File sources for gdrive, gcs, onedata, basespace

lib/galaxy/dependencies/__init__.py

@@ -232,6 +232,18 @@ def check_s3fs(self):
         # use s3fs directly (skipping pyfilesystem) for direct access to more options
         return 's3fs' in self.file_sources
 
+    def check_fs_googledrivefs(self):
+        return 'googledrive' in self.file_sources
+
+    def check_fs_gcsfs(self):
+        return 'googlecloudstorage' in self.file_sources
+
+    def check_fs_onedatafs(self):
+        return 'onedata' in self.file_sources
+
+    def check_fs_basespace(self):
+        return 'basespace' in self.file_sources
+
     def check_watchdog(self):
         install_set = {'auto', 'True', 'true', 'polling', True}
         return (self.config['watch_tools'] in install_set

lib/galaxy/dependencies/conditional-requirements.txt

@@ -20,6 +20,10 @@ fs.sshfs  # type: ssh
 fs-s3fs  # type: s3
 s3fs  # type: s3fs
 fs.anvilfs # type: anvil
+fs.googledrivefs # type: googledrive
+fs-gcsfs # type: googlecloudstorage
+fs-onedatafs # type: onedata
+fs-basespace # type: basespace
 
 # Chronos client
 chronos-python==1.2.1

lib/galaxy/files/sources/basespace.py

@@ -0,0 +1,20 @@
+try:
+    from fs_basespace import BASESPACEFS
+except ImportError:
+    BASESPACEFS = None
+
+from ._pyfilesystem2 import PyFilesystem2FilesSource
+
+
+class BaseSpaceFilesSource(PyFilesystem2FilesSource):
+    plugin_type = 'basespace'
+    required_module = BASESPACEFS
+    required_package = "fs-basespace"
+
+    def _open_fs(self, user_context):
+        props = self._serialization_props(user_context)
+        handle = BASESPACEFS(**props)
+        return handle
+
+
+__all__ = ('BaseSpaceFilesSource',)

lib/galaxy/files/sources/googlecloudstorage.py

@@ -0,0 +1,30 @@
+try:
+    from fs_gcsfs import GCSFS
+    from google.cloud.storage import Client
+    from google.oauth2.credentials import Credentials
+except ImportError:
+    GCSFS = None
+
+from ._pyfilesystem2 import PyFilesystem2FilesSource
+
+
+class GoogleCloudStorageFilesSource(PyFilesystem2FilesSource):
+    plugin_type = 'googlecloudstorage'
+    required_module = GCSFS
+    required_package = "fs-gcsfs"
+
+    def _open_fs(self, user_context):
+        props = self._serialization_props(user_context)
+        bucket_name = props.pop('bucket_name', None)
+        root_path = props.pop('root_path', None)
+        project = props.pop('project', None)
+        args = {}
+        if props.get('anonymous'):
+            args['client'] = Client.create_anonymous_client()
+        elif props.get('token'):
+            args['client'] = Client(project=project, credentials=Credentials(**props))
+        handle = GCSFS(bucket_name, root_path=root_path, retry=0, **args)
+        return handle
+
+
+__all__ = ('GoogleCloudStorageFilesSource',)

lib/galaxy/files/sources/googledrive.py

@@ -0,0 +1,22 @@
+try:
+    from fs.googledrivefs import GoogleDriveFS
+    from google.oauth2.credentials import Credentials
+except ImportError:
+    GoogleDriveFS = None
+
+from ._pyfilesystem2 import PyFilesystem2FilesSource
+
+
+class GoogleDriveFilesSource(PyFilesystem2FilesSource):
+    plugin_type = 'googledrive'
+    required_module = GoogleDriveFS
+    required_package = "fs.googledrivefs"
+
+    def _open_fs(self, user_context):
+        props = self._serialization_props(user_context)
+        credentials = Credentials(**props)
+        handle = GoogleDriveFS(credentials)
+        return handle
+
+
+__all__ = ('GoogleDriveFilesSource',)

lib/galaxy/files/sources/onedata.py

@@ -0,0 +1,20 @@
+try:
+    from fs.onedatafs import OnedataFS
+except ImportError:
+    OnedataFS = None
+
+from ._pyfilesystem2 import PyFilesystem2FilesSource
+
+
+class OneDataFilesSource(PyFilesystem2FilesSource):
+    plugin_type = 'onedata'
+    required_module = OnedataFS
+    required_package = "fs-onedatafs"
+
+    def _open_fs(self, user_context):
+        props = self._serialization_props(user_context)
+        handle = OnedataFS(**props)
+        return handle
+
+
+__all__ = ('OneDataFilesSource',)

packages/files/test-requirements.txt

@@ -1 +1,2 @@
 pytest
+fs-gcsfs

test/unit/files/_util.py

@@ -4,6 +4,7 @@
 
 from galaxy.files import (
     ConfiguredFileSources,
+    ConfiguredFileSourcesConfig,
     DictFileSourcesUserContext,
 )
 
