Public API¶

The only public yarl class is URL:

>>> from yarl import URL

class yarl.URL(arg, *, encoded=False)¶

Represents URL as

[scheme:]//[user[:password]@]host[:port][/path][?query][#fragment]

for absolute URLs and

[/path][?query][#fragment]

for relative ones (Absolute and relative URLs).

The URL structure is:

 http://user:pass@example.com:8042/over/there?name=ferret#nose
 \__/   \__/ \__/ \_________/ \__/\_________/ \_________/ \__/
  |      |    |        |       |      |           |        |
scheme  user password host    port   path       query   fragment

Internally all data are stored as percent-encoded strings for user, path, query and fragment URL parts and IDNA-encoded (RFC 5891) for host.

Constructor and modification operators perform encoding for all parts automatically. The library assumes all data uses UTF-8 for percent-encoded tokens.

>>> URL('http://example.com/path/to/?arg1=a&arg2=b#fragment')
URL('http://example.com/path/to/?arg1=a&arg2=b#fragment')

Unless URL contain the only ascii characters there is no differences.

But for non-ascii case encoding is applied.

>>> str(URL('http://εμπορικόσήμα.eu/шлях/這裡'))
'http://xn--jxagkqfkduily1i.eu/%D1%88%D0%BB%D1%8F%D1%85/%E9%80%99%E8%A3%A1'

The same is true for user, password, query and fragment parts of URL.

Already encoded URL is not changed:

>>> URL('http://xn--jxagkqfkduily1i.eu')
URL('http://xn--jxagkqfkduily1i.eu')

Use human_repr() for getting human readable representation:

>>> url = URL('http://εμπορικόσήμα.eu/шлях/這裡')
>>> str(url)
'http://xn--jxagkqfkduily1i.eu/%D1%88%D0%BB%D1%8F%D1%85/%E9%80%99%E8%A3%A1'
>>> url.human_repr()
'http://εμπορικόσήμα.eu/шлях/這裡'

Note

Sometimes encoding performed by yarl is not acceptable for certain WEB server.

Passing encoded=True parameter prevents URL auto-encoding, user is responsible about URL correctness.

Don’t use this option unless there is no other way for keeping URL attributes not touched.

Any URL manipulations don’t guarantee correct encoding, URL parts could be re-quoted even if encoded parameter was explicitly set.

URL properties¶

There are two kinds of properties: decoded and encoded (with raw_ prefix):

URL.scheme¶

Scheme for absolute URLs, empty string for relative URLs or URLs starting with '//' (Absolute and relative URLs).

>>> URL('http://example.com').scheme
'http'
>>> URL('//example.com').scheme
''
>>> URL('page.html').scheme
''

URL.user¶

Decoded user part of URL, None if user is missing.

>>> URL('http://john@example.com').user
'john'
>>> URL('http://бажан@example.com').user
'бажан'
>>> URL('http://example.com').user is None
True

URL.raw_user¶

Encoded user part of URL, None if user is missing.

>>> URL('http://довбуш@example.com').raw_user
'%D0%B4%D0%BE%D0%B2%D0%B1%D1%83%D1%88'
>>> URL('http://example.com').raw_user is None
True

URL.password¶

Decoded password part of URL, None if user is missing.

>>> URL('http://john:pass@example.com').password
'pass'
>>> URL('http://степан:пароль@example.com').password
'пароль'
>>> URL('http://example.com').password is None
True

URL.raw_password¶

Encoded password part of URL, None if user is missing.

>>> URL('http://user:пароль@example.com').raw_password
'%D0%BF%D0%B0%D1%80%D0%BE%D0%BB%D1%8C'

URL.host¶

Encoded host part of URL, None for relative URLs (Absolute and relative URLs).

Brackets are stripped for IPv6. Host is converted to lowercase, address is validated and converted to compressed form.

>>> URL('http://example.com').host
'example.com'
>>> URL('http://хост.домен').host
'хост.домен'
>>> URL('page.html').host is None
True
>>> URL('http://[::1]').host
'::1'

URL.raw_host¶

IDNA decoded host part of URL, None for relative URLs (Absolute and relative URLs).

