Refactor root Dataverse filter and fix tests

Author: rliebz

dataverse/connection.py

@@ -2,7 +2,7 @@
 import requests
 
 from dataverse import Dataverse
-from utils import get_elements
+from utils import get_elements, is_not_root_dataverse
 
 
 class Connection(object):
@@ -36,21 +36,19 @@ def connect(self):
         self.connected = True if self.status == 200 else False
         self.service_document = etree.XML(resp.content)
 
-    def get_dataverses(self):
+    def get_dataverses(self, allow_root=False):
         # Get latest dataverse information
         self.connect()
 
-        # Dataverse 4.0 beta currently returns collection for "Harvard" DV,
-        # which the user may not have permissions on. Could cause problems.        
         collections = get_elements(
             self.service_document[0],
             tag="collection",
         )
-        # removes the Harvard dataverse; this is gross
-        # also removes the root dataverse (which users do not have read permission on)
-        def is_not_harvard_or_root(collection):
-            return (collection.attrib['href'].split('/')[-1] != 'harvard') and (collection.attrib['href'].split('/')[-1] != 'root')
-        collections = filter(is_not_harvard_or_root, collections)
+
+        # Remove root Dataverses, which may cause permission issues
+        # See https://github.com/IQSS/dataverse/issues/1070
+        if not allow_root:
+            collections = filter(is_not_root_dataverse, collections)
 
         return [Dataverse(self, col) for col in collections]
 

dataverse/test/test_dataverse.py

@@ -1,4 +1,3 @@
-import sys
 import unittest
 
 import logging
@@ -103,10 +102,6 @@ def setUpClass(self):
             )
 
         self.dv = dataverses[0]
-        if not self.dv.is_published:
-            raise utils.DataverseException(
-                'You must publish "{0}" to run these tests.'.format(self.dv.title)
-            )
 
         print "Removing any existing studies."
         studies = self.dv.get_studies()
@@ -199,12 +194,8 @@ def test_publish_dataset(self):
         self.assertTrue(self.s.get_state() == "PUBLISHED")
         self.dv.delete_dataset(self.s)
         self.assertTrue(self.s.get_state(refresh=True) == "DEACCESSIONED")
+
 
-    def test_dataverse_published(self):
-        self.assertTrue(self.dv.is_published)
-    
-if __name__ == "__main__":
-    __file__ = sys.argv[0]
-    suite = unittest.TestLoader().loadTestsFromTestCase(TestDatasetOperations)
-    unittest.TextTestRunner(verbosity=2).run(suite)
+if __name__ == '__main__':
+    unittest.main()
 

dataverse/utils.py

@@ -10,6 +10,12 @@ class DataverseException(Exception):
     pass
 
 
+def is_not_root_dataverse(collection):
+    col_alias = collection.attrib['href'].split('/')[-1]
+    # Root dataverse may be named differently per host
+    return col_alias != 'root' and col_alias != 'harvard'
+
+
 # factor out xpath operations so we don't have to look at its ugliness
 def get_element(root, tag='*', namespace=None, attribute=None, attribute_value=None):
     elements = get_elements(root, tag, namespace, attribute, attribute_value)

readme.md

@@ -28,7 +28,10 @@ Do not commit this file.
 
 ## Testing
 
-Not all tests are functional. 
-
-In order to run any tests, you must first create and publish a Dataverse on the
+In order to run any tests, you must first create a Dataverse on the
 host you wish to test. Do not run tests on the production server.
+
+To run tests
+
+    $ cd dataverse/test
+    $ python -m unittest test_dataverse