- Notifications
You must be signed in to change notification settings - Fork12
A helper library full of URL-related heuristics.
License
medialab/ural
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
A python helper library full of URL-related heuristics.
You can installural with pip with the following command:
pip install uralural is published onZenodo as
You can cite it thusly:
Guillaume Plique, Jules Farjas, Oubine Perrin, Benjamin Ooghe-Tabanou, Martin Delabre, Pauline Breteau, Jean Descamps, Béatrice Mazoyer, Amélie Pellé, Laura Miguel, & César Pichon. Ural, a python helper library full of URL-related heuristics. (2018). Zenodo.https://doi.org/10.5281/zenodo.8160139
Generic functions
- canonicalize_url
- could_be_html
- could_be_rss
- ensure_protocol
- fingerprint_hostname
- fingerprint_url
- force_protocol
- format_url
- get_domain_name
- get_hostname
- get_fingerprinted_hostname
- get_normalized_hostname
- has_special_host
- has_valid_suffix
- has_valid_tld
- infer_redirection
- is_homepage
- is_shortened_url
- is_special_host
- is_typo_url
- is_url
- is_valid_tld
- links_from_html
- normalize_hostname
- normalize_url
- should_follow_href
- should_resolve
- split_suffix
- strip_protocol
- urlpathsplit
- urls_from_html
- urls_from_text
Utilities
Classes
LRU-related functions (What on earth is a LRU?)
- lru.url_to_lru
- lru.lru_to_url
- lru.lru_stems
- lru.canonicalized_lru_stems
- lru.normalized_lru_stems
- lru.fingerprinted_lru_stems
- lru.serialize_lru
- lru.unserialize_lru
LRU-related classes
Platform-specific functions
ural comes with three different url deduplication schemes, targeted to different use-cases and ordered hereafter by ascending aggressiveness:
- canonicalize_url: we clean the url by performing some light preprocessing usually done by web browsers before hitting them, e.g. lowercasing the hostname, decoding punycode, ensuring we have a protocol, dropping leading and trailing whitespace etc. The clean url is guaranteed to still lead to the same place.
- normalize_url: we apply more advanced preprocessing that will drop some parts of the url that are irrelevant to where the url leads, such as technical artifacts and SEO tricks. For instance, we will drop typical query items used by marketing campaigns, reorder the query items, infer some redirections, strip trailing slash or fragment when advisable etc. At that point, the url should be clean enough that one can perform meaningful statistical aggregation when counting them, all while ensuring with some good probability that the url still works and still leads to the same place, at least if the target server follows most common conventions.
- fingerprint_url: we go a step further and we perform destructive preprocessing that cannot guarantee that the resulting url will still be valid. But the result might be even more useful for statistical aggregation, especially when counting urls from large platforms having multiple domains (e.g.
facebook.com,facebook.fretc.)
| Function | Use-cases | Url validity | Deduplication strength |
|---|---|---|---|
| canonicalize_url | web crawler | Technically the same | + |
| normalize_url | web crawler, statistical aggregation | Probably the same | ++ |
| fingerprint_url | statistical aggregation | Potentially invalid | +++ |
Example
fromuralimportcanonicalize_url,normalize_url,fingerprint_urlurl='https://www.FACEBOOK.COM:80/index.html?utc_campaign=3&id=34'canonicalize_url(url)>>>'https://www.facebook.com/index.html?utc_campaign=3&id=34'# The same url, cleaned up a littlenormalize_url(url)>>>'facebook.com?id=34'# Still a valid url, with implicit protocol, where all the cruft has been discardedfingerprint_url(url,strip_suffix=True)>>>'facebook?id=34'# Not a valid url anymore, but useful to match more potential# candidates such as: http://facebook.co.uk/index.html?id=34
Function returning a clean and safe version of the url by performing the same kind of preprocessing as web browsers.
For more details about this be sure to readthis section of the docs.
fromuralimportcanonicalize_urlcanonicalize_url('www.LEMONDE.fr')>>>'https://lemonde.fr'
Arguments
- urlstring: url to canonicalize.
- quoted?bool [
False]: by default the function will unquote the url as much as possible all while keeping the url safe. If this kwarg is set toTrue, the function will instead quote the url as much as possible all while ensuring nothing will be double-quoted. - default_protocol?str [
https]: default protocol to add when the url has none. - strip_fragment?str [
False]: whether to strip the url's fragment.
