generalize tag extraction

Author: mfeurer

openml/flows/flow.py

@@ -4,6 +4,7 @@
 import xmltodict
 
 from .._api_calls import _perform_api_call
+from ..utils import extract_xml_tags
 
 
 class OpenMLFlow(object):
@@ -278,10 +279,8 @@ def _from_dict(cls, xml_dict):
         if 'oml:parameter' in dic:
             # In case of a single parameter, xmltodict returns a dictionary,
             # otherwise a list.
-            if isinstance(dic['oml:parameter'], dict):
-                oml_parameters = [dic['oml:parameter']]
-            else:
-                oml_parameters = dic['oml:parameter']
+            oml_parameters = extract_xml_tags('oml:parameter', dic,
+                                              allow_none=False)
 
             for oml_parameter in oml_parameters:
                 parameter_name = oml_parameter['oml:name']
@@ -299,16 +298,14 @@ def _from_dict(cls, xml_dict):
         if 'oml:component' in dic:
             # In case of a single component xmltodict returns a dict,
             # otherwise a list.
-            if isinstance(dic['oml:component'], dict):
-                oml_components = [dic['oml:component']]
-            else:
-                oml_components = dic['oml:component']
+            oml_components = extract_xml_tags('oml:component', dic,
+                                              allow_none=False)
 
             for component in oml_components:
                 flow = OpenMLFlow._from_dict(component)
                 components[component['oml:identifier']] = flow
         arguments['components'] = components
-        arguments['tags'] = extract_tags(dic)
+        arguments['tags'] = extract_xml_tags('oml:tag', dic)
 
         arguments['model'] = None
         flow = cls(**arguments)
@@ -375,15 +372,3 @@ def _add_if_nonempty(dic, key, value):
         dic[key] = value
 
 
-def extract_tags(dic):
-    if 'oml:tag' in dic and dic['oml:tag'] is not None:
-        if isinstance(dic['oml:tag'], six.string_types):
-            oml_tags = [dic['oml:tag']]
-        elif isinstance(dic['oml:tag'], list):
-            oml_tags = dic['oml:tag']
-        else:
-            raise ValueError('Received not string and non list as tag item')
-
-        return oml_tags
-    else:
-        return None

openml/runs/functions.py

@@ -12,6 +12,7 @@
 import xmltodict
 
 import openml
+import openml.utils
 from ..exceptions import PyOpenMLError
 from .. import config
 from ..flows import sklearn_to_flow, get_flow, flow_exists, _check_n_jobs, \
@@ -578,10 +579,13 @@ def _create_run_from_xml(xml):
             # only one result.. probably due to an upload error
             file_dict = run['oml:output_data']['oml:file']
             files[file_dict['oml:name']] = int(file_dict['oml:file_id'])
-        else:
+        elif isinstance(run['oml:output_data']['oml:file'], list):
             # multiple files, the normal case
             for file_dict in run['oml:output_data']['oml:file']:
                 files[file_dict['oml:name']] = int(file_dict['oml:file_id'])
+        else:
+            raise TypeError(type(run['oml:output_data']['oml:file']))
+
         if 'oml:evaluation' in run['oml:output_data']:
             # in normal cases there should be evaluations, but in case there
             # was an error these could be absent
@@ -614,14 +618,7 @@ def _create_run_from_xml(xml):
         raise ValueError('No prediction files for run %d in run '
                          'description XML' % run_id)
 
-    tags = None
-    if 'oml:tag' in run:
-        if isinstance(run['oml:tag'], six.string_types):
-            tags = [run['oml:tag']]
-        elif isinstance(run['oml:tag'], list):
-            tags = run['oml:tag']
-        else:
-            raise ValueError('Received not string and non list as tag item')
+    tags = openml.utils.extract_xml_tags('oml:tag', run)
 
     return OpenMLRun(run_id=run_id, uploader=uploader,
                      uploader_name=uploader_name, task_id=task_id,

openml/utils.py

@@ -0,0 +1,36 @@
+import six
+
+
+def extract_xml_tags(xml_tag_name, node, allow_none=True):
+    """Helper to extract xml tags from xmltodict.
+
+    Parameters
+    ----------
+    xml_tag_name : str
+        Name of the xml tag to extract from the node.
+
+    node : object
+        Node object returned by ``xmltodict`` from which ``xml_tag_name``
+        should be extracted.
+
+    Returns
+    -------
+    object
+    """
+    if xml_tag_name in node and node[xml_tag_name] is not None:
+        if isinstance(node[xml_tag_name], dict):
+            rval = [node[xml_tag_name]]
+        elif isinstance(node[xml_tag_name], six.string_types):
+            rval = [node[xml_tag_name]]
+        elif isinstance(node[xml_tag_name], list):
+            rval = node[xml_tag_name]
+        else:
+            raise ValueError('Received not string and non list as tag item')
+
+        return rval
+    else:
+        if allow_none:
+            return None
+        else:
+            raise ValueError("Could not find tag '%s' in node '%s'" %
+                             (xml_tag_name, str(node)))

tests/test_flows/test_flow.py

@@ -27,13 +27,13 @@
 from openml.testing import TestBase
 from openml._api_calls import _perform_api_call
 import openml
+import openml.utils
 from openml.flows.sklearn_converter import _format_external_version
 import openml.exceptions
 
 
 class TestFlow(TestBase):
 
-
     def test_get_flow(self):
         # We need to use the production server here because 4024 is not the test
         # server
@@ -324,11 +324,11 @@ def test_sklearn_to_upload_to_flow(self):
     def test_extract_tags(self):
         flow_xml = "<oml:tag>study_14</oml:tag>"
         flow_dict = xmltodict.parse(flow_xml)
-        tags = openml.flows.flow.extract_tags(flow_dict)
+        tags = openml.utils.extract_xml_tags('oml:tag', flow_dict)
         self.assertEqual(tags, ['study_14'])
 
         flow_xml = "<oml:flow><oml:tag>OpenmlWeka</oml:tag>\n" \
                    "<oml:tag>weka</oml:tag></oml:flow>"
         flow_dict = xmltodict.parse(flow_xml)
-        tags = openml.flows.flow.extract_tags(flow_dict['oml:flow'])
+        tags = openml.utils.extract_xml_tags('oml:tag', flow_dict['oml:flow'])
         self.assertEqual(tags, ['OpenmlWeka', 'weka'])