Accessing data

Uploading a new dataset

Click on the Datasets button in the tab on the left-hand side of your workspace. This will lead you to the Datasets page:


Select the directory where you would like to upload data and click the Upload file button at the top right-hand side of the page.


A dialog box will appear and selecting Choose file will direct you to a file picker.


Pick the file you wish to upload and confirm your choice by clicking Upload.

Moving data to the workspace

Files can be moved from the Datasets page to the workspace using the sherlockml.filesystem Python package.

For example, you can move a data.csv file from the input folder in datasets to the workspace by opening a notebook and using:

import sherlockml.filesystem as sfs

sfs.get('/input/data.csv', 'data.csv')

For a detailed explanation of SFS check out this quickstart tutorial.

Downloading data

Files can be downloaded from the workspace into your local computer by using SherlockML’s command line interface sml. To find out how to install this utility, take a look at this quickstart tutorial.

As a first step, you need to sign into SherlockML from the terminal of your local computer, via sml. Then, you make a note of the path in SherlockML corresponding to the file that you wish to download (e.g. path/on/sherlockml). In addition, you select the path for the directory on your computer where files will be placed (e.g. path/on/local). Finally, in the terminal of your local machine you run:

$ sml file get project server path/on/sherlockml path/on/local

Here, project is the name of the project that contains the file, and server is the name of any server running on that project.