Install your packages

Mercury Cloud is using separate container for each notebook, which comes with Python 3.9 and the following libraries preinstalled:

mercury mljar-supervised numpy pandas matplotlib seaborn altair vega_datasets plotly scikit-learn

Install public packages

You can also use other libraries in your notebook. You have to create a requirements.txt file with libraries names:

package_name1
package_name2

You can include versions of libraries:

package_name1==version1
package_name2>=version2

Upload requirements.txt file

Please upload your requirements.txt file just like you do it with your notebook:

Mercury will automatically read that file and will install all of your packages before opening the notebook.

Install private packages

You can use secrets in the requirements.txt file. You can also install packages from private github repositories. Please read about Secrets Management in our docs for more details.

For example you can install packages from github private repository.

git+https://${GITHUB_TOKEN}@github.com/user/project.git@{version}