@@ -55,6 +56,21 @@ def user_context_fixture(user_ftp_dir=None, role_names=None, group_names=None, i
         preferences={
             'webdav|password': 'secret1234',
             'dropbox|access_token': os.environ.get('GALAXY_TEST_DROPBOX_ACCESS_TOKEN'),
+            'googledrive|client_id': os.environ.get('GALAXY_TEST_GOOGLE_DRIVE_CLIENT_ID'),
+            'googledrive|client_secret': os.environ.get('GALAXY_TEST_GOOGLE_DRIVE_CLIENT_SECRET'),
+            'googledrive|access_token': os.environ.get('GALAXY_TEST_GOOGLE_DRIVE_ACCESS_TOKEN'),
+            'googledrive|refresh_token': os.environ.get('GALAXY_TEST_GOOGLE_DRIVE_REFRESH_TOKEN'),
+            'googlecloudstorage|project': os.environ.get('GALAXY_TEST_GCS_PROJECT'),
+            'googlecloudstorage|bucket_name': os.environ.get('GALAXY_TEST_GCS_BUCKET'),
+            'googlecloudstorage|client_id': os.environ.get('GALAXY_TEST_GCS_CLIENT_ID'),
+            'googlecloudstorage|client_secret': os.environ.get('GALAXY_TEST_GCS_CLIENT_SECRET'),
+            'googlecloudstorage|access_token': os.environ.get('GALAXY_TEST_GCS_ACCESS_TOKEN'),
+            'googlecloudstorage|refresh_token': os.environ.get('GALAXY_TEST_GCS_REFRESH_TOKEN'),
+            'onedata|provider_host': os.environ.get('GALAXY_TEST_ONEDATA_PROVIDER_HOST'),
+            'onedata|access_token': os.environ.get('GALAXY_TEST_ONEDATA_ACCESS_TOKEN'),
+            'basespace|client_id': os.environ.get('GALAXY_TEST_ONEDATA_CLIENT_ID'),
+            'basespace|client_secret': os.environ.get('GALAXY_TEST_ONEDATA_CLIENT_SECRET'),
+            'basespace|access_token': os.environ.get('GALAXY_TEST_ONEDATA_ACCESS_TOKEN'),
         },
         role_names=role_names or set(),
         group_names=group_names or set(),
@@ -63,43 +79,43 @@ def user_context_fixture(user_ftp_dir=None, role_names=None, group_names=None, i
     return user_context
 
 
-def assert_realizes_as(file_sources, uri, expected, user_context=None):
+def realize_to_temp_file(file_sources, uri, user_context=None):
     file_source_path = file_sources.get_file_source_path(uri)
     with tempfile.NamedTemporaryFile(mode='r') as temp:
         file_source_path.file_source.realize_to(file_source_path.path, temp.name, user_context=user_context)
         with open(temp.name) as f:
             realized_contents = f.read()
-        if realized_contents != expected:
-            message = "Expected to realize contents at [{}] as [{}], instead found [{}]".format(
-                uri,
-                expected,
-                realized_contents,
-            )
-            raise AssertionError(message)
+            return realized_contents
+
+
+def assert_realizes_as(file_sources, uri, expected, user_context=None):
+    realized_contents = realize_to_temp_file(file_sources, uri, user_context=user_context)
+    if realized_contents != expected:
+        message = "Expected to realize contents at [{}] as [{}], instead found [{}]".format(
+            uri,
+            expected,
+            realized_contents,
+        )
+        raise AssertionError(message)
 
