Jupyter Official DockerHub Notes

Described in the hope that it will become knowledge that I have researched to some extent for my own understanding

Relationships between repositories

See Jupyter Official Read the Docs.

Image Relationships

The derivation relationship is as follows.

• base-notebook
  • minimal-notebook
    • scipy-notebook
      • tensorflow-notebook
      • datascience-notebook
      • pyspark-notebook
        • all-spark-notebook
    • r-notebook

It's hard to tell from the image name, but the minimal-notebook has a larger image size than the base-notebook.

Difference between scipy-notebook and datascience-notebook

The one that is (probably) often used

• scipy-notebook --Additional library sets used for data analysis (pandas, scikit-learn, etc.)
• datascience-notebook --scipy-notebook with additional R and Julia

So if you only use Python, scipy-notebook is enough (misunderstood)

About Image Tag

There are many image tags in each repository, and it is difficult to know which one to use at first glance. The rules for image tags seem to have small quirks for each repository, but for example, base-notebook is as follows as of July 29, 2020.

--The yellow highlighted part (b5abe43c6d31) is the SHA digest of the container image. There are 6 tags for the same digest. --In other words, these six are all aliases that point to the same container image. --The last image Tag (b90cce83f37b) is also a SHA digest. This is the commit hash on GitHub side. --The Git commit corresponding to this image Tag is here

As you can see from these, the image tag of base-notebook is registered as a Python version and version updated for each Jupyter component (Jupyter Notebook, JupyterLab, JupyterHub).

What i don't know

In base-notebook, various aliases point to one container image, and you can find the image Tag corresponding to the past Python version.

Here, the most used (probably) scipy-notebook image tags are only commit hashes, and it is difficult to find the image tag corresponding to the target Python version.

• Conversely, I don't know how to find the image tags of base-notebook and minimal-notebook, which are the base images for each image tag of scipy-notebook.

After trying some images, it seems that an image tag is created for each minor version of Python (3.6.6, 3.6.7, 3.8.0, ...)?