H5View: a light, h5py/IPython-friendly HDF5 viewer in text mode

This light Python module lets you explore interactively any HDF5 file in a Python console or in IPython. Pretty print the file structure and get the shape, data type, and size of any dataset. Tab completion works in IPython to let you access any subgroup or dataset.

H5View is based on the h5py package.

Usage

import h5view
with h5view.open('test.h5') as f:
    # display the whole structure of the file
    print(f)
    # access a group and display its information
    print(f.MyGroup1.SubGroup)
    # access a dataset
    X = f.MyGroup2.MyDataset[0,:]
    # access an attribute
    val = f.MyGroup3.myAttr
    # access to the corresponding h5py object
    item = f.MyDataset.item()
    # get a descendant from its relative path
    print(f.MyGroup4.get('MySubgroup/MyDataset'))

In IPython, tab completion shows all direct children (groups, datasets, attributes) of the file or any item (group/dataset), making it quite convenient to interactively explore a HDF5 file.

Installation

The code is only in h5view.py. The create.py file is used to create a test HDF5 file.

Download the package and type in a shell (a test HDF5 file is created automatically in the same folder):

$ python h5view.py
Let's display the file.
<HDF5 file "/home/me/h5view/test.h5", 50.2 KB>
/
  * MyAttr: 23.0
/MyGroup1
  * MyAttr2: my string
/MyGroup1/MyGroup11
/MyGroup1/MyGroup11/MyGroup111
/MyGroup1/MyGroup11/MyGroup111/MyDataset2: shape (2, 3, 4, 5), dtype "float32", 480 B
/MyGroup2
/MyGroup2/MyDataset1: shape (100, 100), dtype "int32", 39.1 KB

Now, let's access some attributes, groups, and datasets.
23.0
/MyGroup2
/MyGroup2/MyDataset1: shape (100, 100), dtype "int32", 39.1 KB
[[42 42]
 [42 42]]
[ 0.  0.]