Loader¶

Loader fetches data from data sources. Each Loader class is implemented for its corresponding data source. For example, you can use CSVLoader class to load CSV dataset.

from jubakit.loader.csv import CSVLoader

loader = CSVLoader('/path/to/dataset.csv')

The Loader outputs a dict-like (key-value) object for each record loaded from the data source:

>>> for record in loader:
...   print(record)
{'name': 'John', 'age': '24', 'gender': 'male'}
{'name': 'Jane', 'age': '35', 'gender': 'female'}
{'name': 'Mary', 'age': '19', 'gender': 'female'}

List of Loaders¶

Loaders for the following data sources are bundled with Jubakit.

Plain line-based streams and files – jubakit.loader.core
Array-like (Python lists, NumPy arrays, etc.) data structure – jubakit.loader.array
SciPy sparse matrix data structure – jubakit.loader.sparse
CSV files – jubakit.loader.csv
PostgreSQL data – jubakit.loader.postgresql
Twitter streams – jubakit.loader.twitter
Chain Loaders (that wraps other Loaders) – jubakit.loader.chain

You can extend these Loaders or even write your own one. See Loader Development Guide for details.