Show HN: Dietnb – Prevent Jupyter notebooks from bloating with Base64 images

2 days ago 4

dietnb addresses the issue of large .ipynb file sizes caused by matplotlib figures being embedded as Base64 data. By saving plots as external PNG files and embedding only image links, dietnb keeps your notebooks lightweight and improves manageability.

Minimized Notebook Size: Significantly reduces .ipynb file bulk by storing matplotlib figures as external PNG files.
Automatic Image Folder Management: Creates and manages image storage directories (e.g., [NotebookFileName]_dietnb_imgs) relative to the notebook's location. (Applies when path detection is successful, e.g., in VS Code; defaults to dietnb_imgs in the current working directory if detection fails.)
Automatic Image Updates: When a notebook cell is re-executed, image files generated from its previous run are automatically deleted, ensuring only the latest output is retained.
Image Cleanup Function: The dietnb.clean_unused() function allows easy removal of unreferenced image files from the current session.
Simple Auto-Activation: The dietnb install command configures dietnb to activate automatically when IPython and Jupyter environments start.

Installation and Activation

1. Install the dietnb package

Execute the following command in your terminal:

2. Choose an Activation Method

A. Automatic Activation (Recommended) Run the following command in your terminal once:

This creates a startup script (00-dietnb.py) in your IPython profile directory. After restarting your Jupyter kernel, dietnb will be activated automatically. Images will be saved to a folder based on the notebook's path or to the default dietnb_imgs directory.

To disable automatic activation later, run:

This removes the startup script.

B. Manual Activation (Per Notebook) If you prefer to use dietnb only for specific notebooks or do not want automatic activation, add the following code at the top of your notebook to activate it manually:

import dietnb dietnb.activate()

With dietnb active, use your matplotlib code as usual.

import matplotlib.pyplot as plt import numpy as np # Create a plot x = np.linspace(0, 10, 100) plt.plot(x, np.sin(x), label='sin(x)') plt.plot(x, np.cos(x), label='cos(x)') plt.title("Trigonometric Functions") plt.xlabel("X-axis") plt.ylabel("Y-axis") plt.legend() plt.show() # On show(), the image is saved to a file, and a link is displayed in the notebook.

Generated images can be found in the [NotebookFileName]_dietnb_imgs folder alongside your notebook, or in the dietnb_imgs folder.