Box.com

You can read and write files on Box.com. This capability is provided by the “Box.com” plugin, which you need to install. Please see Installing plugins.

How to set up

Install the Box.com plugin in Dataiku from Administration > Plugins. It requires the installation of a code environment.

In order to use the plugin, the administrator of the box.com account will have to create an box.com app:

  • As administrator, go to your box.com account page and upgrade to developer.

  • From the administrator’s developers console, create a new app, select Limited Access App

  • Name your app.

    Note

    People given access to this app will be able to access all files shared to it. Where access to dataset must be restricted to a given group, make and name a separate app for this purpos

  • From the App configuration panel, create and copy a primary and / or secondary access token.

  • Go to the plugin’s setting page (Plugins > Installed > Box.com > Settings > Box.com connection or Project > Settings> Plugins presets). Create a new preset, and paste the access token copied from step 4.

  • Go to the plugin’s setting page and create a preset. In the preset, paste the access token.

How to use

The plugin does not have direct access to the user’s files, but to the app service account. Data can be share with Dataiku by sharing files or directory with this service account. To do so, you need first to retrieve the app’s sharing email address:

  • Inside a Dataiku project, go to the Macros menu

  • Select the Get Box.com sharing email macro

  • Pick a previously created preset, or fill in the access token using Manually defined. Then press Run Macro

  • Copy the email address returned.

To make files visible to Dataiku, you will need to share them from your box account with this email address:

  • Go to your box.com account, locate the files or directories you want to share with Dataiku and click Share.

  • In the Invite People dialog box, paste the email address copied from step 4. Once this is done, you can create a new dataset using the box.com plugin.

First, create the dataset by picking box.com in the dataset section :

  • Choose the Filesystem provider, and fill in the details copied from the app configuration panel.

  • From the connector’s Settings > Files panel, you should now be able to browser your box directory and select the file or directory you want to access.

Browsing speed can be slightly increased by activating the cache option. However this is not available in MUS context.

Write data into Box.com

  • In your Dataiku project flow, select Dataset > Folder

  • Choose Box.com in the Store into parameter and click Create. An External code failed error may appear because you have not selected your Preset yet.

  • In the Settings tab of the folder, select your Preset in the Box.com connection parameter, and save. You can now export your dataset into your newly created Box.com managed folder.