Every analysis that runs on the Platform, from Jupyter sessions to Model APIs, uses code stored in a Git repository (commonly referred to as a “repo”). Actions like runs, APIs, and sessions all start from a specific version (called a “commit”) of the repo, letting you maintain a transparent record of your process.
It’s good to understand the basics of Git before using the DataScience.com Platform. This page of resources curated by GitHub is a good start.
There is a one-to-one relationship between projects on the platform and Git repositories. Once you’ve connected a repo to a project, your team may not connect it to any other project. This rule keeps each project’s analyses closely tied to changes in the source code so others can retrace your steps.
The Platform supports Git repositories hosted on GitHub (including GitHub Enterprise), Bitbucket (including Bitbucket Server), and GitLab (both Community and Enterprise). You’ll use your credentials for these Git providers to work with repos directly in the DataScience.com Platform.
Before your team can authenticate with a Git provider, an Admin should follow the instructions in the Git Provider Integration docs.
Managing Git Provider Connections¶
After an Admin has registered a Git provider with DataScience.com, you can connect your account with the following steps:
- On the DataScience.com Platform, click on your avatar in the upper right-hand corner, click Settings, and navigate to the Git Integrations tab.
- Click Add Authentication and search for providers registered by your Administrator.
- Complete the form with your Git provider account information.
Your registered Git connections appear on this Git Integrations page. From this list, you can edit connections via the button on the right. To remove a connection from the list, hover over that connection and a Remove option will appear.
Disconnecting from your provider means that the Platform will clear out your authentication information (tokens, passwords, etc.) from its secure storage. Removing your your authentication won’t cause anything to be lost on the Platform, but you will no longer be able to see files or launch analyses in your projects that were connected with that Git provider.
The GitHub integration uses GitHub’s OAuth features to connect your account. On the DataScience.com Platform, you’ll only have to click the Connect button in the GitHub integration form to sign in and approve access.
You’ll be prompted to enter your GitHub username and password. Once you’re successfully connected, you’ll be redirected back to the Platform.
The Bitbucket integration uses Bitbucket’s App Passwords feature to grant repo access. A Bitbucket App Password is just like your account password but meant for other apps to control Bitbucket on your behalf.
Start by creating a Bitbucket App Password inside the Bitbucket web app. Visit Bitbucket Settings > App Passwords and select Create an App Password. Give the app password a name and select the following permissions:
Account: Email + Read
Team membership: Read
Repositories: Read + Write
After you create an App Password on Bitbucket and copy it from the pop-up window, you’ll no longer be able to view the password on Bitbucket. Keep this password safe by immediately copying it over to the DataScience.com Platform.
With your Bitbucket App Password copied, go to the DataScience.com Platform and click on your avatar in the upper right-hand corner to reveal Settings > Git Integrations. Click Add Credentials, search for your Bitbucket integration, and enter your user name, email, and app password.
You can find your Bitbucket username in the Bitbucket app by visiting your account settings. For more information on Bitbucket App Passwords, see their docs.
GitLab Enterprise v9 and Higher¶
The GitLab integration uses GitLab’s Access Tokens feature to grant repo access. A GitLab Access Token is just like your account password but meant for other apps to control GitLab on your behalf.
Start by creating a GitLab Access Token inside the GitLab web app. Visit User Settings > Access Tokens. Give the app password a name, select api and read_user under Scopes, then click Create Personal Access Token.
After you create a token on GitLab and copy it from the confirmation screen, you’ll no longer be able to view it on GitLab. Keep this token safe by immediately copying it over to the DataScience.com Platform.
With your GitLab token copied, go to the DataScience.com Platform and via the avatar drop-down, visit Settings > Git Integrations. Click Add Credentials, search for your GitLab integration, and enter your user name, email, and access token.
Gitlab Enterprise v7 and v8¶
These versions only require your GitLab username and password to allow access to the GitLab repo. Creating Access Tokens is not necessary.