Function returning whether the url could return HTML.
fromuralimportcould_be_htmlcould_be_html('https://www.lemonde.fr')>>>Truecould_be_html('https://www.lemonde.fr/articles/page.php')>>>Truecould_be_html('https://www.lemonde.fr/data.json')>>>Falsecould_be_html('https://www.lemonde.fr/img/figure.jpg')>>>False
Function returning whether the given url could be a rss feed url.
fromuralimportcould_be_rsscould_be_rss('https://www.lemonde.fr/cyclisme/rss_full.xml')>>>Truecould_be_rss('https://www.lemonde.fr/cyclisme/')>>>Falsecould_be_rss('https://www.ecorce.org/spip.php?page=backend')>>>Truecould_be_rss('https://feeds.feedburner.com/helloworld')>>>True
Function checking if the url has a protocol, and adding the given one if there is none.
fromuralimportensure_protocolensure_protocol('www.lemonde.fr',protocol='https')>>>'https://www.lemonde.fr'
Arguments
- urlstring: URL to format.
- protocolstring: protocol to use if there is none inurl. Is 'http' by default.
Function returning a "fingerprinted" version of the given hostname by stripping subdomains irrelevant for statistical aggregation. Be warned that this function is even more aggressive thannormalize_hostname and that the resulting hostname might not be valid anymore.
For more details about this be sure to readthis section of the docs.
fromuralimportfingerprint_hostnamefingerprint_hostname('www.lemonde.fr')>>>'lemonde.fr'fingerprint_hostname('fr-FR.facebook.com')>>>'facebook.com'fingerprint_hostname('fr-FR.facebook.com',strip_suffix=True)>>>'facebook'
Arguments
- hostnamestring: target hostname.
- strip_suffix?bool [
False]: whether to strip the hostname suffix such as.comor.co.uk. This can be useful to aggegate different domains of the same platform.
Function returning a "fingerprinted" version of the given url that can be useful for statistical aggregation. Be warned that this function is even more aggressive thannormalize_url and that the resulting url might not be valid anymore.
For more details about this be sure to readthis section of the docs.
fromuralimportfingerprint_hostnamefingerprint_url('www.lemonde.fr/article.html')>>>'lemonde.fr/article.html'fingerprint_url('fr-FR.facebook.com/article.html')>>>'facebook.com/article.html'fingerprint_url('fr-FR.facebook.com/article.html',strip_suffix=True)>>>'facebook/article.html'
Arguments
- urlstring: target url.
- strip_suffix?bool [
False]: whether to strip the hostname suffix such as.comor.co.uk. This can be useful to aggegate different domains of the same platform. - platform_aware?bool [
False]: whether to take some well-known platforms supported byuralsuch as facebook, youtube etc. into account when normalizing the url.
Function force-replacing the protocol of the given url.
fromuralimportforce_protocolforce_protocol('https://www2.lemonde.fr',protocol='ftp')>>>'ftp://www2.lemonde.fr'
Arguments
- urlstring: URL to format.
- protocolstring: protocol wanted in the output url. Is
'http'by default.
Function formatting a url given some typical parameters.
fromuralimportformat_urlformat_url('https://lemonde.fr',path='/article.html',args={'id':'48675'},fragment='title-2')>>>'https://lemonde.fr/article.html?id=48675#title-2'# Path can be given as an iterableformat_url('https://lemonde.fr',path=['articles','one.html'])>>>'https://lemonde.fr/articles/one.html'# Extensionformat_url('https://lemonde.fr',path=['article'],ext='html')>>>'https://lemonde.fr/articles/article.html'# Query args are formatted/quoted and/or skipped if None/Falseformat_url("http://lemonde.fr",path=["business","articles"],args={"hello":"world","number":14,"boolean":True,"skipped":None,"also-skipped":False,"quoted":"test=ok", },fragment="#test",)>>>'http://lemonde.fr/business/articles?boolean&hello=world&number=14"ed=test%3Dok#test'# Custom argument value formattingdefformat_arg_value(key,value):ifkey=='ids':return','.join(value)returnkeyformat_url('https://lemonde.fr',args={'ids': [1,2]},format_arg_value=format_arg_value)>>>'https://lemonde.fr?ids=1%2C2'# Formatter classfromuralimportURLFormatterformatter=URLFormatter('https://lemonde.fr',args={'id':'one'})formatter(path='/article.html')>>>'https://lemonde.fr/article.html?id=one'# same as:formatter.format(path='/article.html')>>>'https://lemonde.fr/article.html?id=one'# Query arguments are mergedformatter(path='/article.html',args={"user_id":"two"})>>>'https://lemonde.fr/article.html?id=one&user_id=two'# Easy subclassingclassMyCustomFormatter(URLFormatter):BASE_URL='https://lemonde.fr/api'defformat_api_call(self,token):returnself.format(args={'token':token})formatter=MyCustomFormatter()formatter.format_api_call('2764753')>>>'https://lemonde.fr/api?token=2764753'
Arguments
- base_urlstr: Base url.
