Jupyter Storage Explorer launcher

Adds Storage Explorer to JupyterLab's launcher (section Other) and command palette. It opens Jupyter Storage Explorer in its own tab, with a kernel of its own that stops when the tab is closed.

The tab opens on the launcher's folder: the one JupyterLab's file browser showed when the launcher was opened (like the launcher's Notebook and Console items; open a new launcher after changing folder). From the command palette, it opens on the folder the file browser shows. The tab reopens on the storage and folders it showed (kept with JupyterLab's workspace) when the page is reloaded, and when the Jupyter server or the tab's kernel restarts, with a new kernel.

It also lets you drag files from any Storage Explorer, in its tab or in a notebook, onto JupyterLab's file browser: after asking, the explorer's kernel downloads them into the folder dropped on, or the one the file browser shows. This needs the kernel to run where the Jupyter server does, as it does by default.

pip install jupyter-storage-explorer-launcher

Install it in the environment of JupyterLab and of the default kernel (on JupyterHub, usually the single-user image), then restart JupyterLab. It installs jupyter-storage-explorer too. Requires JupyterLab 4.

Configuration

What the tab shows is set by files storage-explorer/launcher.py in Jupyter's configuration directories (jupyter --paths, config section). All the files found are used, as layers, in the reverse of that order: first the system's (/etc/jupyter/storage-explorer/launcher.py), usually last the user's (~/.jupyter/storage-explorer/launcher.py; in a virtual environment, Jupyter searches the environment's <environment>/etc/jupyter first, so that file is applied after the user's). Each layer adds to or changes the settings of those before. Without any file, the tab shows an explorer of the folder it opens on.

A file has some of these functions, all optional:

"." is the folder the tab opens on, and height is "fill" (as tall as the tab) unless set. The tabs keep their bookmarks and the data browser's settings in ~/.jupyter/storage-explorer/states.json (in JUPYTER_CONFIG_DIR if set) unless states is set: another path, or None to keep them in each tab's memory. A change to a final setting is ignored, with a warning shown in the tab. Final settings help administrators set defaults; they are not a security boundary, as users can still create explorers of their own in notebooks.

For example, for everyone, /etc/jupyter/storage-explorer/launcher.py:

def filesystems():
    return {
        "Here": ".",
        "Data lake": "abfs://lake@myaccount.dfs.core.windows.net",
    }


def final():
    return ["filesystems"]  # users cannot change the storages offered

And for one user, ~/.jupyter/storage-explorer/launcher.py, with icons of their own (see Customizing icons):

def parameters():
    return {"max_browser_items": 50_000}


def style():
    # The folders in the top folder as 🗃️ instead of 📁: hide the icon, show another
    return """
    .jse-browser [data-depth="1"] .jse-icon { font-size: 0; }
    .jse-browser [data-depth="1"] .jse-icon::before { font-size: 14px; content: "🗃️"; }
    """

A transfer policy is a parameter too, e.g. in the system's file:

def parameters():
    # No downloads from the lake to users' computers or to the Jupyter server.
    return {
        "transfer_policy": [
            {"rule": "deny", "src": r"abfs://lake@.*", "dest": "browser:///.*"},
            {"rule": "deny", "src": r"abfs://lake@.*", "dest": "jupyterlab://.*"},
        ]
    }


def final():
    return ["transfer_policy"]

An error in a file (a function that fails, an unknown parameter) is shown in the tab, with the file's path.