Jupytext Sync: Pair and Auto Sync Jupyter Notebooks via Jupytext in VSCode

This VSCode extension integrates jupytext's pairing and synchronization features, enabling automatic syncing between notebooks and their version-control-friendly text formats (.py, .md, etc.).

Demos

Paired notebook with self-contained metadata

Paired notebook via config file

Installation

You can install this extension by searching for "jupytext-sync" in the Extensions Marketplace within your VSCode-like IDE.

The extension is available on both Microsoft VSCode Marketplace and on the Open VSX Registry in case you are using an alternative VSCode-based IDE like VSCodium.

Alternatively, you can install it directly by downloading the .vsix file from the release page on the GitHub and drag-and-drop it in your VSCode-based IDE. This has the downside that you will not be notified about available updates.

The extension activates when VS Code has finished starting up (onStartupFinished), so it should be available shortly after launch.

Requirements

• Python: You need a Python installation.
• Jupytext: The jupytext Python package must be installed in the Python environment used by the extension.
• VS Code Microsoft Python Extension (Recommended): For the best experience with automatic Python environment detection, it is recommended to have the Microsoft Python extension installed (open in VSCode / open on marketplace).

Features

This extension solves several common annoyances and provides handy features for a better Jupyter Notebook experience in VSCode:

• Automatic Synchronization:
  • Automatically syncs paired files (.ipynb and text-based formats like .py, .md) when you open, save, or close them. This ensures that your notebook and its text representation are always in sync.
  • Configuration options allow you to customize which events trigger a sync (on open, on save, on close for both text documents and notebook documents). See the jupytextSync.syncDocuments setting for details.
• Seamless Notebook-First Workflow:
  • Auto-open notebooks: When you open a paired text file (e.g., .py, .md), the extension automatically opens the paired .ipynb notebook instead, providing a seamless editing experience. This behavior can be customized via the jupytextSync.autoOpenNotebook setting.
  • Auto-cleanup: Optionally delete the .ipynb file when closing the notebook editor, keeping your workspace clean when the text file is your primary source of truth. Deleted files are moved to trash and can be recovered. See the jupytextSync.deleteOnNotebookClose setting for details.
• Open Text Files as Notebooks:
  • Open Jupytext-compatible text files (e.g., .py, .md with Jupytext metadata) as notebooks using the "Open as paired Notebook via Jupytext" command. A notebook file will be created and all your edits in it will be synced to the text file(s).
• Persistent Cell Outputs: All files are saved to disk, meaning cell outputs are not lost when reopening a notebook, providing a more consistent experience.
• Convenient Pairing Command: The "Pair via Jupytext" command is accessible from multiple convenient locations:
  • Command Palette (Ctrl+Shift+P or Cmd+Shift+P)
  • File Explorer context menu (right-click on a file)
  • Editor context menu (right-click inside the editor or on the editor tab)
  • Icons in the editor and notebook toolbars.
• Broad File Support: Supports all file extensions and formats recognized by your jupytext installation. The extension dynamically fetches the supported formats from jupytext. The jupytextSync.defaultFormats setting allows you to pre-configure default pairing strings for various extensions.
• Raw Cells Support in Notebooks:
  • Easily insert new raw cells via buttons in Notebook editor toolbar or keyboard shortcuts.
  • Toggle cells to raw format and back to default code. Keyboard shortcuts available.
• Compact Notebook Layout: A command to apply a suggested VSCode settings for a more compact notebook layout, similar to traditional Jupyter interfaces.
• Python Interpreter Flexibility:
  • Attempts are made to automatically discover Python executables that are able to invoke Jupytext. If the Microsoft Python extension is installed (open in VSCode / open on marketplace), its selected interpreter and other discovered environments (e.g., venv, conda) are considered.
  • Allows you to configure a custom Python executable path for jupytext if needed (via the jupytextSync.pythonExecutable setting).
  • You can use the "Jupytext: Locate Python and Jupytext" command or check the extension logs ("Jupytext: Show Jupytext Sync Logs" command) to see which Python environment is being used.

Recommended Workflow

1. Pair your Notebook:

   • Open an .ipynb file or a text file (e.g., myscript.py, mymarkdown.md) that you want to use as a notebook.
   • Use the "Pair via Jupytext" command (accessible from the Command Palette, context menus, or toolbar icons).
   • You'll be prompted to choose the Jupytext formats (e.g., ipynb,py:percent) unless disabled via jupytextSync.askFormats.onPairDocuments. The default suggestion is configurable per file extension using jupytextSync.defaultFormats.

