USPS Address Normalizer

A Python library for parsing and normalizing US addresses to USPS standard format using libpostal's intelligent parsing.

Features

✅ Intelligent Parsing - Uses libpostal ML model to understand address components ✅ USPS Compliant - Applies all USPS Publication 28 standardization rules ✅ Complete Abbreviations - Street suffixes, directionals, secondary units, states ✅ Separate Components - Returns city, state, and ZIP as individual fields ✅ Proper Line Formatting - Separates primary (Line 1) and secondary (Line 2) addressing ✅ Clean Output - ALL CAPS, no periods, normalized spacing

Installation

Prerequisites

Install libpostal C library:

brew install libpostal

Install the Package

# From wheel file
pip install usps_address_normalizer-1.0.0-py3-none-any.whl

# Or from source
pip install .

Usage

Basic Usage

from usps_address_normalizer import normalize_address

# Normalize any address format
address = normalize_address("128 king street floor 3 suite 301 san francisco california 94107")

# Access individual components
print(address.line1)           # 128 KING ST
print(address.line2)           # FL 3 STE 301
print(address.city)            # SAN FRANCISCO
print(address.state)           # CA
print(address.zip_code)        # 94107

Computed Properties

# Get combined city, state, ZIP
print(address.city_state_zip)
# Output: SAN FRANCISCO, CA 94107

# Get full formatted address
print(address.full_address)
# Output:
# 128 KING ST
# FL 3 STE 301
# SAN FRANCISCO, CA 94107

Dictionary Export

# Convert to dictionary
data = address.as_dict()
print(data)
# Output:
# {
#     'line1': '128 KING ST',
#     'line2': 'FL 3 STE 301',
#     'city': 'SAN FRANCISCO',
#     'state': 'CA',
#     'zip_code': '94107',
#     'city_state_zip': 'SAN FRANCISCO, CA 94107',
#     'full_address': '128 KING ST\nFL 3 STE 301\nSAN FRANCISCO, CA 94107'
# }

Examples

Simple Address (No Unit)

address = normalize_address("100 Market Street San Francisco CA 94105")

print(address.line1)    # 100 MARKET ST
print(address.line2)    # (empty string)
print(address.city)     # SAN FRANCISCO
print(address.state)    # CA
print(address.zip_code) # 94105

Complex Address with Multiple Units

address = normalize_address("128 N King St, Floor 3, Suite 301, San Francisco, CA 94107")

print(address.line1)    # 128 N KING ST
print(address.line2)    # FL 3 STE 301
print(address.city)     # SAN FRANCISCO
print(address.state)    # CA
print(address.zip_code) # 94107

With Full State Name

address = normalize_address("456 Main Avenue Apartment 2B New York New York 10001")

print(address.line1)    # 456 MAIN AVE
print(address.line2)    # APT 2B
print(address.city)     # NEW YORK
print(address.state)    # NY (automatically converted)
print(address.zip_code) # 10001

PO Box

address = normalize_address("PO Box 1234 San Francisco CA 94107")

print(address.line1)    # PO BOX 1234
print(address.line2)    # (empty string)
print(address.city)     # SAN FRANCISCO
print(address.state)    # CA
print(address.zip_code) # 94107

USPS Standardization Rules

The library applies all official USPS abbreviations:

Street Suffixes

• STREET → ST
• AVENUE → AVE
• BOULEVARD → BLVD
• DRIVE → DR
• And 60+ more...

Directionals

• NORTH → N
• SOUTH → S
• NORTHEAST → NE
• And all 8 directionals...

Secondary Units

• SUITE → STE
• APARTMENT → APT
• FLOOR → FL
• BUILDING → BLDG
• And 20+ more...

States

• CALIFORNIA → CA
• NEW YORK → NY
• All 50 states + territories

Formatting Rules

• ALL CAPS for all address fields
• Remove periods (P.O. → PO)
• Remove extra spaces
• Standardize abbreviations

Address Line Format

Line 1 (Primary Address):

• Format: [Number] [PreDir] [Street Name] [Suffix] [PostDir]
• Example: 128 N KING ST

Line 2 (Secondary Address):

• Format: [Unit Type] [Unit Number]
• Example: FL 3 STE 301
• Empty string if no secondary addressing

City, State, ZIP:

• Individual fields: city, state, zip_code
• Combined: city_state_zip → SAN FRANCISCO, CA 94107

API Reference

normalize_address(address: str) -> USPSAddress

Main function to normalize an address.

Parameters:

• address (str): Raw address string in any format

Returns:

• USPSAddress: Normalized address object

Raises:

• ImportError: If libpostal is not installed

USPSAddress Class

Attributes:

• line1 (str): Primary delivery address
• line2 (str): Secondary address (or empty string)
• city (str): City name in ALL CAPS
• state (str): Two-letter state code
• zip_code (str): ZIP code

Properties:

• city_state_zip (str): Combined city, state, ZIP
• full_address (str): All lines joined with newlines

Methods:

• as_dict(): Returns dictionary with all components
• __str__(): Returns full_address
• __repr__(): Developer-friendly representation

Why Use This Library?

The Problem with Manual Parsing

Without this library, you'd need:

• ❌ Complex regex patterns for each component
• ❌ Manual lookup tables for all abbreviations
• ❌ Logic to distinguish "FL" (Floor) from "FL" (Florida)
• ❌ Handling multi-word street names
• ❌ Context-aware parsing

The Solution

✅ libpostal provides: Intelligent ML-based parsing (knows "FL 3" is a floor, not Florida) ✅ This library provides: Complete USPS normalization rules and formatting ✅ You get: Clean, standardized addresses ready for databases, mailings, or APIs

Requirements

• Python >= 3.8
• libpostal C library (brew install libpostal)
• postal Python package (installed automatically)

Development

Building from Source

# Install development dependencies
pip install -e ".[dev]"

# Build wheel
python -m build

# Output: dist/usps_address_normalizer-1.0.0-py3-none-any.whl

Running Tests

cd tests
python test_normalizer.py

License

MIT License

Credits

• Built on libpostal for intelligent address parsing
• USPS abbreviations from USPS Publication 28

Version

1.0.0