w3lib Package

encoding Module

Functions for handling encoding of web pages

w3lib.encoding.html_body_declared_encoding(html_body_str: str | bytes) → str | None

Return the encoding specified in meta tags in the html body, or None if no suitable encoding was found

>>> import w3lib.encoding
>>> w3lib.encoding.html_body_declared_encoding(
... """<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
...      "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
... <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
... <head>
...     <title>Some title</title>
...     <meta http-equiv="content-type" content="text/html;charset=utf-8" />
... </head>
... <body>
... ...
... </body>
... </html>""")
'utf-8'
>>>

w3lib.encoding.html_to_unicode(content_type_header: str | None, html_body_str: bytes, default_encoding: str = 'utf8', auto_detect_fun: Callable[[bytes], str | None] | None = None) → tuple[str, str]

Convert raw html bytes to unicode

This attempts to make a reasonable guess at the content encoding of the html body, following a similar process to a web browser.

It will try in order:

• BOM (byte-order mark)

• http content type header

• meta or xml tag declarations

• auto-detection, if the auto_detect_fun keyword argument is not None

• default encoding in keyword arg (which defaults to utf8)

If an encoding other than the auto-detected or default encoding is used, overrides will be applied, converting some character encodings to more suitable alternatives.

If a BOM is found matching the encoding, it will be stripped.

The auto_detect_fun argument can be used to pass a function that will sniff the encoding of the text. This function must take the raw text as an argument and return the name of an encoding that python can process, or None. To use chardet, for example, you can define the function as:

auto_detect_fun=lambda x: chardet.detect(x).get('encoding')

or to use UnicodeDammit (shipped with the BeautifulSoup library):

auto_detect_fun=lambda x: UnicodeDammit(x).originalEncoding

If the locale of the website or user language preference is known, then a better default encoding can be supplied.

If content_type_header is not present, None can be passed signifying that the header was not present.

This method will not fail, if characters cannot be converted to unicode, \ufffd (the unicode replacement character) will be inserted instead.

Returns a tuple of (<encoding used>, <unicode_string>)

Examples:

>>> import w3lib.encoding
>>> w3lib.encoding.html_to_unicode(None,
... b"""<!DOCTYPE html>
... <head>
... <meta charset="UTF-8" />
... <meta name="viewport" content="width=device-width" />
... <title>Creative Commons France</title>
... <link rel='canonical' href='http://creativecommons.fr/' />
... <body>
... <p>Creative Commons est une organisation \xc3\xa0 but non lucratif
... qui a pour dessein de faciliter la diffusion et le partage des oeuvres
... tout en accompagnant les nouvelles pratiques de cr\xc3\xa9ation \xc3\xa0 l\xe2\x80\x99\xc3\xa8re numerique.</p>
... </body>
... </html>""")
('utf-8', '<!DOCTYPE html>\n<head>\n<meta charset="UTF-8" />\n<meta name="viewport" content="width=device-width" />\n<title>Creative Commons France</title>\n<link rel=\'canonical\' href=\'http://creativecommons.fr/\' />\n<body>\n<p>Creative Commons est une organisation \xe0 but non lucratif\nqui a pour dessein de faciliter la diffusion et le partage des oeuvres\ntout en accompagnant les nouvelles pratiques de cr\xe9ation \xe0 l\u2019\xe8re numerique.</p>\n</body>\n</html>')
>>>

w3lib.encoding.http_content_type_encoding(content_type: str | None) → str | None

Extract the encoding in the content-type header

>>> import w3lib.encoding
>>> w3lib.encoding.http_content_type_encoding("Content-Type: text/html; charset=ISO-8859-4")
'iso8859-4'

w3lib.encoding.read_bom(data: bytes) → tuple[None, None] | tuple[str, bytes]

Read the byte order mark in the text, if present, and return the encoding represented by the BOM and the BOM.

If no BOM can be detected, (None, None) is returned.

