added more unit tests for read_excel and load_example_data to compensate for skipped doctests

because the later kind cannot be skipped when the required optional dependency is missing

Author: gdementen

larray/tests/data/test.xlsx

@@ -3241,25 +3241,40 @@ def test_read_excel_xlwings():
     arr = read_excel(inputpath('test.xlsx'), '2d')
     assert_array_equal(arr, io_2d)
 
+    arr = read_excel(inputpath('test.xlsx'), '2d_classic')
+    assert_array_equal(arr, ndtest("a=a0..a2; b0..b2"))
+
+    arr = read_excel(inputpath('test.xlsx'), '2d_classic', nb_axes=2)
+    assert_array_equal(arr, ndtest("a=a0..a2; b0..b2"))
+
     arr = read_excel(inputpath('test.xlsx'), '3d')
     assert_array_equal(arr, io_3d)
 
+    # for > 2d, specifying nb_axes is required if there is no name for the horizontal axis
+    arr = read_excel(inputpath('test.xlsx'), '3d_classic', nb_axes=3)
+    assert_array_equal(arr, ndtest("a=1..3; b=b0,b1; c0..c2"))
+
     arr = read_excel(inputpath('test.xlsx'), 'int_labels')
     assert_array_equal(arr, io_int_labels)
 
-    arr = read_excel(inputpath('test.xlsx'), '2d_classic')
-    assert_array_equal(arr, ndtest("a=a0..a2; b0..b2"))
-
     # passing a Group as sheet arg
     axis = Axis('dim=1d,2d,3d,5d')
 
     arr = read_excel(inputpath('test.xlsx'), axis['1d'])
     assert_array_equal(arr, io_1d)
 
+    # missing rows, default fill_value
+    arr = read_excel(inputpath('test.xlsx'), 'missing_values')
+    expected = ndtest("a=1..3; b=b0,b1; c=c0..c2", dtype=float)
+    expected[2, 'b0'] = nan
+    expected[3, 'b1'] = nan
+    assert_array_nan_equal(arr, expected)
+
     # missing rows + fill_value argument
     arr = read_excel(inputpath('test.xlsx'), 'missing_values', fill_value=42)
-    expected = io_missing_values.copy()
-    expected[isnan(expected)] = 42
+    expected = ndtest("a=1..3; b=b0,b1; c=c0..c2", dtype=float)
+    expected[2, 'b0'] = 42
+    expected[3, 'b1'] = 42
     assert_array_equal(arr, expected)
 
     # range
@@ -3325,24 +3340,31 @@ def test_read_excel_pandas():
     arr = read_excel(inputpath('test.xlsx'), '1d', engine='xlrd')
     assert_array_equal(arr, io_1d)
 
-    arr = read_excel(inputpath('test.xlsx'), '2d', nb_axes=2, engine='xlrd')
+    arr = read_excel(inputpath('test.xlsx'), '2d', engine='xlrd')
     assert_array_equal(arr, io_2d)
 
-    arr = read_excel(inputpath('test.xlsx'), '2d', engine='xlrd')
+    arr = read_excel(inputpath('test.xlsx'), '2d', nb_axes=2, engine='xlrd')
     assert_array_equal(arr, io_2d)
 
+    arr = read_excel(inputpath('test.xlsx'), '2d_classic', engine='xlrd')
+    assert_array_equal(arr, ndtest("a=a0..a2; b0..b2"))
+
+    arr = read_excel(inputpath('test.xlsx'), '2d_classic', nb_axes=2, engine='xlrd')
+    assert_array_equal(arr, ndtest("a=a0..a2; b0..b2"))
+
     arr = read_excel(inputpath('test.xlsx'), '3d', index_col=[0, 1], engine='xlrd')
     assert_array_equal(arr, io_3d)
 
     arr = read_excel(inputpath('test.xlsx'), '3d', engine='xlrd')
     assert_array_equal(arr, io_3d)
 
+    # for > 2d, specifying nb_axes is required if there is no name for the horizontal axis
+    arr = read_excel(inputpath('test.xlsx'), '3d_classic', nb_axes=3, engine='xlrd')
+    assert_array_equal(arr, ndtest("a=1..3; b=b0,b1; c0..c2"))
+
     arr = read_excel(inputpath('test.xlsx'), 'int_labels', engine='xlrd')
     assert_array_equal(arr, io_int_labels)
 
-    arr = read_excel(inputpath('test.xlsx'), '2d_classic', engine='xlrd')
-    assert_array_equal(arr, ndtest("a=a0..a2; b0..b2"))
-
     # passing a Group as sheet arg
     axis = Axis('dim=1d,2d,3d,5d')
 
@@ -3370,7 +3392,7 @@ def test_read_excel_pandas():
     # missing rows + fill_value argument
     arr = read_excel(inputpath('test_narrow.xlsx'), 'missing_values',
                      fill_value=42, wide=False, engine='xlrd')
-    expected = io_narrow_missing_values
+    expected = io_narrow_missing_values.copy()
     expected[isnan(expected)] = 42
     assert_array_equal(arr, expected)
 

larray/tests/test_array.py

@@ -0,0 +1,55 @@
+from __future__ import absolute_import, division, print_function
+
+import pytest
+
+from larray.tests.common import needs_pytables
+from larray import load_example_data, Array, Session, Axis, Group
+
+
+def obj_summary(obj):
+    if isinstance(obj, Session):
+        return {obj_name: obj_summary(child_obj) for obj_name, child_obj in obj.items()}
+    elif isinstance(obj, Array):
+        return [obj_summary(axis) for axis in obj.axes]
+    elif isinstance(obj, Axis):
+        return obj.name, len(obj)
+    elif isinstance(obj, Group):
+        return obj.name, obj.axis.name, len(obj)
+    else:
+        raise TypeError('unsupported object')
+
+
+@needs_pytables
+def test_load_example_data():
+    demo = load_example_data('demography')
+    expected_summary = {
+        'hh': [('time', 26), ('geo', 3), ('hh_type', 7)],
+        'pop': [('time', 26), ('geo', 3), ('age', 121), ('sex', 2), ('nat', 2)],
+        'qx': [('time', 26), ('geo', 3), ('age', 121), ('sex', 2), ('nat', 2)]
+    }
+    assert obj_summary(demo) == expected_summary
+
+    demo = load_example_data('demography_eurostat')
+    expected_summary = {
+        'births': [('country', 3), ('gender', 2), ('time', 5)],
+        'citizenship': ('citizenship', 3),
+        'country': ('country', 3),
+        'country_benelux': ('country', 3),
+        'deaths': [('country', 3), ('gender', 2), ('time', 5)],
+        'even_years': ('even_years', 'time', 2),
+        'gender': ('gender', 2),
+        'immigration': [('country', 3),
+                        ('citizenship', 3),
+                        ('gender', 2),
+                        ('time', 5)],
+        'odd_years': ('odd_years', 'time', 3),
+        'population': [('country', 3), ('gender', 2), ('time', 5)],
+        'population_5_countries': [('country', 5), ('gender', 2), ('time', 5)],
+        'population_benelux': [('country', 3), ('gender', 2), ('time', 5)],
+        'time': ('time', 5)
+    }
+    assert obj_summary(demo) == expected_summary
+
+
+if __name__ == "__main__":
+    pytest.main()