Figure Explorer is a VS Code extension for browsing, searching, organizing, and navigating figures generated in Jupyter notebooks.
When working with notebooks containing many plots, it can be difficult to keep track of which cell produced which figure. Figure Explorer collects figures from your open notebooks and provides a dedicated gallery for quickly browsing them, searching their contents, organizing them with tags, comparing figures, and jumping back to their source cells.
Features
- Automatically discovers PNG figures generated by open Jupyter notebooks.
- Synchronizes with notebook changes as cells are edited or outputs are regenerated.
- Browse figures in a dedicated thumbnail gallery.
- Search figures by title, notebook, cell number, figure number, tags, or source code.
- Add custom tags to figures for organization and filtering.
- Displays a larger preview of the selected figure.
- Jump directly from a figure to the notebook cell that generated it.
- Select multiple figures for side-by-side comparison.
- Save figures as PNG files.
- Export figures as PDF files.
Gallery
The Gallery view provides a thumbnail overview of the figures available from the selected notebook or from all open notebooks.
Selecting a figure displays a larger preview along with information about its source cell. The gallery can be searched and filtered without leaving the notebook workflow.
Figures can be navigated with the keyboard using the arrow keys. Double-clicking a thumbnail provides a shortcut for revealing the corresponding source cell in the notebook.
Search and Filters
The Gallery toolbar provides controls for searching and filtering figures.
Search can be performed across figure metadata, including notebook names, titles, source code, cell information, and user-defined tags.
The gallery can also be switched between the current notebook and all open notebooks, and figures can be filtered based on their title status or tags.
Figure Titles
You can optionally give a figure a title by adding a # figure: comment as the first non-empty line of a notebook cell:
# figure: Residual distribution
plt.hist(residuals)
plt.xlabel("Residual")
plt.ylabel("Count")
plt.show()
The title is displayed throughout Figure Explorer and can be used when searching for figures.
Cells without an explicit figure title are automatically assigned a number such as Figure 1, Figure 2, etc.
Tags
Figures can be assigned custom tags directly from Figure Explorer, or tags can be defined in the notebook source using a # tags: comment.
For example:
# figure: Residual distribution
# tags: residuals, metallicity, z=0
plt.hist(residuals)
plt.xlabel("Residual")
plt.ylabel("Count")
plt.show()
Multiple tags can be separated by commas:
# tags: MZR, metallicity, paper, final
Tags are displayed with the figure and can be used for searching and filtering. This makes it possible to organize figures according to their scientific content, purpose, redshift, analysis stage, or any other user-defined category.
Tags assigned through Figure Explorer are stored as figure metadata and persist independently of the notebook output itself.
Search
The Gallery supports several structured search modes.
cell:12
Find the figure generated by cell 12.
figure:5
Find figure 5.
title:residual
Search figure titles.
code:plt.scatter
Search the source code associated with figures.
You can also enter a general search term to search across the figure's notebook, title, source code, and tags.
Figure Comparison
Figure Explorer includes a comparison mode for viewing multiple figures together.
Figures can be selected from the gallery and opened in a dedicated comparison view. This allows related figures to be examined side by side without losing the normal gallery workflow.
Comparison mode is particularly useful for comparing figures from different cells or notebooks, such as different redshifts, model configurations, parameter choices, or versions of a plot.
Explorer
The Figure Explorer activity-bar view provides an overview of the notebooks currently being tracked by the extension.
Each notebook can be expanded to view a list of its discovered figures. The list is synchronized with the notebooks currently open in VS Code, so closing a notebook removes it from the Explorer automatically.
This provides a convenient way to move between figures across multiple open notebooks.
Figure Navigation
The Gallery supports keyboard navigation through the currently visible set of figures.
Use the arrow keys to move between figures. Navigation follows the current filtered and searched gallery contents rather than the complete collection, making it possible to quickly move through a specific subset of figures.
Double-clicking a thumbnail reveals the notebook cell that generated the selected figure.
Saving and Exporting Figures
Figures can be saved directly from Figure Explorer as PNG files.
Figures can also be exported as PDF files when a publication-ready or vector-based output is needed.
Getting Started
- Install Figure Explorer from the VS Code Marketplace.
- Open one or more Jupyter notebooks in VS Code.
- Run cells that generate figures.
- Open the Figure Explorer activity-bar view.
- Open the Gallery.
- Browse, search, filter, tag, compare, or reveal figures as needed.
Figure Explorer updates as notebook cells and outputs change, so you can continue working in your notebooks without manually rescanning them.
Requirements
- VS Code 1.125.0 or later
- Jupyter notebooks (
.ipynb)
Current Limitations
Figure Explorer currently focuses on PNG image outputs generated by Jupyter notebooks.
Other output types and image formats are not currently indexed by the figure scanner.
Feedback
Figure Explorer is designed with scientific and data-analysis workflows in mind. Feedback, bug reports, and feature suggestions are welcome.
The preferred way to report issues or suggest features is through the project's GitHub repository:
GitHub: [figure-explorer/issues]
When reporting a problem, please include:
- Figure Explorer version
- VS Code version
- Operating system
- A description of the notebook or output that caused the problem
- Steps to reproduce the problem
- Any relevant error message from VS Code
- Screenshots, if applicable
Feature requests and workflow suggestions are also welcome.
License
Figure Explorer is released under the MIT License.