Warning

This document is for an old release of Galaxy. You can alternatively view this page in the latest release if it exists or view the top of the latest release's documentation.

galaxy.objectstore package

objectstore package, abstraction for storing blobs of data for use in Galaxy.

all providers ensure that data can be accessed on the filesystem for running tools

class galaxy.objectstore.ObjectStore

Bases: object

ObjectStore interface.

FIELD DESCRIPTIONS (these apply to all the methods in this class):

Parameters

• obj (StorableObject) – A Galaxy object with an assigned database ID accessible via the .id attribute.

• base_dir (string) – A key in self.extra_dirs corresponding to the base directory in which this object should be created, or None to specify the default directory.

• dir_only (boolean) – If True, check only the path where the file identified by obj should be located, not the dataset itself. This option applies to extra_dir argument as well.

• extra_dir (string) – Append extra_dir to the directory structure where the dataset identified by obj should be located. (e.g., 000/extra_dir/obj.id). Valid values include ‘job_work’ (defaulting to config.jobs_directory = ‘$GALAXY_ROOT/database/jobs_directory’); ‘temp’ (defaulting to config.new_file_path = ‘$GALAXY_ROOT/database/tmp’).

• extra_dir_at_root (boolean) – Applicable only if extra_dir is set. If True, the extra_dir argument is placed at root of the created directory structure rather than at the end (e.g., extra_dir/000/obj.id vs. 000/extra_dir/obj.id)

• alt_name (string) – Use this name as the alternative name for the created dataset rather than the default.

• obj_dir (boolean) – Append a subdirectory named with the object’s ID (e.g. 000/obj.id)

abstract exists(obj, base_dir=None, dir_only=False, extra_dir=None, extra_dir_at_root=False, alt_name=None)

Return True if the object identified by obj exists, False otherwise.

abstract construct_path(obj, base_dir=None, dir_only=False, extra_dir=None, extra_dir_at_root=False, alt_name=None)

abstract create(obj, base_dir=None, dir_only=False, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Mark the object (obj) as existing in the store, but with no content.

This method will create a proper directory structure for the file if the directory does not already exist.

abstract empty(obj, base_dir=None, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Test if the object identified by obj has content.

If the object does not exist raises ObjectNotFound.

abstract size(obj, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Return size of the object identified by obj.

If the object does not exist, return 0.

abstract delete(obj, entire_dir=False, base_dir=None, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Delete the object identified by obj.

Parameters

entire_dir (boolean) – If True, delete the entire directory pointed to by extra_dir. For safety reasons, this option applies only for and in conjunction with the extra_dir or obj_dir options.

abstract get_data(obj, start=0, count=-1, base_dir=None, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Fetch count bytes of data offset by start bytes using obj.id.

If the object does not exist raises ObjectNotFound.

Parameters

• start (int) – Set the position to start reading the dataset file

• count (int) – Read at most count bytes from the dataset

abstract get_filename(obj, base_dir=None, dir_only=False, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Get the expected filename with absolute path for object with id obj.id.

This can be used to access the contents of the object.

abstract update_from_file(obj, base_dir=None, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False, file_name=None, create=False)

Inform the store that the file associated with obj.id has been updated.

If file_name is provided, update from that file instead of the default. If the object does not exist raises ObjectNotFound.

Parameters

• file_name (string) – Use file pointed to by file_name as the source for updating the dataset identified by obj

• create (boolean) – If True and the default dataset does not exist, create it first.

abstract get_object_url(obj, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Return the URL for direct access if supported, otherwise return None.

Note: need to be careful to not bypass dataset security with this.

abstract get_concrete_store_name(obj)

Return a display name or title of the objectstore corresponding to obj.

To accommodate nested objectstores, obj is passed in so this metadata can be returned for the ConcreteObjectStore corresponding to the object.

If the dataset is in a new or discarded state and an object_store_id has not yet been set, this may return None.

abstract get_concrete_store_description_markdown(obj)

Return a longer description of how data ‘obj’ is stored.

To accommodate nested objectstores, obj is passed in so this metadata can be returned for the ConcreteObjectStore corresponding to the object.

