Address Toolkit

Introduction

The address toolkit package supports the cleaning and processing of address data registered as a pyspark.sql.DataFrame. The package includes functions for cleaning, validating, extracting and contextualising addresses and address components. Additionally, workflows have been created as an 'out-of-the-box' application combining the functions across cleaning, validating and extracting.

To install the package, run:

pip install address-toolkit

To register a spark dataframe from a CSV, the following code can be run.

from pyspark.sql import SparkSession


spark = SparkSession.builder.master("local").appName("test").getOrCreate()
df = spark.read.csv("test.csv")

A recommended first step in using this package is to apply the clean_punctuation function from address_toolkit.cleaning.

from address_toolkit.cleaning import clean_punctuation

df = clean_punctuation(df, 'address_column', create_flag = True, overwrite = True)

Main Package Contents:

Supplementary Package Contents (Resources):

Note: For unit address level RegEx patterns, alternatives i.e. flat_regex_alternative are available which are less 'strict' in their matching.

Address Component lists and lookups have been created from Ordnance Survey Open Names. Contains OS data © Crown copyright and database right 2026. Licensed under the Open Government License v3.0. This dataset is updated quarterly in January, April, July and October.

Postcode processing is based upon the Annex C - Valid post code format specification.

Example Usage

See tutorial.ipynb for full use of all the functions within cleaning, validating, extracting, contextualising and workflows.