Simplifying implementation of "G5compare". Updating docs. (#16)

GooseHDF5/__init__.py

@@ -3,7 +3,7 @@
 
 import h5py
 
-__version__ = '0.3.0'
+__version__ = '0.3.1'
 
 # ==================================================================================================
 
@@ -338,6 +338,14 @@ def filter_datasets(data, paths):
 
         datasets = GooseHDF5.filter_datasets(data,
             GooseHDF5.getpaths(data, max_depth=2, fold='/data'))
+
+:arguments:
+
+    **data** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **paths** (``<list<str>>``)
+        A list of paths to datasets.
     '''
 
     import re
@@ -353,6 +361,20 @@ def verify(data, datasets, error=False):
     r'''
 Try reading each dataset of a list of datasets. Return a list with only those datasets that can be
 successfully opened.
+
+:arguments:
+
+    **data** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **datasets** (``<list<str>>``)
+        A list of paths to datasets.
+
+:option:
+
+    **error** ([``False``] | ``True``)
+        If true, the function raises an error if reading failed. If false, the function just
+        continues.
     '''
 
     out = []
@@ -376,6 +398,14 @@ def verify(data, datasets, error=False):
 def exists(data, path):
     r'''
 Check if a path exists in the HDF5-archive.
+
+:arguments:
+
+    **data** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **path** (``<str>``)
+        A path to datasets.
     '''
 
     if path in data:
@@ -388,6 +418,14 @@ def exists(data, path):
 def exists_any(data, paths):
     r'''
 Check if any of the input paths exists in the HDF5-archive.
+
+:arguments:
+
+    **data** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **paths** (``<list<str>>``)
+        A list of paths to datasets.
     '''
 
     if type(paths) == str:
@@ -404,6 +442,14 @@ def exists_any(data, paths):
 def exists_all(data, paths):
     r'''
 Check if all of the input paths exists in the HDF5-archive.
+
+:arguments:
+
+    **data** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **paths** (``<list<str>>``)
+        A list of paths to datasets.
     '''
 
     if type(paths) == str:
@@ -424,6 +470,23 @@ def copydatasets(source, dest, source_datasets, dest_datasets=None, root=None):
 'source_datasets').
 
 In addition a 'root' (path prefix) for the destination datasets name can be specified.
+
+:arguments:
+
+    **source, dest** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **source_datatsets** (``<list<str>>``)
+        A list of paths to datasets in "source".
+
+:options:
+
+    **dest_datasets**  (``<list<str>>``)
+        A list of paths to datasets in "dest".
+        If not specified, it is taken equal to "source_datasets".
+
+    **root** (``<str>``)
+        Path prefix for all 'dest_datasets'.
     '''
 
     import posixpath
@@ -492,6 +555,23 @@ def _equal(a, b):
 # --------------------------------------------------------------------------------------------------
 
 def equal(source, dest, source_dataset, dest_dataset=None):
+    r'''
+Check that a dataset is equal in both files.
+
+:arguments:
+
+    **source, dest** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **source_datatset** (``<str>``)
+        The path to a dataset in ``source``.
+
+:options:
+
+    **dest_dataset**  (``<str>``)
+        The path to a dataset in ``dest``.
+        If not specified, it is taken equal to ``source_dataset``.
+    '''
 
     if not dest_dataset:
         dest_dataset = source_dataset
@@ -509,6 +589,20 @@ def equal(source, dest, source_dataset, dest_dataset=None):
 def allequal(source, dest, source_datasets, dest_datasets=None):
     r'''
 Check that all listed datasets are equal in both files.
+
+:arguments:
+
+    **source, dest** (``<h5py.File>``)
+        A HDF5-archive.
+
+    **source_datatsets** (``<list<str>>``)
+        A list of paths to datasets in "source".
+
+:options:
+
+    **dest_datasets**  (``<list<str>>``)
+        A list of paths to datasets in "dest".
+        If not specified, it is taken equal to "source_datasets".
     '''
 
     if not dest_datasets:

GooseHDF5/cli/G5compare.py

@@ -31,6 +31,7 @@
 
 from .. import __version__
 from .. import getpaths
+from .. import equal
 
 # --------------------------------------------------------------------------------------------------
 # Check if a file exists, quit otherwise.
@@ -41,41 +42,16 @@ def check_isfile(fname):
     if not os.path.isfile(fname):
         raise IOError('"{0:s}" does not exist'.format(fname))
 
-# --------------------------------------------------------------------------------------------------
-# Support function for "check_dataset".
-# --------------------------------------------------------------------------------------------------
-
-def not_equal(path):
-
-    print('!= {0:s}'.format(path))
-    return False
-
 # --------------------------------------------------------------------------------------------------
 # Check if the datasets (read outside) "a" and "b" are equal. If not print a message with the "path"
 # to the screen and return "False".
 # --------------------------------------------------------------------------------------------------
 
-def check_dataset(path, a, b):
-
-    if np.issubdtype(a.dtype, np.number) and np.issubdtype(b.dtype, np.number):
-        if np.allclose(a, b):
-            return True
-        else:
-            return not_equal(path)
-
-    if a.size != b.size:
-        return not_equal(path)
-
-    if a.size == 1:
-        if a[...] == b[...]:
-            return True
-        else:
-            return not_equal(path)
+def check_dataset(source, dest, source_dataset, dest_dataset=None):
 
-    if list(a) == list(b):
-        return True
-    else:
-        return not_equal(path)
+    if not equal(source, dest, source_dataset, dest_dataset):
+        print('!= {0:s}'.format(source_dataset))
+        return False
 
     return True
 
@@ -95,7 +71,7 @@ def _check_plain(source, other):
 
     for path in getpaths(source):
         if path in other:
-            check_dataset(path, source[path][...], other[path][...])
+            check_dataset(source, other, path)
 
 # --------------------------------------------------------------------------------------------------
 
@@ -128,7 +104,7 @@ def _check_renamed(source, other, renamed):
 
     for new_path, path in s2o.items():
         if new_path in o2s:
-            check_dataset(path, source[path][...], other[new_path][...])
+            check_dataset(source, other, path, new_path)
 
 # --------------------------------------------------------------------------------------------------