2. Work in the Notebook:

   • With jupytextSync.autoOpenNotebook enabled (default), simply open your paired text file (e.g., .py, .md), and the extension will automatically open the .ipynb notebook for you.
   • The extension will automatically keep the paired files in sync upon saving (or other configured events via jupytextSync.syncDocuments).
   • When you close the notebook, you can optionally have the .ipynb file automatically deleted (configurable via jupytextSync.deleteOnNotebookClose), keeping your workspace clean while preserving your text file as the source of truth.

3. Version Control:

   • Commit the text-based file (e.g., .py, .md) to your Git repository. This file is human-readable and diff-friendly.
   • Optionally, you can also commit the .ipynb file if you prefer to version control the outputs too.
   • If using auto-delete on close, consider adding *.ipynb to your .gitignore to avoid tracking generated notebook files.

4. Pre-commit Hook (Recommended): To ensure your paired files are always synchronized before committing, it's highly recommended to use the jupytext pre-commit hook. This prevents accidental commits of unsynced files.

   Refer to the Jupytext documentation on pre-commit hooks for detailed configuration.

Configuration

Extension Settings

You can configure the extension's behavior via VSCode settings (search for jupytextSync in the Settings UI or edit your settings.json):

• jupytextSync.pythonExecutable:

  • Description: The path to the Python executable used to invoke jupytext.
  • Details: Jupytext Sync requires a Python executable with the jupytext package installed.
    • Automatic discovery: If not specified, the extension attempts to find a suitable Python executable. If the Microsoft Python extension is installed, its selected interpreter and other known environments are checked. Otherwise, it looks for python and python3 in your system PATH. The one providing the highest jupytext version is preferred.
    • Manual override: Specify an absolute path or a command (e.g., python3). If using a command, ensure your VS Code instance inherits the correct PATH (launching from an activated terminal might be necessary for virtual environments).
  • Tip: Use the "Jupytext: Show Jupytext Sync Logs" or "Jupytext: Locate Python and Jupytext" commands to verify which Python executable is being used.
  • Default: "" (empty string, for automatic discovery)

• jupytextSync.autoOpenNotebook:

  • Description: Automatically open the paired .ipynb notebook when opening a text file (e.g., .py, .md) that is paired with a notebook via Jupytext.
  • Details: When enabled, opening a paired text file will:
    1. Check if the file has a paired .ipynb format.
    2. Sync the files.
    3. Open the .ipynb file in the Notebook Editor instead.
  • Tip: This provides a seamless notebook-first editing experience for paired files. You can bypass this behavior using the "Open With..." context menu to open the file as a regular text file.
  • Note: This only affects files that jupytext reports as paired. Unpaired files will open normally in the text editor.
  • Default: true

• jupytextSync.deleteOnNotebookClose:

  • Description: Control whether to delete the .ipynb notebook file when closing a paired notebook editor.
  • Options:
    • "never": Never delete the notebook file when closing.
    • "ask": Always ask for confirmation before deleting.
    • "yes": Always delete if the notebook has paired formats.
    • "if auto created": Only delete if the notebook was auto-created by this extension via "Open as paired Notebook" or by opening a paired text file with the custom editor.
  • Details: Deleted files are moved to the system trash/recycle bin and can be recovered. This feature is designed for workflows where the .ipynb file is generated from source text files (e.g., .py, .md) and the text file is the primary source of truth.
  • Default: "ask"

• jupytextSync.syncDocuments:

  • Description: Controls on which events to attempt to jupytext --sync previously paired documents. This applies even if pairing was done externally.
  • Properties:
    • onNotebookDocumentOpen (boolean): Sync when opening a notebook document.
    • onNotebookDocumentSave (boolean): Sync when saving a notebook document. (Default: true)
    • onNotebookDocumentClose (boolean): Sync when closing a notebook document.
    • onTextDocumentOpen (boolean): Sync when opening a supported text document.
    • onTextDocumentSave (boolean): Sync when saving a supported text document. (Default: true)
    • onTextDocumentClose (boolean): Sync when closing a supported text document.
  • Default: See individual property defaults above.