>>> URL('http://хост.домен').raw_host
'xn--n1agdj.xn--d1acufc'
>>> URL('http://[::1]').raw_host
'::1'

URL.host_subcomponent¶

RFC 3986 Section 3.2.2 host subcomponent part of URL, None for relative URLs (Absolute and relative URLs).

>>> URL('http://хост.домен').host_subcomponent
'xn--n1agdj.xn--d1acufc'
>>> URL('http://[::1]').host_subcomponent
'[::1]'

Added in version 1.13.

URL.host_port_subcomponent¶

RFC 3986 Section 3.2.2 host and RFC 3986 Section 3.2.3 port subcomponent part of URL, None for relative URLs (Absolute and relative URLs).

Trailing dots are stripped from the host to ensure this value can be used for an HTTP Host header.

The port is omitted if it is the default port for the scheme.

>>> URL('http://хост.домен:81').host_port_subcomponent
'xn--n1agdj.xn--d1acufc:81'
>>> URL('https://[::1]:8443').host_port_subcomponent
'[::1]:8443'
>>> URL('http://example.com./').host_port_subcomponent
'example.com'
>>> URL('http://[::1]').host_port_subcomponent
'[::1]'

Added in version 1.17.

URL.port¶

port part of URL, with scheme-based fallback.

None for relative URLs (Absolute and relative URLs) or for URLs without explicit port and URL.scheme without default port substitution.

>>> URL('http://example.com:8080').port
8080
>>> URL('http://example.com').port
80
>>> URL('page.html').port is None
True

URL.explicit_port¶

explicit_port part of URL, without scheme-based fallback.

None for relative URLs (Absolute and relative URLs) or for URLs without explicit port.

>>> URL('http://example.com:8080').explicit_port
8080
>>> URL('http://example.com').explicit_port is None
True
>>> URL('page.html').explicit_port is None
True

Added in version 1.3.

URL.authority¶

Decoded authority part of URL, a combination of user, password, host, and port.

authority = [ user [ ":" password ] "@" ] host [ ":" port ].

authority is empty string if all parts are missing.

>>> URL('http://john:pass@example.com:8000').authority
'john:pass@example.com:8000'

Added in version 1.5.

URL.raw_authority¶

Encoded authority part of URL, a combination of user, password, host, and port. empty string if all parts are missing.

>>> URL('http://john:pass@хост.домен:8000').raw_authority
'john:pass@xn--n1agdj.xn--d1acufc:8000'

Added in version 1.5.

URL.path¶

Decoded path part of URL, '/' for absolute URLs without path part.

>>> URL('http://example.com/path/to').path
'/path/to'
>>> URL('http://example.com/шлях/сюди').path
'/шлях/сюди'
>>> URL('http://example.com').path
'/'

Warning

In many situations it is important to distinguish between path separators (a literal /) and other forward slashes (a literal %2F). Use URL.path_safe for these cases.

URL.path_safe¶

Similar to URL.path except it doesn’t decode %2F or %25. This allows to distinguish between path separators (/) and encoded slashes (%2F).

Note that %25 is also not decoded to avoid issues with double unquoting of values. e.g. You can unquote the value with URL.path_safe.replace("%2F", "/").replace("%25", %") to get the same result as URL.path(). If the %25 was unquoted, it would be impossible to tell the difference between %2F and %252F.

Added in version 1.12.

URL.path_qs¶

Decoded path part of URL and query string, '/' for absolute URLs without path part.

>>> URL('http://example.com/path/to?a1=a&a2=b').path_qs
'/path/to?a1=a&a2=b'

URL.raw_path_qs¶

Encoded path part of URL and query string, '/' for absolute URLs without path part.

>>> URL('http://example.com/шлях/сюди?ключ=знач').raw_path_qs
'/%D1%88%D0%BB%D1%8F%D1%85/%D1%81%D1%8E%D0%B4%D0%B8?%D0%BA%D0%BB%D1%8E%D1%87=%D0%B7%D0%BD%D0%B0%D1%87'

Added in version 0.15.

