n5-utils

https://github.com/user-attachments/assets/62d3415e-7f7c-4b43-886f-5e6de8a5b8a1

Collection of standalone command line tools to work with HDF5/ Zarr/ N5 datasets. Including:

• A simple standalone BigDataViewer for multiple datasets or mipmap pyramids.
• A copy, re-chunking, and re-compressing tool for individual datasets or groups.
• A unique tool to get the set of unique numbers in a dataset.
• An equals tool to compare two datasets.

Installation requires maven, OpenJDK, lib-hdf5, [and libblosc] on Ubuntu:

sudo apt-get install openjdk-21-jdk maven hdf5-tools libblosc1

On other platforms, please find your way and report back if interested.

Install into your favorite local binary $PATH:

./install $HOME/bin

All dependencies will be downloaded and managed by maven automatically.

This installs the tools, n5-view, n5-copy, n5-equals, and n5-unique

Run the viewer

n5-view \
  -i '/path/file.h5' \
  -i '/path/file.hdf5' \
  -d /volumes/raw \
  -d /prediction \
  -c 0,255 \
  -c -1,1 \
  -r 1,1,10

to look at two datasets or mipmap pyramids from an HDF5 and N5 container at 1x1x10 resolution and contrast ranges [0,255] and [-1,1]. The viewer maps both contrast ranges into [0,1000] because BDV cannot yet deal with negative intensities and the sliders only show integers. Sorry for the hack.

Copy from N5/HDF5 to N5/HDF5:

n5-copy \
  -i '/path/file.hdf5' \
  -o '/path/file.n5' \
  -b 256,256,26
  -c gzip

or for one or more groups/ datasets:

n5-copy \
  -i '/path/file.hdf5' \
  -o '/path/file.n5' \
  -b 256,256,26
  -c gzip
  -d /volumes

Use as a library in Fiji

This project can be used to deploy the most useful n5 libraries into an existing Fiji installation where they can be used for scripting. Thanks to the scijava-maven-plugin, you can do this by simply passing your Fiji installation path to maven

mvn -Dimagej.app.directory=$HOME/packages/Fiji.app

Then try the experiments in this script with your own data.