 
 def assert_realizes_contains(file_sources, uri, expected, user_context=None):
-    file_source_path = file_sources.get_file_source_path(uri)
-    with tempfile.NamedTemporaryFile(mode='r') as temp:
-        file_source_path.file_source.realize_to(file_source_path.path, temp.name, user_context=user_context)
-        realized_contents = temp.read()
-        if expected not in realized_contents:
-            message = "Expected to realize contents at [{}] to contain [{}], instead found [{}]".format(
-                uri,
-                expected,
-                realized_contents,
-            )
-            raise AssertionError(message)
+    realized_contents = realize_to_temp_file(file_sources, uri, user_context=user_context)
+    if expected not in realized_contents:
+        message = "Expected to realize contents at [{}] to contain [{}], instead found [{}]".format(
+            uri,
+            expected,
+            realized_contents,
+        )
+        raise AssertionError(message)
 
 
 def assert_realizes_throws_exception(file_sources, uri, user_context=None) -> Exception:
-    file_source_path = file_sources.get_file_source_path(uri)
     exception = None
-    with tempfile.NamedTemporaryFile(mode='r') as temp:
-        try:
-            file_source_path.file_source.realize_to(file_source_path.path, temp.name, user_context=user_context)
-        except Exception as e:
-            exception = e
+    try:
+        realize_to_temp_file(file_sources, uri, user_context=user_context)
+    except Exception as e:
+        exception = e
     assert exception
     return exception
 
@@ -110,3 +126,25 @@ def write_from(file_sources, uri, content, user_context=None):
         f.write(content)
         f.flush()
         file_source_path.file_source.write_from(file_source_path.path, f.name, user_context=user_context)
+
+
+def configured_file_sources(conf_file):
+    file_sources_config = ConfiguredFileSourcesConfig()
+    return ConfiguredFileSources(file_sources_config, conf_file=conf_file)
+
+
+def assert_simple_file_realize(conf_file, recursive=False, filename="a", contents="a\n", contains=False):
+    user_context = user_context_fixture()
+    file_sources = configured_file_sources(conf_file)
+    file_source_pair = file_sources.get_file_source_path("gxfiles://test1")
+
+    assert file_source_pair.path == "/"
+    file_source = file_source_pair.file_source
+    res = file_source.list("/", recursive=recursive, user_context=user_context)
+    a_file = find(res, class_="File", name=filename)
+    assert a_file
+
+    if contains:
+        assert_realizes_contains(file_sources, f"gxfiles://test1/{filename}", contents, user_context=user_context)
+    else:
+        assert_realizes_as(file_sources, f"gxfiles://test1/{filename}", contents, user_context=user_context)

test/unit/files/basespace_file_sources_conf.yml

@@ -0,0 +1,7 @@
+- type: basespace
+  id: test1
+  doc: Test access to Illumina BaseSpace
+  basespace_server: https://api.basespace.illumina.com
+  client_id: ${user.preferences['basespace|client_id']}
+  client_secret: ${user.preferences['basespace|client_secret']}
+  access_token: ${user.preferences['basespace|access_token']}

test/unit/files/dropbox_file_sources_conf.yml

@@ -1,4 +1,4 @@
 - type: dropbox
   id: test1
-  doc: Test WebDAV server.
+  doc: Test access to a dropbox account.
   accessToken: ${user.preferences['dropbox|access_token']}

test/unit/files/gcsfs_file_sources_conf.yml

@@ -0,0 +1,11 @@
+- type: googlecloudstorage
+  id: test1
+  doc: Test access to Google Cloud Storage.
+  project: ${user.preferences['googlecloudstorage|project']}
+  bucket_name: 'genomics-public-data'
+  token_uri: "https://www.googleapis.com/oauth2/v4/token"
+  client_id: ${user.preferences['googlecloudstorage|client_id']}
+  client_secret: ${user.preferences['googlecloudstorage|client_secret']}
+  token: ${user.preferences['googlecloudstorage|access_token']}
+  refresh_token: ${user.preferences['googlecloudstorage|refresh_token']}
+  anonymous: true

test/unit/files/googledrive_file_sources_conf.yml

@@ -0,0 +1,8 @@
+- type: googledrive
+  id: test1
+  doc: Test access to a Google drive.
+  token: ${user.preferences['googledrive|access_token']}
+  refresh_token: ${user.preferences['googledrive|refresh_token']}
+  token_uri: "https://www.googleapis.com/oauth2/v4/token"
+  client_id: ${user.preferences['googledrive|client_id']}
+  client_secret: ${user.preferences['googledrive|client_secret']}

