♻️ REFACTOR: NodeRepositoryMixin -> NodeRepository

.pre-commit-config.yaml

@@ -97,6 +97,7 @@ repos:
                 aiida/orm/nodes/data/jsonable.py|
                 aiida/orm/nodes/node.py|
                 aiida/orm/nodes/process/.*py|
+                aiida/orm/nodes/repository.py|
                 aiida/orm/utils/links.py|
                 aiida/plugins/entry_point.py|
                 aiida/plugins/factories.py|

aiida/cmdline/commands/cmd_calcjob.py

@@ -111,7 +111,7 @@ def calcjob_inputcat(calcjob, path):
 
     try:
         # When we `cat`, it makes sense to directly send the output to stdout as it is
-        with calcjob.open(path, mode='rb') as fhandle:
+        with calcjob.base.repository.open(path, mode='rb') as fhandle:
             copyfileobj(fhandle, sys.stdout.buffer)
     except OSError as exception:
         # The sepcial case is breakon pipe error, which is usually OK.
@@ -163,7 +163,7 @@ def calcjob_outputcat(calcjob, path):
 
     try:
         # When we `cat`, it makes sense to directly send the output to stdout as it is
-        with retrieved.open(path, mode='rb') as fhandle:
+        with retrieved.base.repository.open(path, mode='rb') as fhandle:
             copyfileobj(fhandle, sys.stdout.buffer)
     except OSError as exception:
         # The sepcial case is breakon pipe error, which is usually OK.

aiida/cmdline/commands/cmd_node.py

@@ -44,7 +44,7 @@ def repo_cat(node, relative_path):
     import sys
 
     try:
-        with node.open(relative_path, mode='rb') as fhandle:
+        with node.base.repository.open(relative_path, mode='rb') as fhandle:
             copyfileobj(fhandle, sys.stdout.buffer)
     except OSError as exception:
         # The sepcial case is breakon pipe error, which is usually OK.
@@ -96,7 +96,7 @@ def _copy_tree(key, output_dir):  # pylint: disable=too-many-branches
         Recursively copy the content at the ``key`` path in the given node to
         the ``output_dir``.
         """
-        for file in node.list_objects(key):
+        for file in node.base.repository.list_objects(key):
             # Not using os.path.join here, because this is the "path"
             # in the AiiDA node, not an actual OS - level path.
             file_key = file.name if not key else f'{key}/{file.name}'
@@ -110,7 +110,7 @@ def _copy_tree(key, output_dir):  # pylint: disable=too-many-branches
                 assert file.file_type == FileType.FILE
                 out_file_path = output_dir / file.name
                 assert not out_file_path.exists()
-                with node.open(file_key, 'rb') as in_file:
+                with node.base.repository.open(file_key, 'rb') as in_file:
                     with out_file_path.open('wb') as out_file:
                         shutil.copyfileobj(in_file, out_file)
 

aiida/cmdline/utils/repository.py

@@ -20,6 +20,6 @@ def list_repository_contents(node, path, color):
     """
     from aiida.repository import FileType
 
-    for entry in node.list_objects(path):
+    for entry in node.base.repository.list_objects(path):
         bold = bool(entry.file_type == FileType.DIRECTORY)
         echo.echo(entry.name, bold=bold, fg='blue' if color and entry.file_type == FileType.DIRECTORY else None)

aiida/engine/daemon/execmanager.py

@@ -175,12 +175,12 @@ def upload_calculation(
     for code in input_codes:
         if code.is_local():
             # Note: this will possibly overwrite files
-            for filename in code.list_object_names():
+            for filename in code.base.repository.list_object_names():
                 # Note, once #2579 is implemented, use the `node.open` method instead of the named temporary file in
                 # combination with the new `Transport.put_object_from_filelike`
                 # Since the content of the node could potentially be binary, we read the raw bytes and pass them on
                 with NamedTemporaryFile(mode='wb+') as handle:
-                    handle.write(code.get_object_content(filename, mode='rb'))
+                    handle.write(code.base.repository.get_object_content(filename, mode='rb'))
                     handle.flush()
                     transport.put(handle.name, filename)
             transport.chmod(code.get_local_executable(), 0o755)  # rwxr-xr-x
@@ -212,14 +212,14 @@ def upload_calculation(
             filepath_target = pathlib.Path(folder.abspath) / filename_target
             filepath_target.parent.mkdir(parents=True, exist_ok=True)
 
