reduce pytest warnings (#3674)

 * Update docs for pytest.

 * Bring pytest warnings from 423 down to 31
   * fix usage of deprecated unittest api
   * fix tests skipped because of `__init__` constructor
   * fix classes incorrectly identified as tests
   * fix deprecated usage of external modules
   * fix incorrect escape sequences in regular expressions
   * ignore unnecessary warnings from external packages
     **Note:** this could be made more flexible (e.g. narrowing down filters), since such warnings can sometimes be useful, e.g. if we are using a deprecated API of the package (say, django). Most of the time, it is the external package that is using some deprecated API.

 * add individual test timings

.github/workflows/tests.sh

@@ -3,6 +3,8 @@ set -ev
 
 # Make sure the folder containing the workchains is in the python path before the daemon is started
 export PYTHONPATH="${PYTHONPATH}:${GITHUB_WORKSPACE}/.ci"
+# show timings of tests
+export PYTEST_ADDOPTS=" --durations=0"
 
 verdi daemon start 4
 verdi -p test_${AIIDA_TEST_BACKEND} run .ci/test_daemon.py

aiida/backends/tests/cmdline/commands/test_data.py

@@ -32,9 +32,6 @@
 class TestVerdiDataExportable:
     """Test exportable data objects."""
 
-    def __init__(self):
-        pass
-
     NODE_ID_STR = 'node_id'
     EMPTY_GROUP_ID_STR = 'empty_group_id'
     EMPTY_GROUP_NAME_STR = 'empty_group'
@@ -87,7 +84,7 @@ def data_export_test(self, datatype, ids, supported_formats):
                 # Try to export it again. It should fail because the
                 # file exists
                 res = self.cli_runner.invoke(export_cmd, options, catch_exceptions=False)
-                self.assertNotEquals(res.exit_code, 0, 'The command should fail because the file already exists')
+                self.assertNotEqual(res.exit_code, 0, 'The command should fail because the file already exists')
 
                 # Now we force the export of the file and it should overwrite
                 # existing files
@@ -104,9 +101,6 @@ def data_export_test(self, datatype, ids, supported_formats):
 class TestVerdiDataListable:
     """Test listable data objects."""
 
-    def __init__(self):
-        pass
-
     NODE_ID_STR = 'node_id'
     EMPTY_GROUP_ID_STR = 'empty_group_id'
     EMPTY_GROUP_NAME_STR = 'empty_group'

aiida/backends/tests/common/test_extendeddicts.py

@@ -19,14 +19,14 @@
 from aiida.common import extendeddicts
 
 
-class TestFFADExample(extendeddicts.FixedFieldsAttributeDict):
+class FFADExample(extendeddicts.FixedFieldsAttributeDict):
     """
     An example class that accepts only the 'alpha', 'beta' and 'gamma' keys/attributes.
     """
     _valid_fields = ('alpha', 'beta', 'gamma')
 
 
-class TestDFADExample(extendeddicts.DefaultFieldsAttributeDict):
+class DFADExample(extendeddicts.DefaultFieldsAttributeDict):
     """
     An example class that has 'alpha', 'beta' and 'gamma' as default keys.
     """
@@ -270,7 +270,7 @@ class TestFFAD(unittest.TestCase):
 
     def test_insertion(self):
         """Test insertion."""
-        dictionary = TestFFADExample()
+        dictionary = FFADExample()
         dictionary['alpha'] = 1
         dictionary.beta = 2
         # Not a valid key.
@@ -281,14 +281,14 @@ def test_insertion(self):
 
     def test_insertion_on_init(self):
         """Test insertion in constructor."""
-        TestFFADExample({'alpha': 1, 'beta': 2})
+        FFADExample({'alpha': 1, 'beta': 2})
         with self.assertRaises(KeyError):
             # 'delta' is not a valid key
-            TestFFADExample({'alpha': 1, 'delta': 2})
+            FFADExample({'alpha': 1, 'delta': 2})
 