• jupytextSync.askFormats:

  • Description: Controls whether to ask for pairing file formats before executing commands that may require pairing. If false, jupytextSync.defaultFormats are used.
  • Properties:
    • onOpenPairedNotebook (boolean): Ask for formats before opening a text document as a paired notebook. (Default: false)
    • onPairDocuments (boolean): Ask for formats before pairing documents. (Default: true)

• jupytextSync.defaultFormats:

  • Description: Define default Jupytext pairing formats (e.g., ipynb,py:percent) used as suggestions or defaults.
  • Syntax: Uses Jupytext's --set-formats string.
    • ${ext} can be used as a placeholder for the file extension (without the leading dot).
    • default key: An extension set to "default" in the configuration will inherit the format string from this key.
    • Subdirectories: Prefix with dir_name// (e.g., notebooks//ipynb,scripts//py:percent). Paths are relative to the source file's parent directory.
  • Activation: Applies to file extensions recognized by your installed jupytext.
  • Example Default Entry: "default": "ipynb,${ext}:percent"

• jupytextSync.setFormatsArgs:

  • Description: Customization for the command-line arguments of jupytext --set-formats invocation. The order is preserved.
  • Examples: ["--set-formats"], ["--set-formats", "--some-flag"]
  • Default: ["--set-formats"]

• jupytextSync.syncArgs:

  • Description: Customization for the command-line arguments of jupytext --sync invocation. The order is preserved.
  • Example: ["--sync", "--use-source-timestamp"]
  • Default: ["--sync"]

• jupytextSync.enabledMenus:

  • Description: Enable or disable Jupytext menus and buttons in various VS Code UI locations.
  • Properties (all boolean, default true):
    • explorerContext: "Pair via Jupytext" in Explorer context menu.
    • editorContext: "Pair via Jupytext" in Editor context menu.
    • editorTitle: "Pair via Jupytext" button in Editor title bar.
    • editorTitleContext: "Pair via Jupytext" in Editor title context menu.
    • notebookToolbar: "Pair via Jupytext" button in Notebook toolbar.
    • notebookToolbarInsertRaw: "Insert Raw Cell" button in Notebook toolbar.
    • notebookToolbarToRaw: "Toggle Raw/Code Cell" button in Notebook toolbar.

Jupytext Behavior

To configure Jupytext's own behavior (e.g., metadata filters, default formats), consult the official Jupytext documentation on configuring Jupytext via pyproject.toml or jupytext.toml files.

For example, to avoid noisy metadata in your text notebooks, you might add this to your pyproject.toml:

[tool.jupytext]
notebook_metadata_filter = "-kernelspec,-jupytext.text_representation.jupytext_version"
cell_metadata_filter = "-all"

Commands

This extension provides the following commands, accessible via the Command Palette (Ctrl+Shift+P or Cmd+Shift+P) under the "Jupytext" category:

• Jupytext: Show Jupytext Sync Logs: Opens the output channel for Jupytext Sync, showing logs which can be helpful for troubleshooting, especially for Python executable discovery.
• Jupytext: Locate Python and Jupytext: Manually triggers Python and Jupytext discovery, showing detailed information about which Python executable and Jupytext version is being used. Useful for troubleshooting environment issues.
• Jupytext: Pair via Jupytext: Initiates the pairing process for the active file or a file selected from the Explorer. Prompts for Jupytext formats based on jupytextSync.askFormats.onPairDocuments and jupytextSync.defaultFormats.
• Jupytext: Open as paired Notebook via Jupytext: Opens a Jupytext-compatible text file (e.g., .py, .md) as a VS Code notebook. This is a key feature for working with text-based versions of notebooks.
• Jupytext: Insert Raw Code Cell Below and Focus Container: Inserts a new raw cell below the active cell in a notebook.
• Jupytext: Insert Raw Code Cell Above and Focus Container: Inserts a new raw cell above the active cell in a notebook.
• Jupytext: Change Cell to default Code: Changes the selected notebook cell(s) to the default code type.
• Jupytext: Change Cell to Raw Code: Changes the selected notebook cell(s) to raw format.
• Jupytext: Toggle Cell between Raw Code and default Code: Toggles the selected notebook cell(s) between raw and default code formats.
• Jupytext: Set Suggested Compact Notebook Layout: Applies a set of VS Code settings to achieve a more compact notebook UI.

Default Keybindings

The following keybindings are available when a notebook editor is focused and you are not currently typing in a cell input or output area:

• y: Jupytext: Change Cell to default Code
• r: Jupytext: Toggle Cell between Raw Code and default Code
• t: Jupytext: Insert Raw Code Cell Below and Focus Container
• e: Jupytext: Insert Raw Code Cell Above and Focus Container
• i i (press i twice): jupyter.interruptkernel (VS Code built-in)
• 0 0 (press 0 twice): jupyter.restartkernel (VS Code built-in)

Acknowledgements

This extension was inspired by and builds upon the ideas from the following projects:

• marius311/jupytext-paired-vscode
• parmentelat/vscode-jupytext
  • itself a soft fork of notebookPowerTools/vscode-jupytext
  • with fixes from congyiwu/vscode-jupytext

Related Projects

• fmilanese-1/ds_utils: A VSCode extension that has some jupytext-based capabilities and supports exporting notebooks to HTML via nbconvert, but does not include automatic synchronization of paired files.