url-normalize

A Python library for standardizing and normalizing URLs. Ideal for database deduplication, caching, web crawling, and anywhere you need to ensure that equivalent URLs resolve to the exact same string.

from url_normalize import url_normalize

# Fixes IDN, lowercases host/scheme, removes default ports, resolves path segments
url_normalize("HTTP://User:Pass@www.FOO.com:80///foo/../bar/./baz?q=1#frag")
# -> 'http://User:Pass@www.foo.com/bar/baz?q=1#frag'

Features

url-normalize provides a robust URI normalization function that handles IDN domains, scheme/host lowercasing, and RFC-compliant path normalization.

• IDN Support: Full internationalized domain name handling (using IDNA2008 with UTS46).
• Humanization: Convert normalized URLs to a readable display format while preserving round-trip normalization.
• RFC Compliance:
  • Proper percent-encoding (minimal, uppercase hex).
  • Dot-segment removal in paths.
  • Default port and authority handling.
  • UTF-8 NFC normalization.
• Configurable Defaults:
  • Customizable default scheme (https by default).
  • Configurable default domain for absolute paths.
• Query Parameter Control:
  • Parameter filtering with allowlists.
  • Support for domain-specific parameter rules.
• Versatile URL Handling: Handles empty strings, double-slash URLs (//domain.tld), and shebang (#!) URLs.
• Developer Friendly:
  • Python 3.10+ compatibility.
  • 100% test coverage.
  • Modern type hints and string handling.

Inspired by Sam Ruby's urlnorm.py.

Installation

Install as a library:

pip install url-normalize

Or install as a standalone CLI tool using uv:

uv tool install url-normalize

Usage

Python API

Basic Normalization

from url_normalize import url_normalize

# Basic normalization (uses https by default)
print(url_normalize("www.foo.com:80/foo"))
# Output: https://www.foo.com/foo

# With custom default scheme
print(url_normalize("www.foo.com/foo", default_scheme="http"))
# Output: http://www.foo.com/foo

Query Parameter Filtering

You can strip out tracking parameters and only keep the ones you care about using allowlists.

# With query parameter filtering enabled (strips all params by default)
print(url_normalize("www.google.com/search?q=test&utm_source=test", filter_params=True))
# Output: https://www.google.com/search?q=test

# With custom parameter allowlist as a list
print(url_normalize(
    "example.com?page=1&id=123&ref=test",
    filter_params=True,
    param_allowlist=["page", "id"]
))
# Output: https://example.com?page=1&id=123

# With domain-specific parameter allowlists
print(url_normalize(
    "example.com?page=1&id=123&ref=test",
    filter_params=True,
    param_allowlist={"example.com": ["page", "id"]}
))
# Output: https://example.com?page=1&id=123

Default Domain & Scheme

Useful for resolving relative URLs found on a specific page.

# With default domain for absolute paths
print(url_normalize("/images/logo.png", default_domain="example.com"))
# Output: https://example.com/images/logo.png

# With default domain and custom scheme
print(url_normalize("/images/logo.png", default_scheme="http", default_domain="example.com"))
# Output: http://example.com/images/logo.png

Humanizing URLs

Convert normalized URLs back into a user-friendly format for display, particularly useful for IDN domains and percent-encoded paths.

from url_normalize import url_humanize

# Human-readable display form that still normalizes back to the same URL
print(url_humanize("https://xn--e1afmkfd.xn--80akhbyknj4f/%D0%A1%D0%BB%D1%83%D0%B6%D0%B5%D0%B1%D0%BD%D0%B0%D1%8F"))
# Output: https://пример.испытание/Служебная

# Humanization accepts the same normalization options
print(url_humanize("/%D0%A1%D0%BB%D1%83%D0%B6%D0%B5%D0%B1%D0%BD%D0%B0%D1%8F", default_domain="xn--e1afmkfd.xn--80akhbyknj4f"))
# Output: https://пример.испытание/Служебная

Command-line Usage

You can also use url-normalize directly from the terminal to process URLs.

$ url-normalize "www.foo.com:80/foo"
# Output: https://www.foo.com/foo

# With custom default scheme
$ url-normalize -s http "www.foo.com/foo"
# Output: http://www.foo.com/foo

# With query parameter filtering
$ url-normalize -f "www.google.com/search?q=test&utm_source=test"
# Output: https://www.google.com/search?q=test

# With custom allowlist
$ url-normalize -f -p page,id "example.com?page=1&id=123&ref=test"
# Output: https://example.com/?page=1&id=123

# With default domain for absolute paths
$ url-normalize -d example.com "/images/logo.png"
# Output: https://example.com/images/logo.png

# With default domain and custom scheme
$ url-normalize -d example.com -s http "/images/logo.png"
# Output: http://example.com/images/logo.png

# Human-readable display form
$ url-normalize -H "https://xn--e1afmkfd.xn--80akhbyknj4f/%D0%A1%D0%BB%D1%83%D0%B6%D0%B5%D0%B1%D0%BD%D0%B0%D1%8F"
# Output: https://пример.испытание/Служебная

# Via uv tool/uvx
$ uvx url-normalize www.foo.com:80/foo
# Output: https://www.foo.com:80/foo

Documentation

For a complete history of changes, see CHANGELOG.md.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License