Command-line tool and python library for visualising small 3D datasets
Simple matplotlib-based tool for viewing small amounts of 3D image data; helpful for debugging. Supports python 2.7 and 3.4+.
Adapted from this matplotlib recipe.
pip install smalldataviewer[full]
full installation includes all of these optional extras:
hdf5: HDF5 file support via h5py
img: volumetric and animated images via imageio
fits: FITS images via
imageio[fits], which uses astropy
itk: ITK images via
imageio[simpleitk], which uses SimpleITK
Support for N5 and zarr arrays is also available via z5py.
This must be installed with conda (
conda install -c conda-forge -c cpape z5py).
DataViewer opens a
matplotlib figure of the data volume.
- Dimension 0 can be scrolled through with the mouse wheel
- Dimension 1 is shown on the vertical axis
- Dimension 2 is shown on the horizontal axis
- Dimension 3, if it exists, is a colour tuple
Available as a command-line utility at
usage: smalldataviewer [-h] [-i INTERNAL_PATH] [-t TYPE] [-o ORDER] [-f OFFSET] [-s SHAPE] [-v] path positional arguments: path Path to file containing a 3D dataset optional arguments: -h, --help show this help message and exit -i INTERNAL_PATH, --internal_path INTERNAL_PATH Internal path of dataset inside HDF5, N5, zarr or npz file. If JSON, assumes the outer object is a dict, and internal_path is the key of the array -t TYPE, --type TYPE Dataset file type. Inferred from extension if not given. -o ORDER, --order ORDER Order of non-channel axes for axis labelling purposes (data is not transposed): dimension 0 will be scrolled through, dimension 1 will be on the up-down axis, dimension 2 will be on the left-right axis, and dimension 3, if it exists, will be used as the colour channels. Default "zyx". -f OFFSET, --offset OFFSET 3D offset of ROI from (0, 0, 0) in pixels, in the form "<scroll>,<vertical>,<horizontal>" -s SHAPE, --shape SHAPE 3D shape of ROI in pixels, in the form "<scroll>,<vertical>,<horizontal>" -v, --verbose Increase logging verbosity
smalldataviewer my_data.hdf5 -i /my_group/my_volume
Note: because of the circumstances under which python holds file
descriptors open, and under which matplotlib blocks, the executable form
reads the data into memory in its entirety. If your data are too big for
this, look at small chunks with the
import smalldataviewer as sdv import numpy as np data = np.random.random((30, 100, 100)) viewer = sdv.DataViewer(data) viewer.show() # or matplotlib.pyplot.show() viewer2 = sdv.DataViewer.from_file( "my_data.npz", offset=(10, 20, 30), shape=(256, 512, 512), internal_path="volume" ) viewer2.show() reader = sdv.FileReader("my_cat_video.gif") data2 = reader.read() # returns a numpy array viewer3 = sdv.DataViewer(data2) viewer3.show()
FileReader (and by extension
Dataviewer.from_file) reads the requested data
from the file into memory.
Passing an indexable representation of a file, like a numpy memmap or an hdf5 dataset,
However, you may need to copy it into memory for performance, or depending on the rest of your script.
Install a development environment (not including z5py) with
Run tests in your current python environment with
Run tests against all supported python versions with
If you would like to add support for a new file type:
tests/commona function which creates such a file and returns whether it needs an internal path, and add it to
smalldataviewer.files.FileReadera method which reads such a file, returning a numpy array, and add a mapping from likely file extensions to a single file type in
NORMALISED_TYPES(see existing methods for examples).
Don't forget to specify any dependencies in
Release history Release notifications | RSS feed
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Hashes for smalldataviewer-0.8.1-py2.py3-none-any.whl