Version Control

File Previews

You can browse the files in a project’s Git repo on the Files tab in the project menu. On this tab, you can navigate through folders and preview files. Jupyter Notebooks (.ipynb), Markdown files (.md), and images (.jpg, .png, .gif) are rendered as HTML for easy previewing.

Warning

To ensure a fast and responsive app experience, data files (.csv, .tsv, .pkl, .hdf, .npy) and any other file larger than 1MB can’t be previewed.

Files in a project repo are identified by their branch and commit. Above every list of files or file preview on the Files tab, you’ll find a branch selector and a commit (represented by a short commit SHA). Any actions you take on the project, like publishing a report or launching an interactive session, will be associated to this branch and commit.

../_images/users-version-control-1.png

The Files tab is meant to be a lightweight way to preview files. To manage branches, explore the Git log, or take any other more advanced actions on your Git repo, use any of the following tools:

Git Actions in the Platform

Every process launched in the Platform, from Jupyter sessions to Model APIs, uses the code you select from your Git repo. When your analysis launches, the Docker container running it has Git configured with your credentials. From any running process, you could run commands such as git status just as you would on your own computer.

In interactive sessions (like Jupyter, RStudio, and Zeppelin), you don’t have to run Git commands yourself. Instead, the Session Sync feature lets you save changes back to your Git remote from the browser without needing the Git command line interface. For detailed instructions on Session Sync, see the Working in a Session section.

../_images/Sync-changes-menu1.png