Skip to main content

Overview

Jupyterlab is a fully-featured python notebook environment.

We provide it to authors when JupyterLite has insufficient compatibility.

Should you use JupyterLab?​

If possible, we strongly recommend you use JupyterLite instead of JupyterLab.

For more information on why JupyterLite is preferred over JupyterLab see JupyterLite vs JupyterLab.

For information on JupyterLite's limitations and suitability for your lab, visit Should you use JupyterLite?.

You should only use JupyterLab if you have confirmed, based on the above resources, that JupyterLite is unsuitable for your purposes.

JupyterLab Types​

We offer two distinct versions of JupyterLab: JupyterLab Current and JupyterLab Classic. If you cannot use JupyterLite for your Jupyter notebook lab, we strongly recommend using JupyterLab Current unless you need some of the unique features of JupyterLab Classic. JupyterLab Current offers:

In contrast, JupyterLab Classic offers:

  • Python 3.7
  • An older kernel for R
  • Kernels for Julia and Swift
  • Slower loading than JupyterLab Current
  • Free access to the OpenAI API, but not the watsonx.ai and Anthropic APIs; see Free APIs Available in Skills Network Labs for details and limitations

Since JupyterLab Current is strongly recommended over JupyterLab Classic, we have provided a guide to help you convert JupyterLab Classic labs to JupyterLab Current labs. You can find the guide at Upgrading to the Latest Version of JupyterLab.

Using JupyterLab​

Installing Packages​

Install from PyPI​

Packages can be installed from PyPI using pip:

!pip install optuna

To ensure that labs are not broken when PyPI updates library versions we strongly recommend pinning the version you wish to install:

!pip install optuna==4.2.0

When installing packages from PyPI, a substantial amount of output can be generated. To hide this output from learners, you can use the %%capture magic command. Be aware that using %%capture might cause a cell to run for an extended period without producing output. In such cases, it is advisable to inform learners that the cell will take several minutes to execute:

%%capture
!pip install optuna==4.2.0

After installing packages in JupyterLab, we recommend restarting your kernel and informing your learners to do the same. Please see JupyterLab Best Practices for details.

Install using conda or mamba​

Packages can be installed using !conda, %conda, !mamba, or %mamba. Since mamba is significantly faster than conda, it is strongly recommended to use !mamba or %mamba. Remember to use the -y option to automatically confirm all prompts. For example, to install matplotlib from the conda-forge channel, you can use the following command:

!mamba install -c conda-forge -y matplotlib

To ensure that labs are not broken when library versions are updated on conda channels we strongly recommend pinning the version you wish to install:

!mamba install -c conda-forge -y matplotlib==3.10.0

When installing packages using conda or mamba, a substantial amount of output can be generated. To hide this output from learners, you can use the %%capture magic command. Be aware that using %%capture might cause a cell to run for an extended period without producing output. In such cases, it is advisable to inform learners that the cell will take several minutes to execute:

%%capture
!mamba install -c conda-forge -y matplotlib==3.10.0

After installing packages in JupyterLab, we recommend restarting your kernel and informing your learners to do the same. Please see JupyterLab Best Practices for details.

Downloading Data​

JupyterLab has access to the internet, meaning that you can source datasets from online sources. For instance, .csv files can be downloaded and loaded as pandas dataframes using:

import pandas as pd
URL = 'https://www.url.to/my/dataset.csv'
df = pd.read_csv(URL)

Moreover, with JupyterLab you have access to wget with !wget and curl with !curl. For instance, the following downloads and loads a dataset from Kaggle:

!curl -L -o netflix-shows.zip\
  https://www.kaggle.com/api/v1/datasets/download/shivamb/netflix-shows
import zipfile
import pandas as pd
archive = zipfile.ZipFile('netflix-shows.zip', 'r')
df = pd.read_csv(archive.open('netflix_titles.csv'))
df.sample(5)

Relying on online sources carries the risk of labs becoming non-functional if those sources are moved or become unavailable. Therefore, we strongly recommend transferring these data sources to the Skills Network File Library before using them in your labs. To upload files to the Skills Network File Library, follow the steps provided in Skills Network File Library - Access File Library within Jupyterlite or Jupyterlab.

JupyterLab Best Practices​

Follow these best practices when creating JupyterLab labs:

  • Pin python package versions when installing from PyPI. For instance, instead of using !pip install numpy, use !pip install numpy==2.2.1
  • Pin python package versions when installing using conda or mamba. For instance, instead of using !mamba install -c conda-forge -y matplotlib, use !mamba install -c conda-forge -y matplotlib==3.10.0
  • Ask learners to restart their kernel between installing Python libraries and importing them. There are several ways that this can be accomplished. You can include a code cell that, when run, restarts the kernel:
import os
os._exit(00)

And, as a workaround in case the above code fails to restart the kernel for a learner, you can include the following image in a markdown cell that asks the learner to click the kernel restart button: An image that shows the location of the jupyterlab restart kernel button

JupyterLab Caveats and Workarounds​

Issue Solution
The pyspark library fails with JVM error

In JupyterLab Classic, Use findspark:

!pip install pyspark==3.4.4
!pip install findspark==2.0.1
import findspark
findspark.init()
from pyspark import SparkContext, SparkConf
from pyspark.sql import SparkSession
# Creating a spark context class
sc = SparkContext()

# Creating a spark session
spark = SparkSession \
    .builder \
    .appName("Python Spark DataFrames basic example") \
    .config("spark.some.config.option", "some-value") \
    .getOrCreate()

Should run with no errors.

Convert from JupyterLab to JupyterLite​

You can easily convert from JupyterLab to JupyterLite by following Convert JupyterLab to JupyterLite.

Convert from JupyterLite to JupyterLab​

You can easily convert from JupyterLite to JupyterLab by following Convert JupyterLite to JupyterLab.