URL.raw_path¶

Encoded path part of URL, '/' for absolute URLs without path part.

>>> URL('http://example.com/шлях/сюди').raw_path
'/%D1%88%D0%BB%D1%8F%D1%85/%D1%81%D1%8E%D0%B4%D0%B8'

URL.query_string¶

Decoded query part of URL, empty string if query is missing.

>>> URL('http://example.com/path?a1=a&a2=b').query_string
'a1=a&a2=b'
>>> URL('http://example.com/path?ключ=знач').query_string
'ключ=знач'
>>> URL('http://example.com/path').query_string
''

URL.raw_query_string¶

Encoded query part of URL, empty string if query is missing.

>>> URL('http://example.com/path?ключ=знач').raw_query_string
'%D0%BA%D0%BB%D1%8E%D1%87=%D0%B7%D0%BD%D0%B0%D1%87'

URL.fragment¶

Encoded fragment part of URL, empty string if fragment is missing.

>>> URL('http://example.com/path#fragment').fragment
'fragment'
>>> URL('http://example.com/path#якір').fragment
'якір'
>>> URL('http://example.com/path').fragment
''

URL.raw_fragment¶

Decoded fragment part of URL, empty string if fragment is missing.

>>> URL('http://example.com/path#якір').raw_fragment
'%D1%8F%D0%BA%D1%96%D1%80'

For path and query yarl supports additional helpers:

URL.parts¶

A tuple containing decoded path parts, ('/',) for absolute URLs if path is missing.

>>> URL('http://example.com/path/to').parts
('/', 'path', 'to')
>>> URL('http://example.com/шлях/сюди').parts
('/', 'шлях', 'сюди')
>>> URL('http://example.com').parts
('/',)

URL.raw_parts¶

A tuple containing encoded path parts, ('/',) for absolute URLs if path is missing.

>>> URL('http://example.com/шлях/сюди').raw_parts
('/', '%D1%88%D0%BB%D1%8F%D1%85', '%D1%81%D1%8E%D0%B4%D0%B8')

URL.name¶

The last part of parts.

>>> URL('http://example.com/path/to').name
'to'
>>> URL('http://example.com/шлях/сюди').name
'сюди'
>>> URL('http://example.com/path/').name
''

URL.raw_name¶

The last part of raw_parts.

>>> URL('http://example.com/шлях/сюди').raw_name
'%D1%81%D1%8E%D0%B4%D0%B8'

URL.suffix¶

The file extension of name.

>>> URL('http://example.com/path/to.txt').suffix
'.txt'
>>> URL('http://example.com/шлях.сюди').suffix
'.сюди'
>>> URL('http://example.com/path').suffix
''

URL.raw_suffix¶

The file extension of raw_name.

>>> URL('http://example.com/шлях.сюди').raw_suffix
'.%D1%81%D1%8E%D0%B4%D0%B8'

URL.suffixes¶

A list of name’s file extensions.

>>> URL('http://example.com/path/to.tar.gz').suffixes
('.tar', '.gz')
>>> URL('http://example.com/шлях.тут.ось').suffixes
('.тут', '.ось')
>>> URL('http://example.com/path').suffixes
()

URL.raw_suffixes¶

A list of raw_name’s file extensions.

>>> URL('http://example.com/шлях.тут.ось').raw_suffixes
('.%D1%82%D1%83%D1%82', '.%D0%BE%D1%81%D1%8C')

URL.query¶

A multidict.MultiDictProxy representing parsed query parameters in decoded representation. Empty value if URL has no query part.

>>> URL('http://example.com/path?a1=a&a2=b').query
<MultiDictProxy('a1': 'a', 'a2': 'b')>
>>> URL('http://example.com/path?ключ=знач').query
<MultiDictProxy('ключ': 'знач')>
>>> URL('http://example.com/path').query
<MultiDictProxy()>

Absolute and relative URLs¶

The module supports both absolute and relative URLs.

