Support Network File Sharing Protocol (SMB Protocol)

[lucasjamar]

Description

I often work with data located on a network drive. If I am working locally, I can easily load such data by specifying the displayed path, e.g.:

companies:
  type: pandas.CSVDataSet
  filepath: O:/department_name/team_name/companies.csv

However, if I want to move my work to the cloud, I need to use the SMB protocol. It is already supported by fsspec. Using normal pandas read/ write csv, I do something like this:

df = pd.read_csv("smb://myuser:mypassword@myserver.com:445/department_name/team_name/companies.csv")

However, when I plug something similar into my data catalog, I get the following error message:

companies:
  type: pandas.CSVDataSet
  filepath: smb://myuser:mypassword@myserver.com:445/department_name/team_name/companies.csv

File "C:\Users\JAMAR_L\Anaconda3\envs\kedro-alpiq\lib\site-packages\kedro\extras\datasets\pandas\csv_dataset.py", line 121, in __init__
    self._fs = fsspec.filesystem(self._protocol, **_credentials, **_fs_args)
  File "C:\Users\JAMAR_L\Anaconda3\envs\kedro-alpiq\lib\site-packages\fsspec\registry.py", line 244, in filesystem
    return cls(**storage_options)
  File "C:\Users\JAMAR_L\Anaconda3\envs\kedro-alpiq\lib\site-packages\fsspec\spec.py", line 69, in __call__
    obj = super().__call__(*args, **kwargs)
TypeError: __init__() missing 1 required positional argument: 'host'

When I look at the code of kedro.io.core, I notice that smb protocol is not listed.

HTTP_PROTOCOLS = ("http", "https")
PROTOCOL_DELIMITER = "://"
CLOUD_PROTOCOLS = ("s3", "gcs", "gs", "adl", "abfs")

Possible Implementation

Has anyone come across this issue before and perhaps have a workaround? If not, I will implement this by adding smb to the list of protocols.

Thanks :)

[merelcht]

Hi @lucasjamar, this sounds like a good addition. We'd be happy to accept a PR for this 🙂

[lorenabalan]

Closing this as per #782 (comment) and related discussion on the PR.