pyhouse 0.0.13

Lighthouse for Python: a package facilitating the creation of data pipelines.

pyhouse

This is a port of Lighthouse, a library written in Scala, that facilitates the creation of data pipelines that are based on Apache Spark. It also comes with some related convenience functions, like integrations to the AWS parameter store.

This port is targeted at Python and PySpark. It is not an exact port of the Scala code: we add what we need as we go along.

Usage

One of this library’s main usages is to build a class-based data catalog, that supports chaining of sources. For example, if you had a dataset in a text file that needed to be transformed (clean, derive some statistic, …) then you could write this as such:

from pyhouse.datalake.file_system_data_link import FileSystemDataLink

link = FileSystemDataLink(
    environment="dev",
    session = get_spark(),
    path = "s3://bucket-foo/file-bar.csv",
    format="csv",
    savemode="errorifexists",
    partitioned_by=("some-key", "another-key"),
    options={"header": True, "sep": "\t"}
)

link.read().groupBy("client").count().show()

The advantage of such data links becomes clear when there are multiple of them that are combined in a module (the “catalog”): there would be one source of truth that many scripts can refer to. Hardcoded paths scattered across scripts would be a thing of the past.