🧪 test_fields: assert structure, not repr

tests/orm/test_fields.py

@@ -8,25 +8,66 @@
 ###########################################################################
 """Test for entity fields"""
 
+import typing as t
 from importlib.metadata import entry_points
 
 import pytest
 
 from aiida import orm
-from aiida.orm.fields import add_field
+from aiida.orm.fields import QbField, add_field
 from aiida.orm.pydantic import OrmMetadataField
 from aiida.plugins import load_entry_point
 
 EPS = entry_points()
 
 
+def format_dtype(dtype: t.Any) -> str | None:
+    """Return a stable, human-readable rendering of a field dtype.
+
+    Plain classes are rendered by qualified name, everything else by ``str`` with the ``typing.`` prefix stripped.
+    The ``get_origin`` guard is required: on Python 3.10 ``isinstance(dict[str, Any], type)`` is ``True``, so without
+    it a parametrized generic would render as a bare ``dict`` there and as ``dict[str, Any]`` from 3.11 on.
+    """
+    if dtype is None:
+        return None
+    if t.get_origin(dtype) is None and isinstance(dtype, type):
+        return dtype.__qualname__
+    return str(dtype).replace('typing.', '')
+
+
+def field_spec(field: QbField) -> str:
+    """Return a one-line description of a field, for comparison against a reference file.
+
+    The field key is not included: it is the mapping key in the reference file. Everything else is only emitted when
+    it carries information, to keep the files readable:
+
+    * ``backend_key`` only when it differs from the key, which is exactly when ``is_attribute`` is set, since a
+      non-attribute field may not be aliased and so always stores under its own key;
+    * ``root``, the ``extract_root_type`` of the dtype, only when it differs from the dtype itself. It selects the
+      ``QbField`` subclass and is otherwise only covered implicitly, via that choice.
+    """
+    parts = []
+    if field.backend_key != field.key:
+        parts.append(f'backend_key={field.backend_key}')
+    if field._is_attribute:
+        parts.append('is_attribute=True')
+    dtype = format_dtype(field._dtype)
+    root = format_dtype(field.dtype)
+    parts.append(f'dtype={dtype}')
+    if root != dtype:
+        parts.append(f'root={root}')
+    if field._doc:
+        parts.append(f'doc={field._doc!r}')
+    return f'{type(field).__name__}({", ".join(parts)})'
+
+
 @pytest.mark.parametrize(
     'entity_cls',
     (orm.AuthInfo, orm.Comment, orm.Computer, orm.Group, orm.Log, orm.User),
 )
 def test_all_entity_fields(entity_cls, data_regression):
     data_regression.check(
-        {key: repr(value) for key, value in entity_cls.fields._dict.items()},
+        {key: field_spec(value) for key, value in entity_cls.fields._dict.items()},
         basename=f'fields_{entity_cls.__name__}',
     )
 
@@ -46,7 +87,7 @@ def test_all_node_fields(node_and_data_entry_points: list[tuple[str, str]], data
     for group, name in node_and_data_entry_points:
         node_cls = load_entry_point(group, name)
         data_regression.check(
-            {key: repr(value) for key, value in node_cls.fields._dict.items()},
+            {key: field_spec(value) for key, value in node_cls.fields._dict.items()},
             basename=f'fields_{group}.{name}.{node_cls.__name__}',
         )
 

tests/orm/test_fields/fields_AuthInfo.yml

@@ -1,9 +1,7 @@
-auth_params: QbDictField('auth_params', dtype=dict[str, typing.Any], doc='Dictionary
-  of authentication parameters')
-computer: QbNumericField('computer', dtype=<class 'int'>, doc='The PK of the computer')
-enabled: QbAnyField('enabled', dtype=<class 'bool'>, doc='Whether the instance is
-  enabled')
-metadata: QbDictField('metadata', dtype=dict[str, typing.Any], doc='Dictionary of
-  metadata')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
-user: QbNumericField('user', dtype=<class 'int'>, doc='The PK of the user')
+auth_params: QbDictField(dtype=dict[str, Any], root=dict, doc='Dictionary of authentication
+  parameters')
+computer: QbNumericField(dtype=int, doc='The PK of the computer')
+enabled: QbAnyField(dtype=bool, doc='Whether the instance is enabled')
+metadata: QbDictField(dtype=dict[str, Any], root=dict, doc='Dictionary of metadata')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')
+user: QbNumericField(dtype=int, doc='The PK of the user')