test/unit/files/onedata_file_sources_conf.yml

@@ -0,0 +1,5 @@
+- type: onedata
+  id: test1
+  doc: Test access to a OneData host
+  provider_host: ${user.preferences['onedata|provider_host']}
+  access_token: ${user.preferences['onedata|access_token']}

test/unit/files/test_basespace.py

@@ -0,0 +1,36 @@
+import os
+
+import pytest
+
+from ._util import (
+    assert_realizes_as,
+    configured_file_sources,
+    find,
+    user_context_fixture,
+)
+SCRIPT_DIRECTORY = os.path.abspath(os.path.dirname(__file__))
+FILE_SOURCES_CONF = os.path.join(SCRIPT_DIRECTORY, "basespace_file_sources_conf.yml")
+
+skip_if_no_basespace_access_token = pytest.mark.skipif(
+    not os.environ.get('GALAXY_TEST_BASESPACE_CLIENT_ID')
+    or not os.environ.get('GALAXY_TEST_BASESPACE_CLIENT_SECRET')
+    or not os.environ.get('GALAXY_TEST_BASESPACE_ACCESS_TOKEN')
+    or not os.environ.get('GALAXY_TEST_BASESPACE_TEST_FILE_PATH'),
+    reason="GALAXY_TEST_BASESPACE_CLIENT_ID and related vars not set"
+)
+
+
+@skip_if_no_basespace_access_token
+def test_file_source():
+    user_context = user_context_fixture()
+    file_sources = configured_file_sources(FILE_SOURCES_CONF)
+    file_source_pair = file_sources.get_file_source_path("gxfiles://test1")
+
+    assert file_source_pair.path == "/"
+    file_source = file_source_pair.file_source
+    test_file = os.environ.get('GALAXY_TEST_BASESPACE_TEST_FILE_PATH', "")
+    res = file_source.list(os.path.dirname(test_file), recursive=False, user_context=user_context)
+    a_file = find(res, class_="File", name=os.path.basename(test_file))
+    assert a_file
+
+    assert_realizes_as(file_sources, a_file['uri'], "a\n", user_context=user_context)

test/unit/files/test_dropbox.py

@@ -2,12 +2,8 @@
 
 import pytest
 
-from galaxy.files import ConfiguredFileSources, ConfiguredFileSourcesConfig
-from ._util import (
-    assert_realizes_as,
-    find_file_a,
-    user_context_fixture,
-)
+from ._util import assert_simple_file_realize
+
 SCRIPT_DIRECTORY = os.path.abspath(os.path.dirname(__file__))
 FILE_SOURCES_CONF = os.path.join(SCRIPT_DIRECTORY, "dropbox_file_sources_conf.yml")
 
@@ -19,19 +15,4 @@
 
 @skip_if_no_dropbox_access_token
 def test_file_source():
-    user_context = user_context_fixture()
-    file_sources = _configured_file_sources()
-    file_source_pair = file_sources.get_file_source_path("gxfiles://test1")
-
-    assert file_source_pair.path == "/"
-    file_source = file_source_pair.file_source
-    res = file_source.list("/", recursive=True, user_context=user_context)
-    a_file = find_file_a(res)
-    assert a_file
-
-    assert_realizes_as(file_sources, "gxfiles://test1/a", "a\n", user_context=user_context)
-
-
-def _configured_file_sources(conf_file=FILE_SOURCES_CONF):
-    file_sources_config = ConfiguredFileSourcesConfig()
-    return ConfiguredFileSources(file_sources_config, conf_file=conf_file)
+    assert_simple_file_realize(FILE_SOURCES_CONF, recursive=True)

test/unit/files/test_gcsfs.py

@@ -0,0 +1,25 @@
+import os
+
+import pytest
+
+from ._util import assert_simple_file_realize
+
+try:
+    from fs_gcsfs import GCSFS
+except ImportError:
+    GCSFS = None
+
+SCRIPT_DIRECTORY = os.path.abspath(os.path.dirname(__file__))
+FILE_SOURCES_CONF = os.path.join(SCRIPT_DIRECTORY, "gcsfs_file_sources_conf.yml")
+
+
+skip_if_no_gcsfs_libs = pytest.mark.skipif(
+    not GCSFS,
+    reason="Required lib to run gcs file source test: fs_gcsfs is not available"
+)
+
+
+@skip_if_no_gcsfs_libs
+def test_file_source():
+    assert_simple_file_realize(FILE_SOURCES_CONF, recursive=False, filename="README", contents="1000genomes",
+                               contains=True)