# (C) Copyright 2004-2023 Enthought, Inc., Austin, TX # All rights reserved. # # This software is provided without warranty under the terms of the BSD # license included in LICENSE.txt and may be redistributed only under # the conditions described in the aforementioned license. The license # is also available online at http://www.enthought.com/licenses/BSD.txt # # Thanks for using Enthought open source! """This demo shows how to use Traits TreeEditors with PyTables to walk the hierarchy of an HDF5 file. This only picks out tables, arrays and groups, but could easily be extended to other structures. In the demo, the path to the selected item is printed whenever the selection changes. An example HDF5 file is provided here, but you could easily change the path given at the bottom of this file to a path to your own HDF5 file. To run this demonstration successfully, you must have the following packages installed: - **PyTables** (``tables``) - **HDF5 for Python** (``h5py``) Note that PyTables can't read HDF5 files created with h5py, but h5py can read HDF5 files created with PyTables. See HDF5_tree_demo2 for an example using h5py. """ import os import tables as tb from traits.api import HasTraits, Str, List, Instance, Any from traitsui.api import TreeEditor, TreeNode, View, Item, Group ROOT = os.path.dirname(__file__) # View for objects that aren't edited no_view = View() # HDF5 Nodes in the tree class Hdf5ArrayNode(HasTraits): name = Str('') path = Str('') parent_path = Str('') class Hdf5GroupNode(HasTraits): name = Str('') path = Str('') parent_path = Str('') # Can't have recursive traits? Really? # groups = List(Hdf5GroupNode) groups = List() arrays = List(Hdf5ArrayNode) groups_and_arrays = List() class Hdf5FileNode(HasTraits): name = Str('') path = Str('/') groups = List(Hdf5GroupNode) arrays = List(Hdf5ArrayNode) groups_and_arrays = List() # Recursively build tree, there is probably a better way of doing this. def _get_sub_arrays(group, h5file): """Return a list of all arrays immediately below a group in an HDF5 file.""" return [ Hdf5ArrayNode( name=array._v_name, path=array._v_pathname, parent_path=array._v_parent._v_pathname, ) for array in group if isinstance(array, (tb.Array, tb.Table)) ] # More pythonic # for array in h5file.iter_nodes(group, classname='Array')] # Old call def _get_sub_groups(group, h5file): """Return a list of all groups and arrays immediately below a group in an HDF5 file.""" subgroups = [] for subgroup in h5file.iter_nodes(group, classname='Group'): subsubgroups = _get_sub_groups(subgroup, h5file) subsubarrays = _get_sub_arrays(subgroup, h5file) subgroups.append( Hdf5GroupNode( name=subgroup._v_name, path=subgroup._v_pathname, parent_path=subgroup._v_parent._v_pathname, groups=subsubgroups, arrays=subsubarrays, groups_and_arrays=subsubgroups + subsubarrays, ) ) return subgroups def _hdf5_tree(filename): """Return a list of all groups and arrays below the root group of an HDF5 file.""" with tb.open_file(filename, 'r') as h5file: subgroups = _get_sub_groups(h5file.root, h5file) subarrays = _get_sub_arrays(h5file.root, h5file) h5_tree = Hdf5FileNode( name=filename, groups=subgroups, arrays=subarrays, groups_and_arrays=subgroups + subarrays, ) return h5_tree # Get a tree editor def _hdf5_tree_editor(selected=''): """Return a TreeEditor specifically for HDF5 file trees.""" return TreeEditor( nodes=[ TreeNode( node_for=[Hdf5FileNode], auto_open=True, children='groups_and_arrays', label='name', view=no_view, ), TreeNode( node_for=[Hdf5GroupNode], auto_open=False, children='groups_and_arrays', label='name', view=no_view, ), TreeNode( node_for=[Hdf5ArrayNode], auto_open=False, children='', label='name', view=no_view, ), ], editable=False, selected=selected, ) class ATree(HasTraits): h5_tree = Instance(Hdf5FileNode) node = Any() traits_view = View( Group( Item( 'h5_tree', editor=_hdf5_tree_editor(selected='node'), resizable=True, show_label=False, ), orientation='vertical', ), title='HDF5 Tree Example', buttons=['Undo', 'OK', 'Cancel'], resizable=True, width=0.3, height=0.3, ) def _node_changed(self): print(self.node.path) def make_test_datasets(): """Makes the test datasets and store it in the current folder""" import h5py import numpy as np import pandas as pd # pandas uses pytables to store datasets in hdf5 format. from random import randrange n = 100 df = pd.DataFrame( dict( [ ("int{0}".format(i), np.random.randint(0, 10, size=n)) for i in range(5) ] ) ) df['float'] = np.random.randn(n) for i in range(10): df["object_1_{0}".format(i)] = [ '%08x' % randrange(16 ** 8) for _ in range(n) ] for i in range(7): df["object_2_{0}".format(i)] = [ '%15x' % randrange(16 ** 15) for _ in range(n) ] df.info() df.to_hdf('test_fixed.h5', 'data', format='fixed') df.to_hdf('test_table_no_dc.h5', 'data', format='table') df.to_hdf('test_table_dc.h5', 'data', format='table', data_columns=True) df.to_hdf( 'test_fixed_compressed.h5', 'data', format='fixed', complib='blosc', complevel=9, ) # h5py dataset time = np.arange(n) x = np.linspace(-7, 7, n) axes_latlon = [ ('time', time), ('coordinate', np.array(['lat', 'lon'], dtype='S3')), ] axes_mag = [ ('time', time), ('direction', np.array(['x', 'y', 'z'], dtype='S1')), ] latlon = np.vstack( (np.linspace(-0.0001, 0.00001, n) + 23.8, np.zeros(n) - 82.3) ).T mag_data = np.vstack( ( -(1 - np.tanh(x) ** 2) * np.sin(2 * x), -(1 - np.tanh(x) ** 2) * np.sin(2 * x), -(1 - np.tanh(x) ** 2), ) ).T datasets = ( axes_mag + axes_latlon + [('magnetic_3_axial', mag_data), ('latlon', latlon)] ) with h5py.File(os.path.join(ROOT, 'test_h5pydata.h5'), "a") as h5file: h5group = h5file.require_group("run1_test1") for data_name, data in datasets: h5group.require_dataset( name=data_name, dtype=data.dtype, shape=data.shape, data=data, # **options ) def main(): import sys filename = os.path.join(ROOT, 'test_fixed.h5') filename = os.path.join(ROOT, 'test_table_no_dc.h5') if len(sys.argv) > 1: filename = sys.argv[1] a_tree = ATree(h5_tree=_hdf5_tree(filename)) return a_tree demo = main() if __name__ == '__main__': # make_test_datasets() demo.configure_traits()