If the dataset is in a new or discarded state and an object_store_id has not yet been set, this may return None.

abstract get_store_usage_percent()

Return the percentage indicating how full the store is.

abstract get_store_by(obj)

Return how object is stored (by ‘uuid’, ‘id’, or None if not yet saved).

Certain Galaxy remote data features aren’t available if objects are stored by ‘id’.

abstract to_dict() → Dict[str, Any]

class galaxy.objectstore.BaseObjectStore(config, config_dict=None, **kwargs)

Bases: ObjectStore

store_by: str

store_type: str

__init__(config, config_dict=None, **kwargs)

Parameters

config (object) –

An object, most likely populated from galaxy/config.ini, having the following attributes:

• object_store_check_old_style (only used by the DiskObjectStore subclass)

• jobs_directory – Each job is given a unique empty directory as its current working directory. This option defines in what parent directory those directories will be created.

• new_file_path – Used to set the ‘temp’ extra_dir.

start()

Call all postfork function(s) here. These functions spawn a thread. We register start(self) with app, so app starts the threads. Override this function in subclasses, as needed.

shutdown()

Close any connections for this ObjectStore.

file_ready(obj, base_dir=None, dir_only=False, extra_dir=None, extra_dir_at_root=False, alt_name=None, obj_dir=False)

Check if a file corresponding to a dataset is ready to be used.

Return True if so, False otherwise

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

classmethod from_xml(config, config_xml, **kwd)

to_dict()

exists(obj, **kwargs)

Return True if the object identified by obj exists, False otherwise.

construct_path(obj, **kwargs)

create(obj, **kwargs)

Mark the object (obj) as existing in the store, but with no content.

This method will create a proper directory structure for the file if the directory does not already exist.

empty(obj, **kwargs)

Test if the object identified by obj has content.

If the object does not exist raises ObjectNotFound.

size(obj, **kwargs)

Return size of the object identified by obj.

If the object does not exist, return 0.

delete(obj, **kwargs)

Delete the object identified by obj.

Parameters

entire_dir (boolean) – If True, delete the entire directory pointed to by extra_dir. For safety reasons, this option applies only for and in conjunction with the extra_dir or obj_dir options.

get_data(obj, **kwargs)

Fetch count bytes of data offset by start bytes using obj.id.

If the object does not exist raises ObjectNotFound.

Parameters

• start (int) – Set the position to start reading the dataset file

• count (int) – Read at most count bytes from the dataset

get_filename(obj, **kwargs)

Get the expected filename with absolute path for object with id obj.id.

This can be used to access the contents of the object.

update_from_file(obj, **kwargs)

Inform the store that the file associated with obj.id has been updated.

If file_name is provided, update from that file instead of the default. If the object does not exist raises ObjectNotFound.

Parameters

• file_name (string) – Use file pointed to by file_name as the source for updating the dataset identified by obj

• create (boolean) – If True and the default dataset does not exist, create it first.

get_object_url(obj, **kwargs)

Return the URL for direct access if supported, otherwise return None.

Note: need to be careful to not bypass dataset security with this.

get_concrete_store_name(obj)

Return a display name or title of the objectstore corresponding to obj.

To accommodate nested objectstores, obj is passed in so this metadata can be returned for the ConcreteObjectStore corresponding to the object.

If the dataset is in a new or discarded state and an object_store_id has not yet been set, this may return None.

get_concrete_store_description_markdown(obj)

Return a longer description of how data ‘obj’ is stored.

To accommodate nested objectstores, obj is passed in so this metadata can be returned for the ConcreteObjectStore corresponding to the object.

If the dataset is in a new or discarded state and an object_store_id has not yet been set, this may return None.

get_store_usage_percent()

Return the percentage indicating how full the store is.

get_store_by(obj, **kwargs)

Return how object is stored (by ‘uuid’, ‘id’, or None if not yet saved).

Certain Galaxy remote data features aren’t available if objects are stored by ‘id’.

class galaxy.objectstore.ConcreteObjectStore(config, config_dict=None, **kwargs)

Bases: BaseObjectStore

Subclass of ObjectStore for stores that don’t delegate (non-nested).

Currently only adds store_by functionality. Which doesn’t make sense for the delegating object stores.