Absolute URL should start from either scheme or '//'.

URL.absolute¶

A check for absolute URLs.

Return True for absolute ones (having scheme or starting with '//'), False otherwise.

>>> URL('http://example.com').absolute
True
>>> URL('//example.com').absolute
True
>>> URL('/path/to').absolute
False
>>> URL('path').absolute
False

Changed in version 1.9.10: The absolute property is preferred over the is_absolute() method.

New URL generation¶

URL is an immutable object, every operation described in the section generates a new URL instance.

classmethod URL.build(*, scheme=..., authority=..., user=..., password=..., host=..., port=..., path=..., query=..., query_string=..., fragment=..., encoded=False)[source]¶

Creates and returns a new URL:

>>> URL.build(scheme="http", host="example.com")
URL('http://example.com')

>>> URL.build(scheme="http", host="example.com", query={"a": "b"})
URL('http://example.com/?a=b')

>>> URL.build(scheme="http", host="example.com", query_string="a=b")
URL('http://example.com/?a=b')

>>> URL.build()
URL('')

Calling build method without arguments is equal to calling __init__ without arguments.

Note

Only one of query or query_string should be passed then ValueError will be raised.

URL.with_scheme(scheme)[source]¶

Return a new URL with scheme replaced:

>>> URL('http://example.com').with_scheme('https')
URL('https://example.com')

Returned URL may have a different port (default port substitution).

URL.with_user(user)[source]¶

Return a new URL with user replaced, auto-encode user if needed.

Clear user/password if user is None.

>>> URL('http://user:pass@example.com').with_user('new_user')
URL('http://new_user:pass@example.com')
>>> URL('http://user:pass@example.com').with_user('олекса')
URL('http://%D0%BE%D0%BB%D0%B5%D0%BA%D1%81%D0%B0:pass@example.com')
>>> URL('http://user:pass@example.com').with_user(None)
URL('http://example.com')

URL.with_password(password)[source]¶

Return a new URL with password replaced, auto-encode password if needed.

Clear password if None is passed.

>>> URL('http://user:pass@example.com').with_password('пароль')
URL('http://user:%D0%BF%D0%B0%D1%80%D0%BE%D0%BB%D1%8C@example.com')
>>> URL('http://user:pass@example.com').with_password(None)
URL('http://user@example.com')

URL.with_host(host)[source]¶

Return a new URL with host replaced, auto-encode host if needed.

Changing host for relative URLs is not allowed, use URL.join() instead.

>>> URL('http://example.com/path/to').with_host('python.org')
URL('http://python.org/path/to')
>>> URL('http://example.com/path').with_host('хост.домен')
URL('http://xn--n1agdj.xn--d1acufc/path')

URL.with_port(port)[source]¶

Return a new URL with port replaced.

Clear port to default if None is passed.

>>> URL('http://example.com:8888').with_port(9999)
URL('http://example.com:9999')
>>> URL('http://example.com:8888').with_port(None)
URL('http://example.com')

URL.with_path(path, *, keep_query=False, keep_fragment=False)[source]¶

Return a new URL with path replaced, encode path if needed.

If keep_query=True or keep_fragment=True it retains the existing query or fragment in the URL.

Changed in version 1.18: Added keep_query and keep_fragment parameters.

>>> URL('http://example.com/').with_path('/path/to')
URL('http://example.com/path/to')

URL.with_query(query)[source]¶

URL.with_query(**kwargs)

Return a new URL with query part replaced.

Unlike update_query() the method replaces all query parameters.

Accepts any Mapping (e.g. dict, MultiDict instances) or str, auto-encode the argument if needed.

A sequence of (key, value) pairs is supported as well.

Also it can take an arbitrary number of keyword arguments.

Clear query if None is passed.

Note

The library accepts str, float, int and their subclasses except bool as query argument values.

If a mapping such as dict is used, the values may also be list or tuple to represent a key has many values.

Please see Why isn’t boolean supported by the URL query API? for the reason why bool is not supported out-of-the-box.

