streamsightv2.metadata

Dataset

The metadata module allows users to include metadata information corresponding to the dataset. The metadata classes are built on top of the Metadata allowing for easy extension and customization. In this module, we provide the a few metadata that is available from public sources. The programmer is free to add more metadatas as they see fit by defining the abstract methods that must be implemented.

It is important to note that userId and itemId in the metadata module is mapped according to Streamsight’s own internal mapping and not the original userId and itemId in the metadata. Hence, developers should not load the metadata from source but instead implement the metadata class and load the metadata while loading the dataset.

Metadata([filename, base_path])

Example

The following example demonstrates how to load the metadata from the MovieLens100K dataset.

from streamsightv2.datasets.movielens import MovieLens100K

dataset = MovieLens100K(fetch_metadata=True)
data = dataset.load()