     def test_pickle(self):
         """Note: pickle works here because self._valid_fields is defined at the class level!"""
-        dictionary_01 = TestFFADExample({'alpha': 1, 'beta': 2})
+        dictionary_01 = FFADExample({'alpha': 1, 'beta': 2})
         dictionary_02 = pickle.loads(pickle.dumps(dictionary_01))
         dictionary_02.gamma = 3
         with self.assertRaises(KeyError):
@@ -299,15 +299,15 @@ def test_class_attribute(self):
         I test that the get_valid_fields() is working as a class method,
         so I don't need to instantiate the class to get the list.
         """
-        self.assertEqual(set(TestFFADExample.get_valid_fields()), set(['alpha', 'beta', 'gamma']))
+        self.assertEqual(set(FFADExample.get_valid_fields()), set(['alpha', 'beta', 'gamma']))
 
 
 class TestDFAD(unittest.TestCase):
     """Test for the default fields attribute dictionary."""
 
     def test_insertion_and_retrieval(self):
         """Test insertion and retrieval."""
-        dictionary = TestDFADExample()
+        dictionary = DFADExample()
         dictionary['alpha'] = 1
         dictionary.beta = 2
         dictionary['delta'] = 3
@@ -319,7 +319,7 @@ def test_insertion_and_retrieval(self):
 
     def test_keylist_method(self):
         """Test keylist retrieval."""
-        dictionary = TestDFADExample()
+        dictionary = DFADExample()
         dictionary['alpha'] = 1
         dictionary.beta = 2
         dictionary['delta'] = 3
@@ -336,11 +336,11 @@ def test_class_attribute(self):
         I test that the get_default_fields() is working as a class method,
         so I don't need to instantiate the class to get the list.
         """
-        self.assertEqual(set(TestDFADExample.get_default_fields()), set(['alpha', 'beta', 'gamma']))
+        self.assertEqual(set(DFADExample.get_default_fields()), set(['alpha', 'beta', 'gamma']))
 
     def test_validation(self):
         """Test validation."""
-        dictionary = TestDFADExample()
+        dictionary = DFADExample()
 
         # Should be ok to have an empty 'alpha' attribute
         dictionary.validate()

aiida/backends/tests/engine/test_process_builder.py

@@ -9,7 +9,7 @@
 ###########################################################################
 """Module to test process builder."""
 # pylint: disable=no-member,protected-access,no-name-in-module
-from collections import Mapping, MutableMapping
+from collections.abc import Mapping, MutableMapping
 
 from aiida import orm
 from aiida.backends.testbase import AiidaTestCase
@@ -20,7 +20,7 @@
 DEFAULT_INT = 256
 
 
-class TestWorkChain(WorkChain):
+class ExampleWorkChain(WorkChain):
     """Defining test work chain."""
 
     @classmethod
@@ -34,7 +34,7 @@ def define(cls, spec):
         spec.input('default', valid_type=orm.Int, default=orm.Int(DEFAULT_INT).store())
 
 
-class TestLazyProcessNamespace(Process):
+class LazyProcessNamespace(Process):
     """Process with basic nested namespaces to test "pruning" of empty nested namespaces from the builder."""
 
     @classmethod
@@ -74,7 +74,7 @@ def setUp(self):
         self.assertIsNone(Process.current())
         self.process_class = CalculationFactory('templatereplacer')
         self.builder = self.process_class.get_builder()
-        self.builder_workchain = TestWorkChain.get_builder()
+        self.builder_workchain = ExampleWorkChain.get_builder()
         self.inputs = {
             'dynamic': {
                 'namespace': {
@@ -95,34 +95,34 @@ def tearDown(self):
 
     def test_builder_inputs(self):
         """Test the `ProcessBuilder._inputs` method to get the inputs with and without `prune` set to True."""
-        builder = TestLazyProcessNamespace.get_builder()
+        builder = LazyProcessNamespace.get_builder()
 
         # When no inputs are specified specifically, `prune=True` should get rid of completely empty namespaces
         self.assertEqual(builder._inputs(prune=False), {'namespace': {'nested': {}}, 'metadata': {}})
         self.assertEqual(builder._inputs(prune=True), {})
 
         # With a specific input in `namespace` the case of `prune=True` should now only remove `metadata`
         integer = orm.Int(DEFAULT_INT)
-        builder = TestLazyProcessNamespace.get_builder()
+        builder = LazyProcessNamespace.get_builder()
         builder.namespace.a = integer
         self.assertEqual(builder._inputs(prune=False), {'namespace': {'a': integer, 'nested': {}}, 'metadata': {}})
         self.assertEqual(builder._inputs(prune=True), {'namespace': {'a': integer}})
 