>>> URL('http://example.com/path?a=b').with_query('c=d')
URL('http://example.com/path?c=d')
>>> URL('http://example.com/path?a=b').with_query({'c': 'd'})
URL('http://example.com/path?c=d')
>>> URL('http://example.com/path?a=b').with_query({'c': [1, 2]})
URL('http://example.com/path?c=1&c=2')
>>> URL('http://example.com/path?a=b').with_query({'кл': 'зн'})
URL('http://example.com/path?%D0%BA%D0%BB=%D0%B7%D0%BD')
>>> URL('http://example.com/path?a=b').with_query(None)
URL('http://example.com/path')
>>> URL('http://example.com/path?a=b&b=1').with_query(b='2')
URL('http://example.com/path?b=2')
>>> URL('http://example.com/path?a=b&b=1').with_query([('b', '2')])
URL('http://example.com/path?b=2')

Changed in version 1.5: Support list and tuple as a query parameter value.

Changed in version 1.6: Support subclasses of int (except bool) and float as a query parameter value.

URL.extend_query(query)[source]¶

URL.extend_query(**kwargs)

Returns a new URL with query part extended.

Unlike update_query(), this method keeps duplicate keys.

Returned URL object will contain query string which extends parts from passed query parts (or parts of parsed query string).

Accepts any Mapping (e.g. dict, MultiDict instances) or str, auto-encode the argument if needed.

A sequence of (key, value) pairs is supported as well.

Also it can take an arbitrary number of keyword arguments.

Returns the same URL if query of None is passed.

Note

The library accepts str, float, int and their subclasses except bool as query argument values.

If a mapping such as dict is used, the values may also be list or tuple to represent a key has many values.

Please see Why isn’t boolean supported by the URL query API? for the reason why bool is not supported out-of-the-box.

>>> URL('http://example.com/path?a=b&b=1').extend_query(b='2')
URL('http://example.com/path?a=b&b=1&b=2')
>>> URL('http://example.com/path?a=b&b=1').extend_query([('b', '2')])
URL('http://example.com/path?a=b&b=1&b=2')
>>> URL('http://example.com/path?a=b&c=e&c=f').extend_query(c='d')
URL('http://example.com/path?a=b&c=e&c=f&c=d')

Added in version 1.11.0.

URL.update_query(query)[source]¶

URL.update_query(**kwargs)

Returns a new URL with query part updated.

Unlike with_query() the method does not replace query completely.

Returned URL object will contain query string which updated parts from passed query parts (or parts of parsed query string).

Accepts any Mapping (e.g. dict, MultiDict instances) or str, auto-encode the argument if needed.

A sequence of (key, value) pairs is supported as well.

Also it can take an arbitrary number of keyword arguments.

Clear query if None is passed.

Mod operator (%) can be used as alternative to the direct call of URL.update_query().

Note

The library accepts str, float, int and their subclasses except bool as query argument values.

If a mapping such as dict is used, the values may also be list or tuple to represent a key has many values.

Please see Why isn’t boolean supported by the URL query API? for the reason why bool is not supported out-of-the-box.

>>> URL('http://example.com/path?a=b').update_query('c=d')
URL('http://example.com/path?a=b&c=d')
>>> URL('http://example.com/path?a=b').update_query({'c': 'd'})
URL('http://example.com/path?a=b&c=d')
>>> URL('http://example.com/path?a=b').update_query({'c': [1, 2]})
URL('http://example.com/path?a=b&c=1&c=2')
>>> URL('http://example.com/path?a=b').update_query({'кл': 'зн'})
URL('http://example.com/path?a=b&%D0%BA%D0%BB=%D0%B7%D0%BD')
>>> URL('http://example.com/path?a=b&b=1').update_query(b='2')
URL('http://example.com/path?a=b&b=2')
>>> URL('http://example.com/path?a=b&b=1').update_query([('b', '2')])
URL('http://example.com/path?a=b&b=2')
>>> URL('http://example.com/path?a=b&c=e&c=f').update_query(c='d')
URL('http://example.com/path?a=b&c=d')
>>> URL('http://example.com/path?a=b').update_query('c=d&c=f')
URL('http://example.com/path?a=b&c=d&c=f')
>>> URL('http://example.com/path?a=b') % {'c': 'd'}
URL('http://example.com/path?a=b&c=d')

Changed in version 1.0: All multiple key/value pairs are applied to the multi-dictionary.