>>> import w3lib.encoding
>>> w3lib.encoding.read_bom(b'\xfe\xff\x6c\x34')
('utf-16-be', b'\xfe\xff')
>>> w3lib.encoding.read_bom(b'\xff\xfe\x34\x6c')
('utf-16-le', b'\xff\xfe')
>>> w3lib.encoding.read_bom(b'\x00\x00\xfe\xff\x00\x00\x6c\x34')
('utf-32-be', b'\x00\x00\xfe\xff')
>>> w3lib.encoding.read_bom(b'\xff\xfe\x00\x00\x34\x6c\x00\x00')
('utf-32-le', b'\xff\xfe\x00\x00')
>>> w3lib.encoding.read_bom(b'\x01\x02\x03\x04')
(None, None)
>>>

w3lib.encoding.resolve_encoding(encoding_alias: str) → str | None

Return the encoding that encoding_alias maps to, or None if the encoding cannot be interpreted

>>> import w3lib.encoding
>>> w3lib.encoding.resolve_encoding('latin1')
'cp1252'
>>> w3lib.encoding.resolve_encoding('gb_2312-80')
'gb18030'
>>>

w3lib.encoding.to_unicode(data_str: bytes, encoding: str) → str

Convert a str object to unicode using the encoding given

Characters that cannot be converted will be converted to \ufffd (the unicode replacement character).

html Module

Functions for dealing with markup text

w3lib.html.get_base_url(text: str | bytes, baseurl: str | bytes = '', encoding: str = 'utf-8', *, max_scan: int | None = None) → str

Return the base url if declared in the given HTML text, relative to the given base url.

Added in version 2.5.0: The max_scan parameter.

If no base url is found, the given baseurl is returned.

max_scan is an upper bound on how much of text to look at, in characters, or in bytes if text is a byte string, defaulting to None, i.e. no bound. It bounds work, not results: a tag that starts before the limit is read to its end. 0 looks at nothing, and a negative value raises ValueError.

w3lib.html.get_meta_refresh(text: str | bytes, baseurl: str = '', encoding: str = 'utf-8', ignore_tags: Iterable[str] = ('script', 'noscript'), *, max_scan: int | None = None) → tuple[None, None] | tuple[float, str]

Return the http-equiv parameter of the HTML meta element from the given HTML text and return a tuple (interval, url) where interval is a float containing the delay in seconds (or zero if not present) and url is a string with the absolute url to redirect.

Added in version 2.5.0: The max_scan parameter.

If no meta redirect is found, (None, None) is returned.

max_scan works as in get_base_url().

w3lib.html.remove_comments(text: str | bytes, encoding: str | None = None) → str

Remove HTML Comments.

>>> import w3lib.html
>>> w3lib.html.remove_comments(b"test <!--textcoment--> whatever")
'test  whatever'
>>>

w3lib.html.remove_tags(text: str | bytes, which_ones: Iterable[str] = (), keep: Iterable[str] = (), encoding: str | None = None) → str

Remove HTML Tags only.

which_ones and keep are both tuples, there are four cases:

which_ones	keep	what it does
not empty	empty	remove all tags in which_ones
empty	not empty	remove all tags except the ones in keep
empty	empty	remove all tags
not empty	not empty	not allowed

Remove all tags:

>>> import w3lib.html
>>> doc = '<div><p><b>This is a link:</b> <a href="http://www.example.com">example</a></p></div>'
>>> w3lib.html.remove_tags(doc)
'This is a link: example'
>>>

Keep only some tags:

>>> w3lib.html.remove_tags(doc, keep=('div',))
'<div>This is a link: example</div>'
>>>

Remove only specific tags:

>>> w3lib.html.remove_tags(doc, which_ones=('a','b'))
'<div><p>This is a link: example</p></div>'
>>>

You can’t remove some and keep some:

>>> w3lib.html.remove_tags(doc, which_ones=('a',), keep=('p',))
Traceback (most recent call last):
    ...
ValueError: Cannot use both which_ones and keep
>>>

w3lib.html.remove_tags_with_content(text: str | bytes, which_ones: Iterable[str] = (), encoding: str | None = None) → str

Remove tags and their content.

which_ones is a tuple of which tags to remove including their content. If is empty, returns the string unmodified.

>>> import w3lib.html
>>> doc = '<div><p><b>This is a link:</b> <a href="http://www.example.com">example</a></p></div>'
>>> w3lib.html.remove_tags_with_content(doc, which_ones=('b',))
'<div><p> <a href="http://www.example.com">example</a></p></div>'
>>>

w3lib.html.replace_entities(text: str | bytes, keep: Iterable[str] = (), remove_illegal: bool = True, encoding: str = 'utf-8') → str

Remove entities from the given text by converting them to their corresponding unicode character.

text can be a unicode string or a byte string encoded in the given encoding (which defaults to ‘utf-8’).

