Demo

Description

h5tui is a vim-style TUI for browsing HDF5 files, inspired by terminal managers like ranger, lf, and yazi.

• Browse: Three-column tree view of HDF5 files with Vim-style keybindings. Includes fuzzy search for quick navigation.
• Plots: Inline 1D line plot or histogram, 2D heatmap with plotext or PDF export via matplotlib.

Installation

The package is on PyPI:

pip install h5tui

Usage

h5tui file.h5

Documentation

File Navigation

h5tui starts at the root HDF5 group. Groups are 📁, datasets 📊.

Key	Action
h / j / k / l	Move left / down / up / right
g / G	Jump to top / bottom
d / u	Page down / up
~	Return to root

Searching

Key	Action
/	Find in group
n / p	Next / previous match
f	Fuzzy search entire file

Viewing Datasets

Datasets are opened in a viewer that supports the same type of VIM keybindings for navigating the cells.

Additional keybindings:

Key	Action
:	Jump to row
y	Copy focused cell to clipboard

Plotting

Inline

h5tui plots in the terminal with plotext.

Key	1D	2D	>2D
v	Line plot	Heatmap	—
V	Histogram	—	—

For compound datatypes, any field can be chosen as the x and y axis.

Note for large datasets, only an evenly sampled subset of the datapoints is plotted.

Export as PDF

Press s to open dialog to Save Plot As PDF via matplotlib.

Attributes

Items with attributes display ▼ and a (count). Press a to view them.

Aggregation

Press A while viewing a dataset for a numerical summary including (e.g., mean, variance, max). This is performed per field for compound types.

Shared navigation

Most screens reuse the same motion and dismissal keys; meaning depends on focus (tree vs table vs modal).

Pattern	Keys
Move	h j k l, arrows
Jump 10 grid cells	H J K L
Start/end	g G
Page	u d, PageUp PageDown
Dismiss/quit	Esc, q, Q

Limitations

h5tui is meant for reading HDF5 files only; it does not provide functionality for modifying them.

Acknowledgements

This project builds on h5py, Textual, plotext, matplotlib, numpy, and RapidFuzz.