collections.collection

Collection

Bases: SlottedDict[Volume]

A collection of volumes and events, corresponding to an XML file in the data/xml/ directory of the Anthology repo.

Provides dictionary-like functionality mapping volume IDs to Volume objects in the collection.

Info

To create a new collection, use CollectionIndex.create().

Required Attributes:

Name	Type	Description
id	str	The ID of this collection (e.g. "L06" or "2022.emnlp").
parent	CollectionIndex	The parent CollectionIndex instance to which this collection belongs.
path	Path	The path of the XML file representing this collection.

Non-Init Attributes:

Name	Type	Description
event	Optional[Event]	An event represented by this collection.
is_data_loaded	bool	A flag indicating whether the XML file has already been loaded.
is_modified	bool	A flag indicating whether any of the data in this collection has been modified after loading.
raised_error_on_load	bool	A flag indicating whether loading the XML file raised an error. If True, calling load() again won't do anything. This is to prevent the same exceptions being triggered over and over again by other functions accessing this collection (and thereby triggering a load) multiple times.

root property

root

The Anthology instance to which this object belongs.

create_event

create_event(id=None, **kwargs)

Create a new (explicit) Event object in this collection.

Parameters:

Name	Type	Description	Default
id	Optional[str]	The ID of the event; must follow RE_EVENT_ID. If None (default), and this collection has a new-style ID, will generate an event ID based on this (e.g., collection "2022.emnlp" will generate event "emnlp-2022").	None
**kwargs	Any	Any valid list or optional attribute of Event.	{}

Returns:

Type	Description
Event	The created Event object.

Raises:

Type	Description
ValueError	If an explicitly defined event already exists in this collection, or if id was None and this collection has an old-style ID.

Danger

If the event index is loaded and an event with the given ID is already implicitly defined, the newly created event will replace that one, but will inherit its co-located IDs. It is currently not possible to explicitly create an event without also explicitly linking all co-located item IDs to it, but for performance reasons (this linking needs to load the entire Anthology data), it only happens when the event index is loaded. This means that e.g. entirely new proceedings can be created without the performance impact of loading everything, but for adding new events to existing proceedings, the event index should probably be loaded first.

create_volume

create_volume(
    id, title, year=None, type=PROCEEDINGS, **kwargs
)

Create a new Volume object in this collection.

Parameters:

Name	Type	Description	Default
id	str	The ID of the new volume.	required
title	MarkupText | str	The title of the new volume. If given as a string, it will be heuristically parsed for markup.	required
year	Optional[str]	The year of the new volume (optional); if None, will infer the year from this collection's ID.	None
type	VolumeType	Whether this is a journal or proceedings volume; defaults to VolumeType.PROCEEDINGS.	PROCEEDINGS
**kwargs	Any	Any valid list or optional attribute of Volume.	{}

Returns:

Type	Description
Volume	The created Volume object.

Raises:

Type	Description
AnthologyDuplicateIDError	If a volume with the given ID already exists.
ValueError	If this collection has an old-style ID.

get_event

get_event()

An Event explicitly defined in this collection, if any.

load

load()

Loads the XML file belonging to this collection.

papers

papers()

An iterator over all Paper objects in all volumes in this collection.

save

save(path=None, minimal_diff=True)

Saves this collection as an XML file.

Parameters:

Name	Type	Description	Default
path	Optional[StrPath]	The filename to save to. If None, defaults to self.path.	None
minimal_diff	bool	If True (default), will compare against an existing XML file in self.path to minimize the difference, i.e., to prevent noise from changes in the XML that make no semantic difference. See utils.xml.ensure_minimal_diff for details.	True

validate_schema

validate_schema()

Validates the XML file belonging to this collection against the RelaxNG schema.

Raises:

Type	Description
DocumentInvalid	If the XML file does not validate against the schema.

volumes

volumes()

An iterator over all Volume objects in this collection.