If keep is passed (with a list of entity names) those entities will be kept (they won’t be removed).

It supports both numeric entities (&#nnnn; and &#hhhh;) and named entities (such as   or >).

If remove_illegal is True, entities that can’t be converted are removed. If remove_illegal is False, entities that can’t be converted are kept “as is”. For more information see the tests.

Always returns a unicode string (with the entities removed).

>>> import w3lib.html
>>> w3lib.html.replace_entities(b'Price: £100')
'Price: \xa3100'
>>> print(w3lib.html.replace_entities(b'Price: £100'))
Price: £100
>>>

w3lib.html.replace_escape_chars(text: str | bytes, which_ones: Iterable[str] = ('\n', '\t', '\r'), replace_by: str | bytes = '', encoding: str | None = None) → str

Remove escape characters.

which_ones is a tuple of which escape characters we want to remove. By default removes \n, \t, \r.

replace_by is the string to replace the escape characters by. It defaults to '', meaning the escape characters are removed.

w3lib.html.replace_tags(text: str | bytes, token: str = '', encoding: str | None = None) → str

Replace all markup tags found in the given text by the given token. By default token is an empty string so it just removes all tags.

text can be a unicode string or a regular string encoded as encoding (or 'utf-8' if encoding is not given.)

Always returns a unicode string.

Examples:

>>> import w3lib.html
>>> w3lib.html.replace_tags('This text contains <a>some tag</a>')
'This text contains some tag'
>>> w3lib.html.replace_tags('<p>Je ne parle pas <b>fran\xe7ais</b></p>', ' -- ', 'latin-1')
' -- Je ne parle pas  -- fran\xe7ais --  -- '
>>>

w3lib.html.strip_html5_whitespace(text: str) → str

Strip all leading and trailing space characters (as defined in https://www.w3.org/TR/html5/infrastructure.html#space-character).

Such stripping is useful e.g. for processing HTML element attributes which contain URLs, like href, src or form action - HTML5 standard defines them as “valid URL potentially surrounded by spaces” or “valid non-empty URL potentially surrounded by spaces”.

>>> strip_html5_whitespace(' hello\n')
'hello'

w3lib.html.unquote_markup(text: str | bytes, keep: Iterable[str] = (), remove_illegal: bool = True, encoding: str | None = None) → str

This function receives markup as a text (always a unicode string or a UTF-8 encoded string) and does the following:

1. removes entities (except the ones in keep) from any part of it

   that is not inside a CDATA

2. searches for CDATAs and extracts their text (if any) without modifying it.

3. removes the found CDATAs

http Module

w3lib.http.basic_auth_header(username: str | bytes, password: str | bytes, encoding: str = 'ISO-8859-1') → bytes

Return an Authorization header field value for HTTP Basic Access Authentication (RFC 2617)

>>> import w3lib.http
>>> w3lib.http.basic_auth_header('someuser', 'somepass')
b'Basic c29tZXVzZXI6c29tZXBhc3M='

w3lib.http.headers_dict_to_raw(headers_dict: Mapping[bytes, Any | Sequence[bytes]]) → bytes

w3lib.http.headers_dict_to_raw(headers_dict: None) → None

Returns a raw HTTP headers representation of headers

For example:

>>> import w3lib.http
>>> w3lib.http.headers_dict_to_raw({b'Content-type': b'text/html', b'Accept': b'gzip'})
'Content-type: text/html\\r\\nAccept: gzip'
>>>

Note that keys and values must be bytes.

Argument is None (returns None):

>>> w3lib.http.headers_dict_to_raw(None)
>>>

w3lib.http.headers_raw_to_dict(headers_raw: bytes) → MutableMapping[bytes, list[bytes]]

w3lib.http.headers_raw_to_dict(headers_raw: None) → None

Convert raw headers (single multi-line bytestring) to a dictionary.