__init__(config, config_dict=None, **kwargs)

Parameters

config (object) –

An object, most likely populated from galaxy/config.ini, having the following attributes:

• object_store_check_old_style (only used by the DiskObjectStore subclass)

• jobs_directory – Each job is given a unique empty directory as its current working directory. This option defines in what parent directory those directories will be created.

• new_file_path – Used to set the ‘temp’ extra_dir.

store_by: str

to_dict()

store_type: str

class galaxy.objectstore.DiskObjectStore(config, config_dict)

Bases: ConcreteObjectStore

Standard Galaxy object store.

Stores objects in files under a specific directory on disk.

>>> from galaxy.util.bunch import Bunch
>>> import tempfile
>>> file_path=tempfile.mkdtemp()
>>> obj = Bunch(id=1)
>>> s = DiskObjectStore(Bunch(umask=0o077, jobs_directory=file_path, new_file_path=file_path, object_store_check_old_style=False), dict(files_dir=file_path))
>>> s.create(obj)
>>> s.exists(obj)
True
>>> assert s.get_filename(obj) == file_path + '/000/dataset_1.dat'

store_type: str = 'disk'

__init__(config, config_dict)

Parameters

• config (object) –

  An object, most likely populated from galaxy/config.ini, having the same attributes needed by ObjectStore plus:

  • file_path – Default directory to store objects to disk in.

  • umask – the permission bits for newly created files.

• file_path (str) – Override for the config.file_path value.

• extra_dirs (dict) – Keys are string, values are directory paths.

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

to_dict()

store_by: str

class galaxy.objectstore.NestedObjectStore(config, config_xml=None)

Bases: BaseObjectStore

Base for ObjectStores that use other ObjectStores.

Example: DistributedObjectStore, HierarchicalObjectStore

__init__(config, config_xml=None)

Extend ObjectStore’s constructor.

shutdown()

For each backend, shuts them down.

file_ready(obj, **kwargs)

Determine if the file for obj is ready to be used by any of the backends.

store_by: str

store_type: str

class galaxy.objectstore.DistributedObjectStore(config, config_dict, fsmon=False)

Bases: NestedObjectStore

ObjectStore that defers to a list of backends.

When getting objects the first store where the object exists is used. When creating objects they are created in a store selected randomly, but with weighting.

store_type: str = 'distributed'

__init__(config, config_dict, fsmon=False)

Parameters

• config (object) –

  An object, most likely populated from galaxy/config.ini, having the same attributes needed by NestedObjectStore plus:

  • distributed_object_store_config_file

• fsmon (bool) – If True, monitor the file system for free space, removing backends when they get too full.

classmethod parse_xml(config_xml, legacy=False)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

classmethod from_xml(config, config_xml, fsmon=False)

to_dict() → Dict[str, Any]

shutdown()

Shut down. Kill the free space monitor if there is one.

store_by: str

class galaxy.objectstore.HierarchicalObjectStore(config, config_dict, fsmon=False)

Bases: NestedObjectStore

ObjectStore that defers to a list of backends.

When getting objects the first store where the object exists is used. When creating objects only the first store is used.

store_type: str = 'hierarchical'

__init__(config, config_dict, fsmon=False)

The default constructor. Extends NestedObjectStore.

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

to_dict()

store_by: str

galaxy.objectstore.type_to_object_store_class(store: str, fsmon: bool = False) → Tuple[Type[BaseObjectStore], Dict[str, Any]]

galaxy.objectstore.build_object_store_from_config(config, fsmon=False, config_xml=None, config_dict=None)

Invoke the appropriate object store.

Will use the object_store_config_file attribute of the config object to configure a new object store from the specified XML file.

Or you can specify the object store type in the object_store attribute of the config object. Currently ‘disk’, ‘s3’, ‘swift’, ‘distributed’, ‘hierarchical’, ‘irods’, and ‘pulsar’ are supported values.

galaxy.objectstore.local_extra_dirs(func)

Non-local plugin decorator using local directories for the extra_dirs (job_work and temp).

galaxy.objectstore.convert_bytes(bytes)

A helper function used for pretty printing disk usage.

galaxy.objectstore.config_to_dict(config)

Dict-ify the portion of a config object consumed by the ObjectStore class and its subclasses.

class galaxy.objectstore.ObjectStorePopulator(has_object_store, user)

Bases: object

Small helper for interacting with the object store and making sure all datasets from a job end up with the same object_store_id.

