lenskit.data

Data abstractions and data set access.

Submodules

accum	Data accumulation support
amazon	Legacy location of the Amazon import functions.
matrix	Classes for working with matrix data.
movielens	Legacy location of the MovieLens import functions.
msweb	Legacy location of the MS Web import functions.
repr	Utility functions for implementing __str__ and __repr__ methods
schema	Pydantic models for LensKit data schemas. These models define define the data
sources	Import code for various data sources.
types	Basic data types used in data representations.

Attributes

GenericKey	A generic collection key with no bounds or type information. Key types must
QueryInput	Types that can be converted to a query by RecQuery.create().
QueryItemSource	Valid sources for query items.

Classes

EntityAttribute	Base class for an attribute associated with an entity class. This class
BatchedRange	Iterator over a range by batches.
DatasetBuilder	Construct data sets from data and tables.
ItemListCollection	A collection of item lists. This protocol defines read access to the
ItemListCollector	Collect item lists with associated keys, as in ItemListCollection.
ListILC	Mutable item list collection backed by a Python list.
MutableItemListCollection	Intersection type of ItemListCollection and
QueryIDKey	Key type for query IDs. This is used for :ref:`item list collections
UserIDKey	Key type for user IDs. This is used for :ref:`item list collections
DataContainer	A general container for the data backing a dataset.
Dataset	Representation of a data set for LensKit training, evaluation, etc. Data can
EntitySet	Representation of a set of entities from the dataset. Obtained from
ItemList	Representation of a (usually ordered) list of items, possibly with scores
RecQuery	Representation of a the data available for a recommendation query. This is
MatrixRelationshipSet	Two-entity relationships without duplicates, accessible in matrix form.
RelationshipSet	Representation for a set of relationship records. This is the class for
Vocabulary	Vocabularies of entity identifiers for the LensKit data model.

Functions

from_interactions_df(df, *[, user_col, item_col, ...])	Create a dataset from a data frame of ratings or other user-item
key_dict(kt)
flatten_dict(data)
unflatten_dict(data, *[, sep])
xopen(…)	Open a file with the appropriate compression mode.

Package Contents

lenskit.data.from_interactions_df(df, *, user_col=None, item_col=None, rating_col=None, timestamp_col=None, users=None, items=None, class_name='rating', timestamp_unit=None)

Create a dataset from a data frame of ratings or other user-item interactions.

Stability:

Caller (see Stability Levels).

Parameters:

• df (pandas.DataFrame) – The user-item interactions (e.g. ratings). The dataset code takes ownership of this data frame and may modify it.

• user_col (str | None) – The name of the user ID column. By default, looks for columns named user, user_id, or userId, with several case variants.

• item_col (str | None) – The name of the item ID column. By default, looks for columns named item, item_id, or itemId, with several case variants.

• rating_col (str | None) – The name of the rating column.

• timestamp_col (str | None) – The name of the timestamp column.

• timestamp_unit (Literal['s', 'ms', 'us', 'ns'] | None) – The unit of numeric timestamp values. If omitted, integer units are inferred and floating-point values are assumed to be seconds.

• user_ids – A vocabulary of user IDs. The data frame is subset to this set of IDs.

• item_ids – A vocabulary of item IDs. The data frame is subset to this set of IDs.

• name – The interaction class name.

• users (lenskit.data.types.IDSequence | pandas.Index | collections.abc.Iterable[lenskit.data.types.ID] | lenskit.data._vocab.Vocabulary | None)

• items (lenskit.data.types.IDSequence | pandas.Index | collections.abc.Iterable[lenskit.data.types.ID] | lenskit.data._vocab.Vocabulary | None)

• class_name (str)

Returns:

The initiated data set.

Return type:

lenskit.data._dataset.Dataset

type lenskit.data.GenericKey = tuple[ID, ...]

A generic collection key with no bounds or type information. Key types must also be named tuples (the Python type system does not allow us to express this).

lenskit.data.key_dict(kt)

Parameters:

kt (tuple[lenskit.data.types.ID, ...])

Return type:

dict[str, Any]

lenskit.data.flatten_dict(data)

Parameters:

data (dict[str, Any])

Return type:

dict[str, Any]

lenskit.data.unflatten_dict(data, *, sep='.')

Parameters:

data (dict[str, Any])

Return type:

dict[str, Any]

type lenskit.data.QueryInput = RecQuery | ID | ItemList | None

Types that can be converted to a query by RecQuery.create().

type lenskit.data.QueryItemSource = Literal['history', 'session', 'context']

Valid sources for query items.

lenskit.data.xopen(path: pathlib.Path | os.PathLike[str] | str, mode: Literal['r', 'rb', 'w', 'wb', 'a', 'ab'] = 'r') → BinaryIO

lenskit.data.xopen(path: pathlib.Path | os.PathLike[str] | str, mode: Literal['rt', 'wt', 'at']) → TextIO

Open a file with the appropriate compression mode.

This is modeled after the xopen package, but implemented directly without as many external dependencies, preferring Python standard library modules.

Note

Unlike open(), but like gzip.open() and friends, the default mode is binary.

Parameters:

• path – The path to the file to open.

• mode – The mode to open the file in.

Exported Aliases

exception lenskit.data.FieldError

Re-exported alias for lenskit.diagnostics.FieldError.

lenskit.data.load_ms_web()

Re-exported alias for lenskit.data.msweb.load_ms_web().

lenskit.data.load_amazon_ratings()

Re-exported alias for lenskit.data.sources.amazon.load_amazon_ratings().

lenskit.data.load_movielens()

Re-exported alias for lenskit.data.sources.movielens.load_movielens().

lenskit.data.load_movielens_df()

Re-exported alias for lenskit.data.sources.movielens.load_movielens_df().

lenskit.data.load_steam()

Re-exported alias for lenskit.data.sources.steam.load_steam().

lenskit.data.ID

Re-exported alias for lenskit.data.types.ID.

lenskit.data.NPID

Re-exported alias for lenskit.data.types.NPID.

lenskit.data.FeedbackType

Re-exported alias for lenskit.data.types.FeedbackType.