Environment management¶

This page outlines how to configure and start the notebook environment.

Manage the notebook environment¶

Before you create and execute code, click the environment icon to configure the notebook's environment. The environment image determines the coding language, dependencies, and open-source libraries used in the notebook. The default image for the DataRobot Notebooks is a pre-built Python image. To see the list of all packages available in the default image, hover over that image in the Environment tab:

The screenshot and table below outline the configuration options available for a notebook environment.

Manage session status¶

To begin a notebook session where you create and run code, start the environment by toggling it on in the toolbar.

Wait a moment for the environment to initialize, and once it displays the Started status, you can begin editing.

If you upgrade any of the existing packages in the notebook environment during your session and want the upgraded version to be recognized, you need to restart the kernel. To do so, click the circular arrow icon in the toolbar.

Note that restarting the kernel is different than restarting the environment session: when you stop the environment session (using the session toggle), this will stop the container your notebook is running in. The notebook state and any packages installed at runtime will be lost, as the next time you start the session, a new container will be spun up.

Note that the session will automatically shut down on inactivity when the timeout is reached. The session is considered inactive when the notebook has no running cells and no changes have been made to the notebook contents in the time set by the session timeout value.

Manage exposed ports¶

To create a port for port forwarding, click + Add port, enter a Port number and optional Description, and then click the accept icon :

Once you have added the port to forward, you can start the container session for the notebook or codespace. After launching a web application, access the web application via the URL for the port listed in the Exposed ports list:

• Click Link to open a new browser tab with your application running.

• Click the copy icon to copy the web application URL.

To change a port number, edit the port description or stop forwarding that port, click the actions menu :

• Click Edit to modify the Port number and Description.

• Click Delete to remove the exposed port, ending session forwarding to that port.

Feature considerations and FAQ¶

Review the following considerations and frequently asked questions before configuring port forwarding:

%%writefile app.py
from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()

@app.get("/")
async def echo():
    return {"message": "Hello, world!"}

streamlit run app.py 

streamlit run app.py --server.port 8080

mlflow ui --port 8080 --host 0.0.0.0

pkill -f mlflow.server

options(shiny.port = 7775)
options(shiny.host = "0.0.0.0")

ui <- fluidPage(
"Hello, world!"
)

server <- function(input, output, session) {
}

shinyApp(ui, server)

!kedro viz run --host 0.0.0.0

tensorboard --bind_all --logdir .

import gradio as gr
from fastapi import FastAPI


def greet(name, intensity):
    return "Hello, " + name + "!" * int(intensity)

demo = gr.Interface(
    fn=greet,
    inputs=["text", "slider"],
    outputs=["text"],
)


fapp = FastAPI(root_path="/notebook-sessions/{NOTEBOOK_ID}/ports/{YOUR_PORT}")

gr.mount_gradio_app(fapp, demo, f"/", root_path="/notebook-sessions/{NOTEBOOK_ID}/ports/{YOUR_PORT}")

uvicorn app:fapp --host 0.0.0.0 --port 7860 --reload

from nicegui import ui

class Demo:
    def __init__(self):
        self.number = 1

demo = Demo()
v = ui.checkbox('visible', value=True)
with ui.column().bind_visibility_from(v, 'value'):
    ui.slider(min=1, max=3).bind_value(demo, 'number')
    ui.toggle({1: 'A', 2: 'B', 3: 'C'}).bind_value(demo, 'number')
    ui.number().bind_value(demo, 'number')

ui.run()

!python app.py

from fastapi import FastAPI
from pydantic import BaseModel
app = FastAPI()
@app.get("/")
async def echo():
    return {"message": "Hello there"}

!uvicorn app:app --host 0.0.0.0 --port 8501 --reload

Custom environment images¶

DataRobot Notebooks is integrated with DataRobot custom environments, allowing you to define reusable custom Docker images for running notebook sessions. You can create a custom environment to use for your notebook sessions if you want full control over the environment, and to leverage reproducible dependencies beyond those available in the built-in images.

Create a custom environment¶

When creating a custom environment, note that only Python and R custom environments are supported for DataRobot Notebooks. Additionally, in the Environment type field, select the Notebooks option to use the environment with DataRobot Notebooks.

DataRobot strongly recommends accessing the notebook built-in environment templates to reference the requirements needed to create a Docker context that is compatible with running DataRobot Notebooks.

After you create a custom environment, you can click the Current Notebooks tab to view a list of any DataRobot notebooks that are configured to use versions of the custom environment as the notebook container image.

Use a custom environment for a notebook session¶

To use a custom environment as the image to run for a notebook container session, open the notebook you want to use and navigate to the Environment tab () in the notebook sidebar.

In the Environment field dropdown, you can see all of the custom environments that you have access to that are compatible with DataRobot Notebooks. Select the custom environment you’d like to use. In the Version field, you can select the version of the environment to use. By default, the latest version of the environment is selected. Once you’ve configured your environment selection, you can start the notebook session.

Built-in environment images¶

DataRobot maintains a set of built-in Docker images that you can select from to use as the container image for a given notebook.

DataRobot provides the following images:

• Python 3.9 image: Contains Python version 3.9, the DataRobot Python client, and suite of common data science libraries.

• Python 3.8 image: Contains Python version 3.8, the DataRobot Python client, and suite of common data science libraries. Previously, this image was needed to work with Snowflake as certain included libraries Snowflake Python Connector and Snowpark required Python 3.8. These libraries now support later versions of Python so this image is being considered for deprecation

• R 4.3 image: Contains R version 4.3, the DataRobot R client, and a suite of common data science libraries.

Environment variables¶

If you need to reference sensitive strings in a notebook, rather than storing them in plain text within the notebook you can use environment variables to securely store the values. These values are stored encrypted by DataRobot. Environment variables are useful if you need to specify credentials for connecting to an external data source within your notebook, for instance.

Whenever you start a notebook session, DataRobot sets the notebook's associated environment variables in the container environment, so you can reference them from your notebook code using the following code:

import os
KEY = os.environ['KEY']  # KEY variable now references your VALUE

KEY = Sys.getenv("KEY")

To access environment variables, click the lock icon in the sidebar.

Click Create new entry.

In the dialog box, enter the key and value for a single entry, and provide an optional description.

If you want to add multiple variables, select Bulk import. Use the following format on each line in the field:

KEY=VALUE # DESCRIPTION

When you have finished adding environment variables, click Save.

Edit existing variables¶

You can also edit and delete a notebook’s associated environment variables from the Environment variables panel:

• Click the pencil icon on a variable to edit it.
• Select the eye icon to view a hidden value.
• Click the trash can icon to delete a variable.
• Click Insert all to insert a code snippet that retrieves all of the notebook's environment variables and includes them in the notebook.