This is an unofficial Docker image for Netflix's recently open sourced polynote.
The base image is alpine to reduce size. For convenience, the following Python libraries and their pre-requisites are also included:
- PIL for displaying images.
- matplotlib for plotting data and rendering charts.
There is an additional Dockerfile, Dockerfile.opencv, which can be used to build a docker image with opencv dependencies already installed. You can pull these images by using the opencv tag.
When running this image, expose port 8192 to access the web application at: http://localhost:8192/
Example: docker run -p 8192:8192 --name=polynote -d -t greglinscheid/polynote:latest
If you would like to mount a persistent volume for your notebooks, include the -v flag. The following example will mount the container's notebooks/ directory to $HOME/poly-notes
Example: docker run -p 8192:8192 -v $HOME/poly-notes:/usr/src/app/polynote/notebooks --name=polynote -d -t greglinscheid/polynote:latest
If you would like to include opencv dependencies in your container, use the opencv tag instead:
Example: docker run -p 8192:8192 -v $HOME/poly-notes:/usr/src/app/polynote/notebooks --name=polynote -d -t greglinscheid/polynote:opencv
To add more packages, you can exec into the running container and install them.
Python Example:
docker exec -it polynote /bin/ash
pip3 install requests
For the above example, requests will be available without needing to restart the container.
For a solution which persists containers, it is recommended to create your own docker image. You can do this by either modifying requirements.txt and building the repository:
echo 'requests==2.22.0' >> requirements.txt
docker build -t greglinscheid/polynote:latest .
Or by creating a new Dockerfile which uses greglinscheid/polynote as the base:
FROM greglinscheid/polynote:latest
...
RUN pip3 install requests
For convenience, this repository includes shell scripts for building and running this image:
build_image.shrun_container.shbuild_opencv_image.shrun_opencv_container.sh
To build an image from the current repository, execute:
sh build_image.sh
To run the container and map the notebook directory to $HOME/poly-notes, execute:
sh run_container.sh
To build an image with opencv dependencies already installed, execute:
sh build_opencv_image.sh
To run the opencv container and map the notebook directory to $HOME/poly-notes, execute:
sh run_opencv_container.sh