-            if data_node.get_object(filename_source).file_type == FileType.DIRECTORY:
+            if data_node.base.repository.get_object(filename_source).file_type == FileType.DIRECTORY:
                 # If the source object is a directory, we copy its entire contents
-                data_node.copy_tree(filepath_target, filename_source)
-                provenance_exclude_list.extend(data_node.list_object_names(filename_source))
+                data_node.base.repository.copy_tree(filepath_target, filename_source)
+                provenance_exclude_list.extend(data_node.base.repository.list_object_names(filename_source))
             else:
                 # Otherwise, simply copy the file
                 with folder.open(target, 'wb') as handle:
-                    with data_node.open(filename, 'rb') as source:
+                    with data_node.base.repository.open(filename, 'rb') as source:
                         shutil.copyfileobj(source, handle)
 
                 provenance_exclude_list.append(target)
@@ -320,12 +320,12 @@ def upload_calculation(
                 dirname not in provenance_exclude_list for dirname in dirnames
             ):
                 with open(filepath, 'rb') as handle:  # type: ignore[assignment]
-                    node._repository.put_object_from_filelike(handle, relpath)  # pylint: disable=protected-access
+                    node.base.repository._repository.put_object_from_filelike(handle, relpath)  # pylint: disable=protected-access
 
     # Since the node is already stored, we cannot use the normal repository interface since it will raise a
     # `ModificationNotAllowed` error. To bypass it, we go straight to the underlying repository instance to store the
     # files, however, this means we have to manually update the node's repository metadata.
-    node._update_repository_metadata()  # pylint: disable=protected-access
+    node.base.repository._update_repository_metadata()  # pylint: disable=protected-access
 
     if not dry_run:
         # Make sure that attaching the `remote_folder` with a link is the last thing we do. This gives the biggest
@@ -465,7 +465,7 @@ def retrieve_calculation(calculation: CalcJobNode, transport: Transport, retriev
         with SandboxFolder() as folder:
             retrieve_files_from_list(calculation, transport, folder.abspath, retrieve_list)
             # Here I retrieved everything; now I store them inside the calculation
-            retrieved_files.put_object_from_tree(folder.abspath)
+            retrieved_files.base.repository.put_object_from_tree(folder.abspath)
 
         # Retrieve the temporary files in the retrieved_temporary_folder if any files were
         # specified in the 'retrieve_temporary_list' key

aiida/engine/processes/calcjobs/calcjob.py

@@ -519,7 +519,7 @@ def parse_scheduler_output(self, retrieved: orm.Node) -> Optional[ExitCode]:
             self.logger.warning('could not determine `stderr` filename because `scheduler_stderr` option was not set.')
         else:
             try:
-                scheduler_stderr = retrieved.get_object_content(filename_stderr)
+                scheduler_stderr = retrieved.base.repository.get_object_content(filename_stderr)
             except FileNotFoundError:
                 scheduler_stderr = None
                 self.logger.warning(f'could not parse scheduler output: the `{filename_stderr}` file is missing')
@@ -528,7 +528,7 @@ def parse_scheduler_output(self, retrieved: orm.Node) -> Optional[ExitCode]:
             self.logger.warning('could not determine `stdout` filename because `scheduler_stdout` option was not set.')
         else:
             try:
-                scheduler_stdout = retrieved.get_object_content(filename_stdout)
+                scheduler_stdout = retrieved.base.repository.get_object_content(filename_stdout)
             except FileNotFoundError:
                 scheduler_stdout = None
                 self.logger.warning(f'could not parse scheduler output: the `{filename_stdout}` file is missing')

aiida/orm/__init__.py

@@ -72,7 +72,7 @@
     'Node',
     'NodeEntityLoader',
     'NodeLinksManager',
-    'NodeRepositoryMixin',
+    'NodeRepository',
     'NumericType',
     'OrbitalData',
     'OrderSpecifier',

aiida/orm/implementation/storage_backend.py

@@ -40,7 +40,7 @@ class StorageBackend(abc.ABC):  # pylint: disable=too-many-public-methods
     - Searchable data, which is stored in the database and can be queried using the QueryBuilder
     - Non-searchable (binary) data, which is stored in the repository and can be loaded using the RepositoryBackend
 
-    The two sources are inter-linked by the ``Node.repository_metadata``.
+    The two sources are inter-linked by the ``Node.base.repository.metadata``.
     Once stored, the leaf values of this dictionary must be valid pointers to object keys in the repository.
 
     The class methods,`version_profile` and `migrate`,

aiida/orm/nodes/__init__.py

@@ -40,7 +40,7 @@
     'KpointsData',
     'List',
     'Node',
-    'NodeRepositoryMixin',
+    'NodeRepository',
     'NumericType',
     'OrbitalData',
     'ProcessNode',

aiida/orm/nodes/data/array/array.py

@@ -45,11 +45,11 @@ def delete_array(self, name):
         :param name: The name of the array to delete from the node.
         """
         fname = f'{name}.npy'
-        if fname not in self.list_object_names():
+        if fname not in self.base.repository.list_object_names():
             raise KeyError(f"Array with name '{name}' not found in node pk= {self.pk}")
 