Added in version 1.5: Support for mod operator (%) to update the URL’s query part.

Changed in version 1.5: Support list and tuple as a query parameter value.

Changed in version 1.6: Support subclasses of int (except bool) and float as a query parameter value.

URL.without_query_params(*query_params)[source]¶

Return a new URL whose query part does not contain specified query_params.

Accepts str for query_params.

It does nothing if none of specified query_params are present in the query.

Added in version 1.10.0.

URL.with_fragment(fragment)[source]¶

Return a new URL with fragment replaced, auto-encode fragment if needed.

Clear fragment to default if None is passed.

>>> URL('http://example.com/path#frag').with_fragment('anchor')
URL('http://example.com/path#anchor')
>>> URL('http://example.com/path#frag').with_fragment('якір')
URL('http://example.com/path#%D1%8F%D0%BA%D1%96%D1%80')
>>> URL('http://example.com/path#frag').with_fragment(None)
URL('http://example.com/path')

URL.with_name(name, *, keep_query=False, keep_fragment=False)[source]¶

Return a new URL with name (last part of path) replaced and cleaned up query and fragment parts.

Name is encoded if needed.

If keep_query=True or keep_fragment=True it retains the existing query or fragment in the URL.

Changed in version 1.18: Added keep_query and keep_fragment parameters.

>>> URL('http://example.com/path/to?arg#frag').with_name('new')
URL('http://example.com/path/new')
>>> URL('http://example.com/path/to').with_name("ім'я")
URL('http://example.com/path/%D1%96%D0%BC%27%D1%8F')

URL.with_suffix(suffix, *, keep_query=False, keep_fragment=False)[source]¶

Return a new URL with suffix (file extension of name) replaced and cleaned up query and fragment parts.

Name is encoded if needed.

If keep_query=True or keep_fragment=True it retains the existing query or fragment in the URL.

Changed in version 1.18: Added keep_query and keep_fragment parameters.

>>> URL('http://example.com/path/to?arg#frag').with_suffix('.doc')
URL('http://example.com/path/to.doc')
>>> URL('http://example.com/path/to').with_suffix('.cуфікс')
URL('http://example.com/path/to.c%D1%83%D1%84%D1%96%D0%BA%D1%81')

URL.parent¶

A new URL with last part of path removed and cleaned up query and fragment parts.

>>> URL('http://example.com/path/to?arg#frag').parent
URL('http://example.com/path')

URL.origin()[source]¶

A new URL with scheme, host and port parts only. user, password, path, query and fragment are removed.

>>> URL('http://example.com/path/to?arg#frag').origin()
URL('http://example.com')
>>> URL('http://user:pass@example.com/path').origin()
URL('http://example.com')

URL.relative()[source]¶

A new relative URL with path, query and fragment parts only. scheme, user, password, host and port are removed.

>>> URL('http://example.com/path/to?arg#frag').relative()
URL('/path/to?arg#frag')

Division (/) operator creates a new URL with appended path parts and cleaned up query and fragment parts.

The path is encoded if needed.

>>> url = URL('http://example.com/path?arg#frag') / 'to/subpath'
>>> url
URL('http://example.com/path/to/subpath')
>>> url.parts
('/', 'path', 'to', 'subpath')
>>> url = URL('http://example.com/path?arg#frag') / 'сюди'
>>> url
URL('http://example.com/path/%D1%81%D1%8E%D0%B4%D0%B8')

URL.joinpath(*other, encoded=False)[source]¶

Construct a new URL by with all other elements appended to path, and cleaned up query and fragment parts.

Passing encoded=True parameter prevents path element auto-encoding, the caller is responsible for taking care of URL correctness.

