add function to check flow equality

Author: mfeurer

openml/flows/functions.py

@@ -127,4 +127,37 @@ def _list_flows(api_call):
                 'uploader': flow_['oml:uploader']}
         flows[fid] = flow
 
-    return flows
+    return flows
+
+
+def are_flows_equal(flow1, flow2):
+    """Check equality of two flows.
+
+    Two flows are equal if their all keys which are not set by the server
+    are equal, as well as all their parameters and components.
+    """
+    if not isinstance(flow2, flow1.__class__):
+        return False
+
+    # Name is actually not generated by the server, but it will be
+    # tested further down with a getter (allows mocking in the tests)
+    generated_by_the_server = ['flow_id', 'uploader', 'version',
+                               'upload_date', ]
+    ignored_by_python_API = ['binary_url', 'binary_format', 'binary_md5',
+                             'model']
+
+    for key in set(flow1.__dict__.keys()).union(flow2.__dict__.keys()):
+        if key in generated_by_the_server + ignored_by_python_API:
+            continue
+        attr1 = getattr(flow1, key, None)
+        attr2 = getattr(flow2, key, None)
+        if key == 'components':
+            for name in set(attr1.keys()).union(attr2.keys()):
+                if not (name in attr1 and name in attr2):
+                    return False
+                if not are_flows_equal(attr1[name], attr2[name]):
+                    return False
+        else:
+            if attr1 != attr2:
+                return False
+    return True

tests/test_flows/test_flow.py

@@ -2,8 +2,6 @@
 import hashlib
 import re
 import time
-import random
-import unittest
 
 import xmltodict
 
@@ -26,38 +24,6 @@
 from openml.flows.sklearn_converter import _format_external_version
 
 
-def are_flows_equal(flow1, flow2):
-    """Check equality of two flows.
-
-    Two flows are equal if their all keys which are not set by the server
-    are equal, as well as all their parameters and components.
-    """
-    if not isinstance(flow2, flow1.__class__):
-        return False
-
-    # Name is actually not generated by the server, but it will be
-    # tested further down with a getter (allows mocking in the tests)
-    generated_by_the_server = ['name', 'flow_id', 'uploader', 'version',
-                               'upload_date', 'source_url',
-                               'binary_url', 'source_format',
-                               'binary_format', 'source_md5',
-                               'binary_md5', 'model']
-
-    for key in set(flow1.__dict__.keys()).union(flow2.__dict__.keys()):
-        if key in generated_by_the_server:
-            continue
-        attr1 = getattr(flow1, key, None)
-        attr2 = getattr(flow2, key, None)
-        if key == 'components':
-            for name in set(attr1.keys()).union(attr2.keys()):
-                if not are_flows_equal(attr1[name], attr2[name]):
-                    return False
-        else:
-            if attr1 != attr2:
-                return False
-    return True
-
-
 def get_sentinel():
     # Create a unique prefix for the flow. Necessary because the flow is
     # identified by its name and external version online. Having a unique
@@ -134,7 +100,7 @@ def test_to_xml_from_xml(self):
         xml = flow._to_xml()
         xml_dict = xmltodict.parse(xml)
         new_flow = openml.flows.OpenMLFlow._from_dict(xml_dict)
-        self.assertTrue(are_flows_equal(new_flow, flow))
+        self.assertTrue(openml.flows.functions.are_flows_equal(new_flow, flow))
         self.assertIsNot(new_flow, flow)
 
     def test_publish_flow(self):
@@ -267,7 +233,7 @@ def test_sklearn_to_upload_to_flow(self):
 
         self.assertEqual(server_xml, local_xml)
 
-        self.assertTrue(are_flows_equal(new_flow, flow))
+        self.assertTrue(openml.flows.functions.are_flows_equal(new_flow, flow))
         self.assertIsNot(new_flow, flow)
 
         fixture_name = '%ssklearn.model_selection._search.RandomizedSearchCV(' \

tests/test_flows/test_flow_functions.py

@@ -1,3 +1,5 @@
+from collections import OrderedDict
+import copy
 import unittest
 
 import openml
@@ -39,4 +41,68 @@ def test_list_flows_paginate(self):
             flows = openml.flows.list_flows(offset=i, size=size)
             self.assertGreaterEqual(size, len(flows))
             for did in flows:
-                self._check_flow(flows[did])
+                self._check_flow(flows[did])
+
+    def test_are_flows_equal(self):
+        flow = openml.flows.OpenMLFlow(name='Test',
+                                       description='Test flow',
+                                       model=None,
+                                       components=OrderedDict(),
+                                       parameters=OrderedDict(),
+                                       parameters_meta_info=OrderedDict(),
+                                       external_version='1',
+                                       tags=['abc', 'def'],
+                                       language='English',
+                                       dependencies='abc',
+                                       class_name='Test',
+                                       custom_name='Test')
+
+        # Test most important values that can be set by a user
+        self.assertTrue(openml.flows.functions.are_flows_equal(flow, flow))
+        for attribute, new_value in [('name', 'Tes'),
+                                     ('description', 'Test flo'),
+                                     ('external_version', '2'),
+                                     ('tags', ['abc', 'de']),
+                                     ('language', 'english'),
+                                     ('dependencies', 'ab'),
+                                     ('class_name', 'Tes'),
+                                     ('custom_name', 'Tes')]:
+            new_flow = copy.deepcopy(flow)
+            setattr(new_flow, attribute, new_value)
+            self.assertNotEqual(getattr(flow, attribute), getattr(new_flow, attribute))
+            self.assertFalse(openml.flows.functions.are_flows_equal(flow, new_flow),
+                             msg=attribute)
+
+        # Test that the API ignores several keys when comparing flows
+        self.assertTrue(openml.flows.functions.are_flows_equal(flow, flow))
+        for attribute, new_value in [('flow_id', 1),
+                                     ('uploader', 1),
+                                     ('version', 1),
+                                     ('upload_date', '18.12.1988'),
+                                     ('binary_url', 'openml.org'),
+                                     ('binary_format', 'gzip'),
+                                     ('binary_md5', '12345'),
+                                     ('model', [])]:
+            new_flow = copy.deepcopy(flow)
+            setattr(new_flow, attribute, new_value)
+            self.assertNotEqual(getattr(flow, attribute), getattr(new_flow, attribute))
+            self.assertTrue(openml.flows.functions.are_flows_equal(flow, new_flow),
+                            msg=attribute)
+
+        # Now test for parameters
+        flow.parameters['abc'] = 1.0
+        flow.parameters['def'] = 2.0
+        self.assertTrue(openml.flows.functions.are_flows_equal(flow, flow))
+        new_flow = copy.deepcopy(flow)
+        new_flow.parameters['abc'] = 2.0
+        self.assertFalse(openml.flows.functions.are_flows_equal(flow, new_flow))
+
+        # Now test for components (subflows)
+        parent_flow = copy.deepcopy(flow)
+        subflow = copy.deepcopy(flow)
+        parent_flow.components['subflow'] = subflow
+        self.assertTrue(openml.flows.functions.are_flows_equal(parent_flow, parent_flow))
+        self.assertFalse(openml.flows.functions.are_flows_equal(parent_flow, subflow))
+        new_flow = copy.deepcopy(parent_flow)
+        new_flow.components['subflow'].name = 'Subflow name'
+        self.assertFalse(openml.flows.functions.are_flows_equal(parent_flow, new_flow))