How to Run Jupyter Notebooks Locally in a Docker Container
We use Jupyter Notebook to quickly run and test simple Python scripts and entire Jupyter notebooks. But installing the Jupyter Notebook on our local PC can be quite redundant if we do not use it often. For that kind of situation, we can run Jupyter Notebook in a Docker container without installing anything, at all.
Just open a new terminal window and run the following command. If you need to access Jupyter through a different port other than default 8888, you can mention it like
docker run -p 8888:8888 jupyter/scipy-notebook
A few seconds later, you will be able to see a URL something similar to below in the terminal. Directly open it or copy and paste the complete
http://127.0.0.1:8888/?token=9e880… URL in the browser.
[C 10:16:14.004 NotebookApp] To access the notebook, open this file in a browser: file:///home/jovyan/.local/share/jupyter/runtime/nbserver-7-open.html Or copy and paste one of these URLs: http://127.0.0.1:8888/?token=9e88021ada57b03fda0a548c9e88f3f5fd6a1ced4ce6405e
Now you should be able to use the Jupyter Notebook.
Make Jupyter Notebooks Persist With Docker Volumes
By default, the files you create with Jupyter Notebook such as
txt will NOT persist in permanent storage. That means, when you stop or remove the Jupyter Notebook Docker container, all the files you created will be permanently lost and will not be able to recover. But we can easily attach an external directory in the host as a Docker volume to the container so we can easily access the project files. For that, create a directory.
At this moment, the Jupyter Notebook Docker container will not be able to write files into the directory. Run the following command to change the ownership of the directory we created. After that, the Jupyter Notebook Docker container will get full permission to read and write files into the directory.
chown -R 1000 work/
Now run the following command, make sure to mention the full path to the directory you created.
docker run -p 8888:8888 -v /full/path/to/work:/home/jovyan/work jupyter/scipy-notebook
How To Fix “ModuleNotFoundError”
When you start to work with many notebooks, occasionally you might an error message similar to this.
ModuleNotFoundError: No module named 'tensorflow'
Under that kind of situation, install the necessary modules by adding them all at the top as a new cell. After running that newly added cell, make sure to restart the notebook kernel to apply changes.
conda install tensorflow keras
Visit Jupyter Docker Stacks for more information about available features.