tests/orm/test_fields/fields_Comment.yml

@@ -1,10 +1,7 @@
-content: QbStrField('content', dtype=<class 'str'>, doc='Content of the comment')
-ctime: QbNumericField('ctime', dtype=<class 'datetime.datetime'>, doc='Creation time
-  of the comment')
-mtime: QbNumericField('mtime', dtype=<class 'datetime.datetime'>, doc='Modified time
-  of the comment')
-node: QbNumericField('node', dtype=<class 'int'>, doc='Node PK that the comment is
-  attached to')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
-user: QbNumericField('user', dtype=<class 'int'>, doc='User PK that created the comment')
-uuid: QbAnyField('uuid', dtype=<class 'uuid.UUID'>, doc='The UUID of the comment')
+content: QbStrField(dtype=str, doc='Content of the comment')
+ctime: QbNumericField(dtype=datetime, doc='Creation time of the comment')
+mtime: QbNumericField(dtype=datetime, doc='Modified time of the comment')
+node: QbNumericField(dtype=int, doc='Node PK that the comment is attached to')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')
+user: QbNumericField(dtype=int, doc='User PK that created the comment')
+uuid: QbAnyField(dtype=UUID, doc='The UUID of the comment')

tests/orm/test_fields/fields_Computer.yml

@@ -1,12 +1,8 @@
-description: QbStrField('description', dtype=<class 'str'>, doc='Description of the
-  computer')
-hostname: QbStrField('hostname', dtype=<class 'str'>, doc='Hostname of the computer')
-label: QbStrField('label', dtype=<class 'str'>, doc='Label for the computer')
-metadata: QbDictField('metadata', dtype=dict[str, typing.Any], doc='Metadata of the
-  computer')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
-scheduler_type: QbStrField('scheduler_type', dtype=<class 'str'>, doc='Scheduler type
-  of the computer')
-transport_type: QbStrField('transport_type', dtype=<class 'str'>, doc='Transport type
-  of the computer')
-uuid: QbAnyField('uuid', dtype=<class 'uuid.UUID'>, doc='The UUID of the computer')
+description: QbStrField(dtype=str, doc='Description of the computer')
+hostname: QbStrField(dtype=str, doc='Hostname of the computer')
+label: QbStrField(dtype=str, doc='Label for the computer')
+metadata: QbDictField(dtype=dict[str, Any], root=dict, doc='Metadata of the computer')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')
+scheduler_type: QbStrField(dtype=str, doc='Scheduler type of the computer')
+transport_type: QbStrField(dtype=str, doc='Transport type of the computer')
+uuid: QbAnyField(dtype=UUID, doc='The UUID of the computer')

tests/orm/test_fields/fields_Group.yml

@@ -1,9 +1,9 @@
-description: QbStrField('description', dtype=<class 'str'>, doc='The group description')
-extras: QbDictField('extras', dtype=dict[str, typing.Any], doc='The group extras')
-label: QbStrField('label', dtype=<class 'str'>, doc='The group label')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
-time: QbNumericField('time', dtype=<class 'datetime.datetime'>, doc='The creation
-  time of the node, defaults to now (timezone-aware)')
-type_string: QbStrField('type_string', dtype=<class 'str'>, doc='The type of the group')
-user: QbNumericField('user', dtype=<class 'int'>, doc='The PK of the group owner')
-uuid: QbAnyField('uuid', dtype=<class 'uuid.UUID'>, doc='The UUID of the group')
+description: QbStrField(dtype=str, doc='The group description')
+extras: QbDictField(dtype=dict[str, Any], root=dict, doc='The group extras')
+label: QbStrField(dtype=str, doc='The group label')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')
+time: QbNumericField(dtype=datetime, doc='The creation time of the node, defaults
+  to now (timezone-aware)')
+type_string: QbStrField(dtype=str, doc='The type of the group')
+user: QbNumericField(dtype=int, doc='The PK of the group owner')
+uuid: QbAnyField(dtype=UUID, doc='The UUID of the group')

tests/orm/test_fields/fields_Log.yml

@@ -1,10 +1,8 @@
-levelname: QbStrField('levelname', dtype=<class 'str'>, doc='The name of the log level')
-loggername: QbStrField('loggername', dtype=<class 'str'>, doc='The name of the logger')
-message: QbStrField('message', dtype=<class 'str'>, doc='The message of the log')
-metadata: QbDictField('metadata', dtype=dict[str, typing.Any], doc='The metadata of
-  the log')
-node: QbNumericField('node', dtype=<class 'int'>, doc='Associated node')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
-time: QbNumericField('time', dtype=<class 'datetime.datetime'>, doc='The time at which
-  the log was created')
-uuid: QbAnyField('uuid', dtype=<class 'uuid.UUID'>, doc='The UUID of the node')
+levelname: QbStrField(dtype=str, doc='The name of the log level')
+loggername: QbStrField(dtype=str, doc='The name of the logger')
+message: QbStrField(dtype=str, doc='The message of the log')
+metadata: QbDictField(dtype=dict[str, Any], root=dict, doc='The metadata of the log')
+node: QbNumericField(dtype=int, doc='Associated node')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')
+time: QbNumericField(dtype=datetime, doc='The time at which the log was created')
+uuid: QbAnyField(dtype=UUID, doc='The UUID of the node')

