Add 'verdi node repo dump' command.

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

@@ -10,7 +10,9 @@
 """Tests for verdi node"""
 
 import os
+import io
 import errno
+import pathlib
 import tempfile
 
 from click.testing import CliRunner
@@ -52,6 +54,17 @@ def setUpClass(cls, *args, **kwargs):
 
         cls.node = node
 
+        # Set up a FolderData for the node repo cp tests.
+        folder_node = orm.FolderData()
+        cls.content_file1 = 'nobody expects'
+        cls.content_file2 = 'the minister of silly walks'
+        cls.key_file1 = 'some/nested/folder/filename.txt'
+        cls.key_file2 = 'some_other_file.txt'
+        folder_node.put_object_from_filelike(io.StringIO(cls.content_file1), key=cls.key_file1)
+        folder_node.put_object_from_filelike(io.StringIO(cls.content_file2), key=cls.key_file2)
+        folder_node.store()
+        cls.folder_node = folder_node
+
     def setUp(self):
         self.cli_runner = CliRunner()
 
@@ -171,6 +184,60 @@ def test_node_extras(self):
                 result = self.cli_runner.invoke(cmd_node.extras, options)
                 self.assertIsNone(result.exception, result.output)
 
+    def test_node_repo_dump(self):
+        """Test 'verdi node repo dump' command."""
+
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            out_path = pathlib.Path(tmp_dir) / 'out_dir'
+            options = [str(self.folder_node.uuid), str(out_path)]
+            res = self.cli_runner.invoke(cmd_node.repo_dump, options, catch_exceptions=False)
+            self.assertFalse(res.stdout)
+
+            for file_key, content in [(self.key_file1, self.content_file1), (self.key_file2, self.content_file2)]:
+                curr_path = out_path
+                for key_part in file_key.split('/'):
+                    curr_path /= key_part
+                    self.assertTrue(curr_path.exists())
+                with curr_path.open('r') as res_file:
+                    self.assertEqual(res_file.read(), content)
+
+    def test_node_repo_dump_to_nested_folder(self):
+        """Test 'verdi node repo dump' command, with an output folder whose parent does not exist."""
+
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            out_path = pathlib.Path(tmp_dir) / 'out_dir' / 'nested' / 'path'
+            options = [str(self.folder_node.uuid), str(out_path)]
+            res = self.cli_runner.invoke(cmd_node.repo_dump, options, catch_exceptions=False)
+            self.assertFalse(res.stdout)
+
+            for file_key, content in [(self.key_file1, self.content_file1), (self.key_file2, self.content_file2)]:
+                curr_path = out_path
+                for key_part in file_key.split('/'):
+                    curr_path /= key_part
+                    self.assertTrue(curr_path.exists())
+                with curr_path.open('r') as res_file:
+                    self.assertEqual(res_file.read(), content)
+
+    def test_node_repo_existing_out_dir(self):
+        """Test 'verdi node repo dump' command, check that an existing output directory is not overwritten."""
+
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            out_path = pathlib.Path(tmp_dir) / 'out_dir'
+            # Create the directory and put a file in it
+            out_path.mkdir()
+            some_file = out_path / 'file_name'
+            some_file_content = 'ni!'
+            with some_file.open('w') as file_handle:
+                file_handle.write(some_file_content)
+            options = [str(self.folder_node.uuid), str(out_path)]
+            res = self.cli_runner.invoke(cmd_node.repo_dump, options, catch_exceptions=False)
+            self.assertIn('exists', res.stdout)
+            self.assertIn('Aborted!', res.stdout)
+
+            # Make sure the directory content is still there
+            with some_file.open('r') as file_handle:
+                assert file_handle.read() == some_file_content
+
 
 def delete_temporary_file(filepath):
     """

aiida/cmdline/commands/cmd_node.py

@@ -9,6 +9,9 @@
 ###########################################################################
 """`verdi node` command."""
 
+import shutil
+import pathlib
+
 import click
 import tabulate
 
@@ -61,6 +64,52 @@ def repo_ls(node, relative_path, color):
         echo.echo_critical(exception)
 
 
+@verdi_node_repo.command('dump')
+@arguments.NODE()
+@click.argument(
+    'output_directory',
+    type=click.Path(),
+    required=True,
+)
+@with_dbenv()
+def repo_dump(node, output_directory):
+    """Copy the repository files of a node to an output directory."""
+    from aiida.orm.utils.repository import FileType
+
+    output_directory = pathlib.Path(output_directory)
+
+    try:
+        output_directory.mkdir(parents=True, exist_ok=False)
+    except FileExistsError:
+        click.echo('Error: Invalid value for "OUTPUT_DIRECTORY": Path "{}" exists.'.format(output_directory))
+        raise click.Abort
+
+    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=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 key + '/' + file.name
+            if file.type == FileType.DIRECTORY:
+                new_out_dir = output_dir / file.name
+                assert not new_out_dir.exists()
+                new_out_dir.mkdir()
+                _copy_tree(key=file_key, output_dir=new_out_dir)
+
+            else:
+                assert 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 out_file_path.open('wb') as out_file:
+                        shutil.copyfileobj(in_file, out_file)
+
+    _copy_tree(key='', output_dir=output_directory)
+
+
 @verdi_node.command('label')
 @arguments.NODES()
 @options.LABEL(help='Set LABEL as the new label for all NODES')