support for reading dataset fillvalues

pyfive/high_level.py

@@ -316,7 +316,7 @@ def fletcher32(self):
     @property
     def fillvalue(self):
         """ fillvalue attribute. """
-        raise NotImplementedError
+        return self._dataobjects.fillvalue
 
     @property
     def dims(self):

pyfive/low_level.py

@@ -619,6 +619,37 @@ def shape(self):
         msg_offset = msg['offset_to_message']
         return determine_data_shape(self.msg_data, msg_offset)
 
+    @property
+    def fillvalue(self):
+        """ Fillvalue of the dataset. """
+        msg = self.find_msg_type(FILLVALUE_MSG_TYPE)[0]
+        offset = msg['offset_to_message']
+        version = struct.unpack_from('<B', self.msg_data, offset)[0]
+        if version == 1 or version == 2:
+            info = _unpack_struct_from(FILLVAL_MSG_V1V2, self.msg_data, offset)
+            offset += FILLVAL_MSG_V1V2_SIZE
+            is_defined = info['fillvalue_defined']
+        elif version == 3:
+            info = _unpack_struct_from(FILLVAL_MSG_V3, self.msg_data, offset)
+            offset += FILLVAL_MSG_V3_SIZE
+            is_defined = info['flags'] & 2**5
+        else:
+            raise InvalidHDF5File(
+                "Unknown fillvalue msg version:" + str(version))
+
+        if is_defined:
+            size = struct.unpack_from('<I', self.msg_data, offset)[0]
+            offset += 4
+        else:
+            size = 0
+
+        if size:
+            payload = self.msg_data[offset:offset+size]
+            fillvalue = np.fromstring(payload, self.dtype, count=1)[0]
+        else:
+            fillvalue = 0
+        return fillvalue
+
     @property
     def dtype(self):
         """ Datatype of the dataset. """
@@ -1318,6 +1349,22 @@ def _unpack_struct_from(structure, buf, offset=0):
     ('heap_address', 'Q'),      # 8 byte addressing
 ))
 
+# IV.A.2.f. The Data Storage - Fill Value Message
+FILLVAL_MSG_V1V2 = OrderedDict((
+    ('version', 'B'),
+    ('space_allocation_time', 'B'),
+    ('fillvalue_write_time', 'B'),
+    ('fillvalue_defined', 'B'),
+))
+FILLVAL_MSG_V1V2_SIZE = _structure_size(FILLVAL_MSG_V1V2)
+
+FILLVAL_MSG_V3 = OrderedDict((
+    ('version', 'B'),
+    ('flags', 'B'),
+))
+FILLVAL_MSG_V3_SIZE = _structure_size(FILLVAL_MSG_V3)
+
+
 # IV.A.2.l The Data Storage - Filter Pipeline message
 FILTER_PIPELINE_DESCR_V1 = OrderedDict((
     ('filter_id', 'H'),

tests/make_fillvalue.py

@@ -0,0 +1,24 @@
+#! /usr/bin/env python
+""" Create two HDF5 files with datasets with and without fillvalues. """
+import h5py
+import numpy as np
+
+
+def make_fillvalue(f):
+
+    common_args = {
+        'shape': (4, ),
+        'data': np.arange(4),
+        'track_times': False,
+    }
+
+    f.create_dataset('dset1', dtype='<i1', fillvalue=42, **common_args)
+    f.create_dataset('dset2', dtype='<i1', **common_args)
+    f.create_dataset('dset3', dtype='<f4', fillvalue=99.5, **common_args)
+
+
+with h5py.File('fillvalue_earliest.hdf5', 'w', libver='earliest') as f:
+    make_fillvalue(f)
+
+with h5py.File('fillvalue_latest.hdf5', 'w', libver='latest') as f:
+    make_fillvalue(f)

tests/test_fillvalue.py

@@ -0,0 +1,24 @@
+""" Test pyfive's Dataset fillvalue attribute. """
+import os
+
+import pyfive
+
+DIRNAME = os.path.dirname(__file__)
+DATASET_FILLVALUE_EARLIEST_HDF5_FILE = os.path.join(
+    DIRNAME, 'fillvalue_earliest.hdf5')
+DATASET_FILLVALUE_LATEST_HDF5_FILE = os.path.join(
+    DIRNAME, 'fillvalue_latest.hdf5')
+
+
+def test_dataset_fillvalue_earliest():
+    with pyfive.File(DATASET_FILLVALUE_EARLIEST_HDF5_FILE) as hfile:
+        assert hfile['dset1'].fillvalue == 42
+        assert hfile['dset2'].fillvalue == 0
+        assert abs(hfile['dset3'].fillvalue - 99.5) < 0.05
+
+
+def test_dataset_fillvalue_latest():
+    with pyfive.File(DATASET_FILLVALUE_LATEST_HDF5_FILE) as hfile:
+        assert hfile['dset1'].fillvalue == 42
+        assert hfile['dset2'].fillvalue == 0
+        assert abs(hfile['dset3'].fillvalue - 99.5) < 0.05