tests/orm/test_fields/fields_User.yml

@@ -1,5 +1,5 @@
-email: QbStrField('email', dtype=<class 'str'>, doc='The user email')
-first_name: QbStrField('first_name', dtype=<class 'str'>, doc='The user first name')
-institution: QbStrField('institution', dtype=<class 'str'>, doc='The user institution')
-last_name: QbStrField('last_name', dtype=<class 'str'>, doc='The user last name')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
+email: QbStrField(dtype=str, doc='The user email')
+first_name: QbStrField(dtype=str, doc='The user first name')
+institution: QbStrField(dtype=str, doc='The user institution')
+last_name: QbStrField(dtype=str, doc='The user last name')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')

tests/orm/test_fields/fields_aiida.data.core.array.ArrayData.yml

@@ -1,21 +1,18 @@
-attributes: QbAttributesField('attributes', dtype=<class 'aiida.orm.nodes.data.array.array.ArrayData.AttributesModel'>,
-  doc='The node attributes')
-computer: QbNumericField('computer', dtype=int | None, doc='The PK of the computer')
-ctime: QbNumericField('ctime', dtype=<class 'datetime.datetime'>, doc='The creation
-  time of the node')
-description: QbStrField('description', dtype=<class 'str'>, doc='The node description')
-extras: QbDictField('extras', dtype=dict[str, typing.Any], doc='The node extras')
-label: QbStrField('label', dtype=<class 'str'>, doc='The node label')
-mtime: QbNumericField('mtime', dtype=<class 'datetime.datetime'>, doc='The modification
-  time of the node')
-node_type: QbStrField('node_type', dtype=typing.Literal['data.core.array.ArrayData.'],
-  doc='The type of the node.')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
-process_type: QbStrField('process_type', dtype=str | None, doc='The process type of
-  the node')
-repository_metadata: QbDictField('repository_metadata', dtype=dict[str, typing.Any],
-  doc='Virtual hierarchy of the file repository')
-source: QbDictField('attributes.source', dtype=dict | None, doc='Source of the data')
-user: QbNumericField('user', dtype=<class 'int'>, doc='The PK of the user who owns
-  the node')
-uuid: QbAnyField('uuid', dtype=<class 'uuid.UUID'>, doc='The UUID of the node')
+attributes: QbAttributesField(dtype=ArrayData.AttributesModel, doc='The node attributes')
+computer: QbNumericField(dtype=int | None, root=int, doc='The PK of the computer')
+ctime: QbNumericField(dtype=datetime, doc='The creation time of the node')
+description: QbStrField(dtype=str, doc='The node description')
+extras: QbDictField(dtype=dict[str, Any], root=dict, doc='The node extras')
+label: QbStrField(dtype=str, doc='The node label')
+mtime: QbNumericField(dtype=datetime, doc='The modification time of the node')
+node_type: QbStrField(dtype=Literal['data.core.array.ArrayData.'], root=Literal, doc='The
+  type of the node.')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')
+process_type: QbStrField(dtype=str | None, root=str, doc='The process type of the
+  node')
+repository_metadata: QbDictField(dtype=dict[str, Any], root=dict, doc='Virtual hierarchy
+  of the file repository')
+source: QbDictField(backend_key=attributes.source, is_attribute=True, dtype=dict |
+  None, root=dict, doc='Source of the data')
+user: QbNumericField(dtype=int, doc='The PK of the user who owns the node')
+uuid: QbAnyField(dtype=UUID, doc='The UUID of the node')