         # A value that is a `Node` instance but also happens to be an "empty mapping" should not be pruned
         empty_node = MappingData()
-        builder = TestLazyProcessNamespace.get_builder()
+        builder = LazyProcessNamespace.get_builder()
         builder.namespace.a = empty_node
         self.assertEqual(builder._inputs(prune=False), {'namespace': {'a': empty_node, 'nested': {}}, 'metadata': {}})
         self.assertEqual(builder._inputs(prune=True), {'namespace': {'a': empty_node}})
 
         # Verify that empty lists are considered as a "value" and are not pruned
-        builder = TestLazyProcessNamespace.get_builder()
+        builder = LazyProcessNamespace.get_builder()
         builder.namespace.c = list()
         self.assertEqual(builder._inputs(prune=False), {'namespace': {'c': [], 'nested': {}}, 'metadata': {}})
         self.assertEqual(builder._inputs(prune=True), {'namespace': {'c': []}})
 
         # Verify that empty lists, even in doubly nested namespace are considered as a "value" and are not pruned
-        builder = TestLazyProcessNamespace.get_builder()
+        builder = LazyProcessNamespace.get_builder()
         builder.namespace.nested.bird = list()
         self.assertEqual(builder._inputs(prune=False), {'namespace': {'nested': {'bird': []}}, 'metadata': {}})
         self.assertEqual(builder._inputs(prune=True), {'namespace': {'nested': {'bird': []}}})
@@ -172,15 +172,15 @@ def test_dynamic_getters_value(self):
 
     def test_dynamic_getters_doc_string(self):
         """Verify that getters have the correct docstring."""
-        builder = TestWorkChain.get_builder()
-        self.assertEqual(builder.__class__.name_spaced.__doc__, str(TestWorkChain.spec().inputs['name_spaced']))
-        self.assertEqual(builder.__class__.boolean.__doc__, str(TestWorkChain.spec().inputs['boolean']))
+        builder = ExampleWorkChain.get_builder()
+        self.assertEqual(builder.__class__.name_spaced.__doc__, str(ExampleWorkChain.spec().inputs['name_spaced']))
+        self.assertEqual(builder.__class__.boolean.__doc__, str(ExampleWorkChain.spec().inputs['boolean']))
 
     def test_builder_restart_work_chain(self):
         """Verify that nested namespaces imploded into flat link labels can be reconstructed into nested namespaces."""
         caller = orm.WorkChainNode().store()
 
-        node = orm.WorkChainNode(process_type=TestWorkChain.build_process_type())
+        node = orm.WorkChainNode(process_type=ExampleWorkChain.build_process_type())
         node.add_incoming(self.inputs['dynamic']['namespace']['alp'], LinkType.INPUT_WORK, 'dynamic__namespace__alp')
         node.add_incoming(self.inputs['name']['spaced'], LinkType.INPUT_WORK, 'name__spaced')
         node.add_incoming(self.inputs['name_spaced'], LinkType.INPUT_WORK, 'name_spaced')
@@ -211,7 +211,7 @@ def test_port_names_overlapping_mutable_mapping_methods(self):  # pylint: disabl
         as attributes, they can overlap with some of the mappings builtin methods, e.g. `values()`, `items()` etc.
         The port names should take precendence in this case and if one wants to access the mapping methods one needs to
         cast the builder to a dictionary first."""
-        builder = TestWorkChain.get_builder()
+        builder = ExampleWorkChain.get_builder()
 
         # The `values` method is obscured by a port that also happens to be called `values`, so calling it should raise
         with self.assertRaises(TypeError):
@@ -232,7 +232,7 @@ def test_port_names_overlapping_mutable_mapping_methods(self):  # pylint: disabl
             self.assertNotIn(method, dir(builder))
 
         # On the other hand, all the port names *should* be present
-        for port_name in TestWorkChain.spec().inputs.keys():
+        for port_name in ExampleWorkChain.spec().inputs.keys():
             self.assertIn(port_name, dir(builder))
 
         # The `update` method is implemented, but prefixed with an underscore to not block the name for a port