__init__(has_object_store, user)

set_object_store_id(data)

set_dataset_object_store_id(dataset)

Subpackages

• galaxy.objectstore.unittest_utils package
  • Config
    • Config.__init__()
    • Config.write()
  • MockConfig
    • MockConfig.__init__()

Submodules

galaxy.objectstore.azure_blob module

Object Store plugin for the Microsoft Azure Block Blob Storage system

galaxy.objectstore.azure_blob.parse_config_xml(config_xml)

class galaxy.objectstore.azure_blob.AzureBlobObjectStore(config, config_dict)

Bases: ConcreteObjectStore

Object store that stores objects as blobs in an Azure Blob Container. A local cache exists that is used as an intermediate location for files between Galaxy and Azure.

store_type: str = 'azure_blob'

__init__(config, config_dict)

Parameters

config (object) –

An object, most likely populated from galaxy/config.ini, having the following attributes:

• object_store_check_old_style (only used by the DiskObjectStore subclass)

• jobs_directory – Each job is given a unique empty directory as its current working directory. This option defines in what parent directory those directories will be created.

• new_file_path – Used to set the ‘temp’ extra_dir.

to_dict()

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

file_ready(obj, **kwargs)

A helper method that checks if a file corresponding to a dataset is ready and available to be used. Return True if so, False otherwise.

store_by: str

galaxy.objectstore.cloud module

Object Store plugin for Cloud storage.

class galaxy.objectstore.cloud.CloudConfigMixin

Bases: object

class galaxy.objectstore.cloud.Cloud(config, config_dict)

Bases: ConcreteObjectStore, CloudConfigMixin

Object store that stores objects as items in an cloud storage. A local cache exists that is used as an intermediate location for files between Galaxy and the cloud storage.

store_type: str = 'cloud'

__init__(config, config_dict)

Parameters

config (object) –

An object, most likely populated from galaxy/config.ini, having the following attributes:

• object_store_check_old_style (only used by the DiskObjectStore subclass)

• jobs_directory – Each job is given a unique empty directory as its current working directory. This option defines in what parent directory those directories will be created.

• new_file_path – Used to set the ‘temp’ extra_dir.

start_cache_monitor()

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

to_dict()

file_ready(obj, **kwargs)

A helper method that checks if a file corresponding to a dataset is ready and available to be used. Return True if so, False otherwise.

shutdown()

Close any connections for this ObjectStore.

store_by: str

galaxy.objectstore.irods module

Object Store plugin for the Integrated Rule-Oriented Data System (iRODS)

galaxy.objectstore.irods.parse_config_xml(config_xml)

class galaxy.objectstore.irods.CloudConfigMixin

Bases: object

class galaxy.objectstore.irods.IRODSObjectStore(config, config_dict)

Bases: DiskObjectStore, CloudConfigMixin

Object store that stores files as data objects in an iRODS Zone. A local cache exists that is used as an intermediate location for files between Galaxy and iRODS.

store_type: str = 'irods'

__init__(config, config_dict)

Parameters

• config (object) –

  An object, most likely populated from galaxy/config.ini, having the same attributes needed by ObjectStore plus:

  • file_path – Default directory to store objects to disk in.

  • umask – the permission bits for newly created files.

• file_path (str) – Override for the config.file_path value.

• extra_dirs (dict) – Keys are string, values are directory paths.

shutdown()

Close any connections for this ObjectStore.

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

start_connection_pool_monitor()

start()

Call all postfork function(s) here. These functions spawn a thread. We register start(self) with app, so app starts the threads. Override this function in subclasses, as needed.

to_dict()

file_ready(obj, **kwargs)

A helper method that checks if a file corresponding to a dataset is ready and available to be used. Return True if so, False otherwise.

store_by: str

galaxy.objectstore.pithos module

galaxy.objectstore.pithos.parse_config_xml(config_xml)

Parse and validate config_xml, return dict for convenience :param config_xml: (lxml.etree.Element) root of XML subtree :returns: (dict) according to syntax :raises: various XML parse errors

class galaxy.objectstore.pithos.PithosObjectStore(config, config_dict)

Bases: ConcreteObjectStore

Object store that stores objects as items in a Pithos+ container. Cache is ignored for the time being.

store_type: str = 'pithos'

__init__(config, config_dict)

Parameters

config (object) –

An object, most likely populated from galaxy/config.ini, having the following attributes:

• object_store_check_old_style (only used by the DiskObjectStore subclass)

• jobs_directory – Each job is given a unique empty directory as its current working directory. This option defines in what parent directory those directories will be created.

