestore.
• Created a onestore module with adapter logic.
• Refactored onenote-converter into multiple sub-packages (managed with cargo workspaces).
  • The packages include: renderer, parser, parser-macros, and parser-utils.
  • This allows creating a #[derive(...)] macro to simplify parsing binary data. For why this needs to be in a separate package, see the relevant section of the Rust book.

Possible follow-up tasks

• Allow selecting .one files from the OneNote importer dialog. Currently, only .zip files can be selected.
• Allow selecting .onepkg files from the OneNote importer dialog.
  • onepkg files are CAB archives containing .one and .onetoc2 files. Joplin would need support for extracting the one files from the onepkg archive.
• Add automated tests that can run in CI.

Testing

Automated testing

An automated test has been created that uses a sample .one file from OfficeDev/Interop-TestSuites. At present, it must be run manually with cargo test from packages/onenote-converter and does not run in CI.

Manual testing

It has been verified that the OneNote importer can successfully import a user-provided .onepkg file by:

1. Converting the .onepkg file to a .zip file.
2. Importing the .zip file from Joplin.

[laurent22]

There's a conflict on lib.rs

[personalizedrefrigerator]

I've found a bug related to how the parser handles . I'm currently encountering the following error when trying to import this file from the OfficeDev/InteropTestSuites and this other file from the Apache Tika test suites:

Malformed OneNote file data: unexpected object type: 0x20046

I'm converting this to a draft until the bug is fixed.

Edit: Notes:

• One possibility is that this is specific to .one files created with older versions of OneNote (the problematic FileSyncWOPI_OneNoteWithODB.one was created 10+ years ago, but the testOneNote1.one seems to have been uploaded more recently).
• The 0x20046 ID corresponds to a version history metadata node.

[personalizedrefrigerator]

Renamed from src/parser/reader.rs, with the following changes: Added tests, added absolute_offset, advance, and with_updated_bounds methods.