For example:

>>> import w3lib.http
>>> w3lib.http.headers_raw_to_dict(b"Content-type: text/html\n\rAccept: gzip\n\n")
{'Content-type': ['text/html'], 'Accept': ['gzip']}

Incorrect input:

>>> w3lib.http.headers_raw_to_dict(b"Content-typt gzip\n\n")
{}
>>>

Argument is None (return None):

>>> w3lib.http.headers_raw_to_dict(None)
>>>

url Module

This module contains general purpose URL functions not found in the standard library.

class w3lib.url.ParseDataURIResult(media_type: str, media_type_parameters: dict[str, str], data: bytes)

Named tuple returned by parse_data_uri().

data: bytes

Data, decoded if it was encoded in base64 format.

media_type: str

MIME type type and subtype, separated by / (e.g. "text/plain").

media_type_parameters: dict[str, str]

MIME type parameters (e.g. {"charset": "US-ASCII"}).

w3lib.url.add_http_if_no_scheme(url: str) → str

Add http as the default scheme if it is missing from url.

w3lib.url.add_or_replace_parameter(url: str, name: str, new_value: str, *, separator: str = '&') → str

Add or remove a parameter to a given url

>>> import w3lib.url
>>> w3lib.url.add_or_replace_parameter('http://www.example.com/index.php', 'arg', 'v')
'http://www.example.com/index.php?arg=v'
>>> w3lib.url.add_or_replace_parameter('http://www.example.com/index.php?arg1=v1&arg2=v2&arg3=v3', 'arg4', 'v4')
'http://www.example.com/index.php?arg1=v1&arg2=v2&arg3=v3&arg4=v4'
>>> w3lib.url.add_or_replace_parameter('http://www.example.com/index.php?arg1=v1&arg2=v2&arg3=v3', 'arg3', 'v3new')
'http://www.example.com/index.php?arg1=v1&arg2=v2&arg3=v3new'
>>> w3lib.url.add_or_replace_parameter('http://www.example.com/index.php?arg1=v1;arg2=v2', 'arg3', 'v3', separator=';')
'http://www.example.com/index.php?arg1=v1;arg2=v2;arg3=v3'
>>>

separator is used both to split the existing query and to join the resulting one.

w3lib.url.add_or_replace_parameters(url: str, new_parameters: dict[str, str], *, separator: str = '&') → str

Add or remove a parameters to a given url

>>> import w3lib.url
>>> w3lib.url.add_or_replace_parameters('http://www.example.com/index.php', {'arg': 'v'})
'http://www.example.com/index.php?arg=v'
>>> args = {'arg4': 'v4', 'arg3': 'v3new'}
>>> w3lib.url.add_or_replace_parameters('http://www.example.com/index.php?arg1=v1&arg2=v2&arg3=v3', args)
'http://www.example.com/index.php?arg1=v1&arg2=v2&arg3=v3new&arg4=v4'
>>>

w3lib.url.any_to_uri(uri_or_path: str) → str

If given a path name, return its File URI, otherwise return it unmodified

w3lib.url.canonicalize_url(url: str | bytes | ParseResult, keep_blank_values: bool = True, keep_fragments: bool = False, encoding: str | None = None, *, query_separator: str = '&') → str

Return a normalized form of url, to compare or deduplicate URLs.

Changed in version 2.5.0: Dot segments (. and ..) in the path are now resolved, and IPv6 addresses in the host are now normalized.

URLs that differ only in ways that servers usually ignore get the same canonical form. The canonical form may point to a different resource than url, so to send a request, use safe_url_string() on url instead.

The following changes are applied:

• make the URL safe, as safe_url_string() does

• lowercase the host

• sort query arguments, first by key, then by value

• normalize all spaces (in query arguments) ‘+’ (plus symbol)

• normalize percent encodings case (%2f -> %2F)

• remove query arguments with blank values (unless keep_blank_values is True)

• remove fragments (unless keep_fragments is True)

• resolve dot segments (. and ..) in the path

• normalize IPv6 addresses in the host

>>> import w3lib.url
>>>
>>> # sorting query arguments
>>> w3lib.url.canonicalize_url('http://www.example.com/do?c=3&b=5&b=2&a=50')
'http://www.example.com/do?a=50&b=2&b=5&c=3'
>>>
>>> # UTF-8 conversion + percent-encoding of non-ASCII characters
>>> w3lib.url.canonicalize_url('http://www.example.com/r\u00e9sum\u00e9')
'http://www.example.com/r%C3%A9sum%C3%A9'
>>>
>>> # a query separator other than the default '&'
>>> w3lib.url.canonicalize_url('http://www.example.com/do?c=3;a=50', query_separator=';')
'http://www.example.com/do?a=50;c=3'
>>>
>>> # resolving dot segments and normalizing an IPv6 address
>>> w3lib.url.canonicalize_url('http://[::0:1]/a/../b')
'http://[::1]/b'
>>>