         # remove both file and attribute
-        self.delete_object(fname)
+        self.base.repository.delete_object(fname)
         try:
             self.delete_attribute(f'{self.array_prefix}{name}')
         except (KeyError, AttributeError):
@@ -71,7 +71,7 @@ def _arraynames_from_files(self):
         Return a list of all arrays stored in the node, listing the files (and
         not relying on the properties).
         """
-        return [i[:-4] for i in self.list_object_names() if i.endswith('.npy')]
+        return [i[:-4] for i in self.base.repository.list_object_names() if i.endswith('.npy')]
 
     def _arraynames_from_properties(self):
         """
@@ -111,11 +111,11 @@ def get_array_from_file(self, name):
             """Return the array stored in a .npy file"""
             filename = f'{name}.npy'
 
-            if filename not in self.list_object_names():
+            if filename not in self.base.repository.list_object_names():
                 raise KeyError(f'Array with name `{name}` not found in ArrayData<{self.pk}>')
 
             # Open a handle in binary read mode as the arrays are written as binary files as well
-            with self.open(filename, mode='rb') as handle:
+            with self.base.repository.open(filename, mode='rb') as handle:
                 return numpy.load(handle, allow_pickle=False)  # pylint: disable=unexpected-keyword-arg
 
         # Return with proper caching if the node is stored, otherwise always re-read from disk
@@ -171,7 +171,7 @@ def set_array(self, name, array):
             handle.seek(0)
 
             # Write the numpy array to the repository, keeping the byte representation
-            self.put_object_from_filelike(handle, f'{name}.npy')
+            self.base.repository.put_object_from_filelike(handle, f'{name}.npy')
 
         # Store the array name and shape for querying purposes
         self.set_attribute(f'{self.array_prefix}{name}', list(array.shape))

aiida/orm/nodes/data/code.py

@@ -93,7 +93,7 @@ def set_files(self, files):
         for filename in files:
             if os.path.isfile(filename):
                 with open(filename, 'rb') as handle:
-                    self.put_object_from_filelike(handle, os.path.split(filename)[1])
+                    self.base.repository.put_object_from_filelike(handle, os.path.split(filename)[1])
 
     def __str__(self):
         local_str = 'Local' if self.is_local() else 'Remote'
@@ -282,12 +282,12 @@ def _validate(self):
                     'You have to set which file is the local executable '
                     'using the set_exec_filename() method'
                 )
-            if self.get_local_executable() not in self.list_object_names():
+            if self.get_local_executable() not in self.base.repository.list_object_names():
                 raise exceptions.ValidationError(
                     f"The local executable '{self.get_local_executable()}' is not in the list of files of this code"
                 )
         else:
-            if self.list_object_names():
+            if self.base.repository.list_object_names():
                 raise exceptions.ValidationError('The code is remote but it has files inside')
             if not self.get_remote_computer():
                 raise exceptions.ValidationError('You did not specify a remote computer')

aiida/orm/nodes/data/data.py

@@ -69,8 +69,8 @@ def clone(self):
 
         backend_clone = self.backend_entity.clone()
         clone = self.__class__.from_backend_entity(backend_clone)
-        clone.reset_attributes(copy.deepcopy(self.attributes))  # pylint: disable=no-member
-        clone._repository.clone(self._repository)  # pylint: disable=no-member,protected-access
+        clone.reset_attributes(copy.deepcopy(self.attributes))
+        clone.base.repository._clone(self.base.repository)  # pylint: disable=protected-access
 
         return clone
 

aiida/orm/nodes/data/folder.py

@@ -37,4 +37,4 @@ def __init__(self, **kwargs):
         tree = kwargs.pop('tree', None)
         super().__init__(**kwargs)
         if tree:
-            self.put_object_from_tree(tree)
+            self.base.repository.put_object_from_tree(tree)

aiida/orm/nodes/data/singlefile.py

@@ -56,7 +56,7 @@ def open(self, path=None, mode='r'):
         if path is None:
             path = self.filename
 
-        with super().open(path, mode=mode) as handle:
+        with self.base.repository.open(path, mode=mode) as handle:
             yield handle
 
     def get_content(self):
@@ -94,7 +94,7 @@ def set_file(self, file, filename=None):
 
         key = filename or key
 
-        existing_object_names = self.list_object_names()
+        existing_object_names = self.base.repository.list_object_names()
 
         try:
             # Remove the 'key' from the list of currently existing objects such that it is not deleted after storing
@@ -103,13 +103,13 @@ def set_file(self, file, filename=None):
             pass
 
         if is_filelike:
-            self.put_object_from_filelike(file, key)
+            self.base.repository.put_object_from_filelike(file, key)
         else:
-            self.put_object_from_file(file, key)
+            self.base.repository.put_object_from_file(file, key)
 
         # Delete any other existing objects (minus the current `key` which was already removed from the list)
         for existing_key in existing_object_names:
-            self.delete_object(existing_key)
+            self.base.repository.delete_object(existing_key)
 
         self.set_attribute('filename', key)
 
@@ -122,7 +122,7 @@ def _validate(self):
         except AttributeError:
             raise exceptions.ValidationError('the `filename` attribute is not set.')
 
-        objects = self.list_object_names()
+        objects = self.base.repository.list_object_names()
 
         if [filename] != objects:
             raise exceptions.ValidationError(