jupyter-storage-explorer-abfs-support

Azure Data Lake Storage support for Storage Explorer, the Jupyter Storage Explorer plugin for Azure Data Lake Storage Gen2: folders moved and deleted in one step, copies within an account on the server, empty folders.

Installation

pip install "jupyter-storage-explorer[abfs]"          # the plugin and its libraries
pip install "jupyter-storage-explorer[abfs-support]"  # the plugin only, with the libraries you have

The libraries are not the plugin's dependencies, so that installing it never changes those of a shared environment. The first command installs them, as does jupyter-storage-explorer-abfs-support[full]:

Without either, the explorer's default operations do everything, file by file.

Prerequisites

• The storage account has a hierarchical namespace (Data Lake Storage Gen2). On other accounts, the plugin only speeds up folder deletions.
• The plugin uses the explorer's adlfs storage with its own account and credential: it needs no settings of its own.

What it does

On storage accounts with a hierarchical namespace (Data Lake Storage Gen2), browsed with adlfs (abfs://), its operation, AzureDataLakeGen2Operation, uses the Data Lake SDK with the adlfs filesystem's own account and credential:

• Moves and renames within an account in one request, however large the folder, never left half moved (adlfs alone copies then deletes file by file).
• Folder deletions in one request, however large the folder, never left half deleted. adlfs alone deletes file by file.
• Copies within an account by the storage itself (Copy Blob): no byte goes through the kernel. A file whose copy fails or is cancelled is deleted, never left partly copied.
• Empty folders, which adlfs alone cannot create.

On accounts without a hierarchical namespace, it deletes folders in batches of files. Their progress moves batch by batch, and Cancel stops them between two batches. adlfs alone checks every file one by one before deleting any.

Anything else is left to the explorer's other operations:

• other storages;
• transfers between accounts;
• moves, copies and new folders on accounts without a hierarchical namespace.

Plugin options

plugin_options gives the plugin's options:

StorageExplorer(
    "abfs://lake@myaccount.dfs.core.windows.net/raw",
    plugin_options={"abfs-support": {"concurrency": 16}},
)

In the launcher's tabs, plugin_options() in launcher.py gives them:

def plugin_options():
    return {"abfs-support": {"concurrency": 16}}

AzureDataLakeGen2Operation

The plugin gives every explorer an AzureDataLakeGen2Operation, made with the plugin's options. To use it in one explorer only, disable the plugin and give the operation yourself:

from jupyter_storage_explorer_abfs_support import AzureDataLakeGen2Operation

StorageExplorer(
    "abfs://lake@myaccount.dfs.core.windows.net/raw",
    plugin_options={"abfs-support": {"enabled": False}},
    filesystem_operations=[AzureDataLakeGen2Operation(concurrency=16)],
)