For more examples, see the tests in tests/test_url.py.

w3lib.url.file_uri_to_path(uri: str) → str

Convert File URI to local filesystem path according to: http://en.wikipedia.org/wiki/File_URI_scheme

w3lib.url.parse_data_uri(uri: str | bytes) → ParseDataURIResult

Parse a data: URI into ParseDataURIResult.

w3lib.url.path_to_file_uri(path: str | PathLike[str]) → str

Convert local filesystem path to legal File URIs as described in: http://en.wikipedia.org/wiki/File_URI_scheme

w3lib.url.safe_url_string(url: str | bytes, encoding: str = 'utf8', path_encoding: str = 'utf8', quote_path: bool = True) → str

Return url with its syntax fixed, ready to be sent in a request.

Use it on URLs found in HTML or given by users before requesting them. To compare or deduplicate URLs, use canonicalize_url() instead.

url is parsed the way web browsers parse it, following the URL living standard, and during serialization additional characters are percent-encoded, so that the result is valid by all of the following URL standards known to be enforced by modern-day web browsers and web servers:

• URL living standard

• RFC 3986

• RFC 2396 and RFC 2732, as interpreted by Java 8’s java.net.URI class.

Only the syntax is fixed: any scheme (e.g. file or javascript), any host (e.g. localhost, a private IP address or a domain name that resolves to one) and relative URLs are allowed. If url comes from an untrusted source, it is up to you to check those.

If url is bytes, it is decoded with encoding. encoding is also used to percent-encode the query and the fragment, so it should be the encoding of the page where url was found. path_encoding is used to percent-encode the path, unless quote_path is False, in which case the path is left as is.

Calling this function on an already safe URL returns the URL unmodified.

w3lib.url.url_query_cleaner(url: str | bytes, parameterlist: str | Sequence[str] = (), sep: str = '&', kvsep: str = '=', remove: bool = False, unique: bool = True, keep_fragments: bool = False) → str

Clean URL arguments leaving only those passed in the parameterlist keeping order

>>> import w3lib.url
>>> w3lib.url.url_query_cleaner("product.html?id=200&foo=bar&name=wired", ('id',))
'product.html?id=200'
>>> w3lib.url.url_query_cleaner("product.html?id=200&foo=bar&name=wired", ['id', 'name'])
'product.html?id=200&name=wired'
>>>

If unique is False, do not remove duplicated keys

>>> w3lib.url.url_query_cleaner("product.html?d=1&e=b&d=2&d=3&other=other", ['d'], unique=False)
'product.html?d=1&d=2&d=3'
>>>

If remove is True, leave only those not in parameterlist.

>>> w3lib.url.url_query_cleaner("product.html?id=200&foo=bar&name=wired", ['id'], remove=True)
'product.html?foo=bar&name=wired'
>>> w3lib.url.url_query_cleaner("product.html?id=2&foo=bar&name=wired", ['id', 'foo'], remove=True)
'product.html?name=wired'
>>>

By default, URL fragments are removed. If you need to preserve fragments, pass the keep_fragments argument as True.

>>> w3lib.url.url_query_cleaner('http://domain.tld/?bla=123#123123', ['bla'], remove=True, keep_fragments=True)
'http://domain.tld/#123123'

w3lib.url.url_query_parameter(url: str | bytes, parameter: str, default: None = None, keep_blank_values: bool | int = 0, *, separator: str = '&') → str | None

w3lib.url.url_query_parameter(url: str | bytes, parameter: str, default: str, keep_blank_values: bool | int = 0, *, separator: str = '&') → str

Return the value of a url parameter, given the url and parameter name

If url contains multiple parameters, the first leftmost one is returned.

General case:

>>> import w3lib.url
>>> w3lib.url.url_query_parameter("product.html?id=200&foo=bar", "id")
'200'
>>>

Return a default value if the parameter is not found:

>>> w3lib.url.url_query_parameter("product.html?id=200&foo=bar", "notthere", "mydefault")
'mydefault'
>>>

Returns None if keep_blank_values not set or 0 (default):

>>> w3lib.url.url_query_parameter("product.html?id=", "id")
>>>

Returns an empty string if keep_blank_values set to 1:

>>> w3lib.url.url_query_parameter("product.html?id=", "id", keep_blank_values=1)
''
>>>