• new_file_path – Used to set the ‘temp’ extra_dir.

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

to_dict()

store_by: str

galaxy.objectstore.pulsar module

class galaxy.objectstore.pulsar.PulsarObjectStore(config, config_xml)

Bases: BaseObjectStore

Object store implementation that delegates to a remote Pulsar server.

This may be more aspirational than practical for now, it would be good to Galaxy to a point that a handler thread could be setup that doesn’t attempt to access the disk files returned by a (this) object store - just passing them along to the Pulsar unmodified. That modification - along with this implementation and Pulsar job destinations would then allow Galaxy to fully manage jobs on remote servers with completely different mount points.

This implementation should be considered beta and may be dropped from Galaxy at some future point or significantly modified.

__init__(config, config_xml)

Parameters

config (object) –

An object, most likely populated from galaxy/config.ini, having the following attributes:

• object_store_check_old_style (only used by the DiskObjectStore subclass)

• jobs_directory – Each job is given a unique empty directory as its current working directory. This option defines in what parent directory those directories will be created.

• new_file_path – Used to set the ‘temp’ extra_dir.

file_ready(obj, **kwds)

Check if a file corresponding to a dataset is ready to be used.

Return True if so, False otherwise

shutdown()

Close any connections for this ObjectStore.

store_by: str

store_type: str

galaxy.objectstore.s3 module

Object Store plugin for the Amazon Simple Storage Service (S3)

galaxy.objectstore.s3.download_directory(bucket, remote_folder, local_path)

galaxy.objectstore.s3.parse_config_xml(config_xml)

class galaxy.objectstore.s3.CloudConfigMixin

Bases: object

class galaxy.objectstore.s3.S3ObjectStore(config, config_dict)

Bases: ConcreteObjectStore, CloudConfigMixin

Object store that stores objects as items in an AWS S3 bucket. A local cache exists that is used as an intermediate location for files between Galaxy and S3.

store_type: str = 's3'

__init__(config, config_dict)

Parameters

config (object) –

An object, most likely populated from galaxy/config.ini, having the following attributes:

• object_store_check_old_style (only used by the DiskObjectStore subclass)

• jobs_directory – Each job is given a unique empty directory as its current working directory. This option defines in what parent directory those directories will be created.

• new_file_path – Used to set the ‘temp’ extra_dir.

start_cache_monitor()

classmethod parse_xml(config_xml)

Parse an XML description of a configuration for this object store.

Return a configuration dictionary (such as would correspond to the YAML configuration) for the object store.

to_dict()

file_ready(obj, **kwargs)

A helper method that checks if a file corresponding to a dataset is ready and available to be used. Return True if so, False otherwise.

shutdown()

Close any connections for this ObjectStore.

store_by: str

class galaxy.objectstore.s3.SwiftObjectStore(config, config_dict)

Bases: S3ObjectStore

Object store that stores objects as items in a Swift bucket. A local cache exists that is used as an intermediate location for files between Galaxy and Swift.

store_type: str = 'swift'

store_by: str

galaxy.objectstore.s3_multipart_upload module

Split large file into multiple pieces for upload to S3. This parallelizes the task over available cores using multiprocessing. Code mostly taken form CloudBioLinux.

galaxy.objectstore.s3_multipart_upload.mp_from_ids(s3server, mp_id, mp_keyname, mp_bucketname)

Get the multipart upload from the bucket and multipart IDs.

This allows us to reconstitute a connection to the upload from within multiprocessing functions.

galaxy.objectstore.s3_multipart_upload.transfer_part(s3server, mp_id, mp_keyname, mp_bucketname, i, part)

Transfer a part of a multipart upload. Designed to be run in parallel.

galaxy.objectstore.s3_multipart_upload.multipart_upload(s3server, bucket, s3_key_name, tarball, mb_size)

Upload large files using Amazon’s multipart upload functionality.