>>> url = URL('http://example.com/path?arg#frag').joinpath('to', 'subpath')
>>> url
URL('http://example.com/path/to/subpath')
>>> url.parts
('/', 'path', 'to', 'subpath')
>>> url = URL('http://example.com/path?arg#frag').joinpath('сюди')
>>> url
URL('http://example.com/path/%D1%81%D1%8E%D0%B4%D0%B8')
>>> url = URL('http://example.com/path').joinpath('%D1%81%D1%8E%D0%B4%D0%B8', encoded=True)
>>> url
URL('http://example.com/path/%D1%81%D1%8E%D0%B4%D0%B8')

Added in version 1.9.

URL.__truediv__(url)[source]¶

Shortcut for URL.joinpath() with a single element and encoded=False.

>>> url = URL('http://example.com/path?arg#frag') / 'to'
>>> url
URL('http://example.com/path/to')
>>> url.parts
('/', 'path', 'to')

Added in version 0.9.

URL.join(url)[source]¶

Construct a full (“absolute”) URL by combining a “base URL” (self) with another URL (url). Informally, this uses components of the base URL, in particular the addressing scheme, the network location and (part of) the path, to provide missing components in the relative URL, e.g.:

>>> base = URL('http://example.com/path/index.html')
>>> base.join(URL('page.html'))
URL('http://example.com/path/page.html')

Note

If url is an absolute URL (that is, starting with // or scheme://), the URL‘s host name and/or scheme will be present in the result, e.g.:

>>> base = URL('http://example.com/path/index.html')
>>> base.join(URL('//python.org/page.html'))
URL('http://python.org/page.html')

Human readable representation¶

All URL data is stored in encoded form internally. It’s pretty good for passing str(url) everywhere URL string is accepted but quite bad for memorizing by humans.

URL.human_repr()[source]¶

Return decoded human readable string for URL representation.

>>> url = URL('http://εμπορικόσήμα.eu/這裡')
>>> str(url)
'http://xn--jxagkqfkduily1i.eu/%E9%80%99%E8%A3%A1'
>>> url.human_repr()
'http://εμπορικόσήμα.eu/這裡'

Default port substitution¶

yarl is aware about the following scheme -> port translations:

scheme	port
`'http'`	80
`'https'`	443
`'ws'`	80
`'wss'`	443

URL.is_default_port()[source]¶

A check for default port.

Return True if URL’s port is default for used scheme, False otherwise.

Relative URLs have no default port.

>>> URL('http://example.com').is_default_port()
True
>>> URL('http://example.com:80').is_default_port()
True
>>> URL('http://example.com:8080').is_default_port()
False
>>> URL('/path/to').is_default_port()
False

Cache control¶

IDNA conversion and host encoding are quite expensive operations, that’s why the yarl library caches these calls by storing results in the global LRU cache.

yarl.cache_clear()¶: Clear IDNA and host encoding cache.

yarl.cache_info()¶

Return a dictionary with "idna_encode", "idna_decode", and "encode_host" keys, each value points to corresponding CacheInfo structure (see functools.lru_cache() for details):

>>> yarl.cache_info()
{'idna_encode': CacheInfo(hits=5, misses=5, maxsize=256, currsize=5),
 'idna_decode': CacheInfo(hits=24, misses=15, maxsize=256, currsize=15),
 'encode_host': CacheInfo(hits=0, misses=0, maxsize=512, currsize=0)}

Changed in version 1.16: ip_address, and host_validate are deprecated in favor of a single encode_host cache.

yarl.cache_configure(*, idna_encode_size=256, idna_decode_size=256, encode_host_size=512)¶

Set the IDNA encode, IDNA decode, and host encode cache sizes.

Pass None to make the corresponding cache unbounded (may speed up host encoding operation a little but the memory footprint can be very high, please use with caution).

Changed in version 1.16: ip_address_size and host_validate_size are deprecated in favor of a single encode_host cache.

References¶

yarl stays on shoulders of giants: several RFC documents and low-level urllib.parse which performs almost all gory work.

The module borrowed design from pathlib in any place where it was possible.