tests/orm/test_fields/fields_aiida.data.core.array.bands.BandsData.yml

@@ -1,40 +1,38 @@
-array_labels: QbArrayField('attributes.array_labels', dtype=list[str] | None, doc='Labels
-  associated with the band arrays')
-attributes: QbAttributesField('attributes', dtype=<class 'aiida.orm.nodes.data.array.bands.BandsData.AttributesModel'>,
-  doc='The node attributes')
-cell: QbArrayField('attributes.cell', dtype=list[list[float]] | None, doc='Unit cell
-  of the crystal, in Angstroms')
-computer: QbNumericField('computer', dtype=int | None, doc='The PK of the computer')
-ctime: QbNumericField('ctime', dtype=<class 'datetime.datetime'>, doc='The creation
-  time of the node')
-description: QbStrField('description', dtype=<class 'str'>, doc='The node description')
-extras: QbDictField('extras', dtype=dict[str, typing.Any], doc='The node extras')
-label: QbStrField('label', dtype=<class 'str'>, doc='The node label')
-label_numbers: QbArrayField('attributes.label_numbers', dtype=list[int] | None, doc='Index
-  of the labels in the list of kpoints')
-labels: QbArrayField('attributes.labels', dtype=list[str] | None, doc='Labels associated
-  with the list of kpoints')
-mesh: QbArrayField('attributes.mesh', dtype=list[int] | None, doc='Mesh of kpoints')
-mtime: QbNumericField('mtime', dtype=<class 'datetime.datetime'>, doc='The modification
-  time of the node')
-node_type: QbStrField('node_type', dtype=typing.Literal['data.core.array.bands.BandsData.'],
+array_labels: QbArrayField(backend_key=attributes.array_labels, is_attribute=True,
+  dtype=list[str] | None, root=list, doc='Labels associated with the band arrays')
+attributes: QbAttributesField(dtype=BandsData.AttributesModel, doc='The node attributes')
+cell: QbArrayField(backend_key=attributes.cell, is_attribute=True, dtype=list[list[float]]
+  | None, root=list, doc='Unit cell of the crystal, in Angstroms')
+computer: QbNumericField(dtype=int | None, root=int, doc='The PK of the computer')
+ctime: QbNumericField(dtype=datetime, doc='The creation time of the node')
+description: QbStrField(dtype=str, doc='The node description')
+extras: QbDictField(dtype=dict[str, Any], root=dict, doc='The node extras')
+label: QbStrField(dtype=str, doc='The node label')
+label_numbers: QbArrayField(backend_key=attributes.label_numbers, is_attribute=True,
+  dtype=list[int] | None, root=list, doc='Index of the labels in the list of kpoints')
+labels: QbArrayField(backend_key=attributes.labels, is_attribute=True, dtype=list[str]
+  | None, root=list, doc='Labels associated with the list of kpoints')
+mesh: QbArrayField(backend_key=attributes.mesh, is_attribute=True, dtype=list[int]
+  | None, root=list, doc='Mesh of kpoints')
+mtime: QbNumericField(dtype=datetime, doc='The modification time of the node')
+node_type: QbStrField(dtype=Literal['data.core.array.bands.BandsData.'], root=Literal,
   doc='The type of the node.')
-offset: QbArrayField('attributes.offset', dtype=list[float] | None, doc='Offset of
-  kpoints')
-pbc1: QbAnyField('attributes.pbc1', dtype=bool | None, doc='Periodicity in the first
-  lattice vector direction')
-pbc2: QbAnyField('attributes.pbc2', dtype=bool | None, doc='Periodicity in the second
-  lattice vector direction')
-pbc3: QbAnyField('attributes.pbc3', dtype=bool | None, doc='Periodicity in the third
-  lattice vector direction')
-pk: QbNumericField('pk', dtype=<class 'int'>, doc='The primary key of the entity')
-process_type: QbStrField('process_type', dtype=str | None, doc='The process type of
-  the node')
-repository_metadata: QbDictField('repository_metadata', dtype=dict[str, typing.Any],
-  doc='Virtual hierarchy of the file repository')
-source: QbDictField('attributes.source', dtype=dict | None, doc='Source of the data')
-units: QbStrField('attributes.units', dtype=str | None, doc='Units in which the data
-  in bands were stored')
-user: QbNumericField('user', dtype=<class 'int'>, doc='The PK of the user who owns
-  the node')
-uuid: QbAnyField('uuid', dtype=<class 'uuid.UUID'>, doc='The UUID of the node')
+offset: QbArrayField(backend_key=attributes.offset, is_attribute=True, dtype=list[float]
+  | None, root=list, doc='Offset of kpoints')
+pbc1: QbAnyField(backend_key=attributes.pbc1, is_attribute=True, dtype=bool | None,
+  root=bool, doc='Periodicity in the first lattice vector direction')
+pbc2: QbAnyField(backend_key=attributes.pbc2, is_attribute=True, dtype=bool | None,
+  root=bool, doc='Periodicity in the second lattice vector direction')
+pbc3: QbAnyField(backend_key=attributes.pbc3, is_attribute=True, dtype=bool | None,
+  root=bool, doc='Periodicity in the third lattice vector direction')
+pk: QbNumericField(dtype=int, doc='The primary key of the entity')
+process_type: QbStrField(dtype=str | None, root=str, doc='The process type of the
+  node')
+repository_metadata: QbDictField(dtype=dict[str, Any], root=dict, doc='Virtual hierarchy
+  of the file repository')
+source: QbDictField(backend_key=attributes.source, is_attribute=True, dtype=dict |
+  None, root=dict, doc='Source of the data')
+units: QbStrField(backend_key=attributes.units, is_attribute=True, dtype=str | None,
+  root=str, doc='Units in which the data in bands were stored')
+user: QbNumericField(dtype=int, doc='The PK of the user who owns the node')
+uuid: QbAnyField(dtype=UUID, doc='The UUID of the node')