- path?str|list: the url's path.
- args?dict: query arguments as a dictionary.
- format_arg_value?callable: function taking a query argument key and value and returning the formatted value.
- fragment?str: the url's fragment.
- ext?str: path extension such as
.html.
Function returning an url's domain name. This function is of course tld-aware and will returnNone if no valid domain name can be found.
fromuralimportget_domain_nameget_domain_name('https://facebook.com/path')>>>'facebook.com'
Function returning the given url's full hostname. It can work on scheme-less urls.
fromuralimportget_hostnameget_hostname('http://www.facebook.com/path')>>>'www.facebook.com'
Function returning the "fingerprinted" hostname of the given url by stripping subdomains irrelevant for statistical aggregation. Be warned that this function is even more aggressive thanget_normalized_hostname and that the resulting hostname might not be valid anymore.
For more details about this be sure to readthis section of the docs.
fromuralimportget_normalized_hostnameget_normalized_hostname('https://www.lemonde.fr/article.html')>>>'lemonde.fr'get_normalized_hostname('https://fr-FR.facebook.com/article.html')>>>'facebook.com'get_normalized_hostname('https://fr-FR.facebook.com/article.html',strip_suffix=True)>>>'facebook'
Arguments
- urlstring: target url.
- strip_suffix?bool [
False]: whether to strip the hostname suffix such as.comor.co.uk. This can be useful to aggegate different domains of the same platform.
Function returning the given url's normalized hostname, i.e. without usually irrelevant subdomains etc. Works a lot likenormalize_url.
For more details about this be sure to readthis section of the docs.
fromuralimportget_normalized_hostnameget_normalized_hostname('http://www.facebook.com/path')>>>'facebook.com'get_normalized_hostname('http://fr-FR.facebook.com/path')>>>'facebook.com'
Arguments
- urlstr: Target url.
- infer_redirectionbool [
True]: whether to attempt resolving common redirects by leveraging well-known GET parameters. - normalize_amp?bool [
True]: Whether to attempt to normalize Google AMP subdomains.
Function returning whether the given url looks like it has a special host.
fromuralimporthas_special_hosthas_special_host('http://104.19.154.83')>>>Truehas_special_host('http://youtube.com')>>>False
Function returning whether the given url has a valid suffix as perMozzila's Public Suffix List.
fromuralimporthas_valid_suffixhas_valid_suffix('http://lemonde.fr')>>>Truehas_valid_suffix('http://lemonde.doesnotexist')>>>False# Also works with hostnameshas_valid_suffix('lemonde.fr')>>>True
Function returning whether the given url has a valid Top Level Domain (TLD) as perIANA's list.
fromuralimporthas_valid_tldhas_valid_tld('http://lemonde.fr')>>>Truehas_valid_tld('http://lemonde.doesnotexist')>>>False# Also works with hostnameshas_valid_tld('lemonde.fr')>>>True
Function attempting to find obvious clues in the given url that it is in fact a redirection and resolving the redirection automatically without firing any HTTP request. If nothing is found, the given url will be returned as-is.
The function is by default recursive and will attempt to infer redirections until none is found, but you can disable this behavior if you need to.
fromuralimportinfer_redirectioninfer_redirection('https://www.google.com/url?sa=t&source=web&rct=j&url=https%3A%2F%2Fm.youtube.com%2Fwatch%3Fv%3D4iJBsjHMviQ&ved=2ahUKEwiBm-TO3OvkAhUnA2MBHQRPAR4QwqsBMAB6BAgDEAQ&usg=AOvVaw0i7y2_fEy3nwwdIZyo_qug')>>>'https://m.youtube.com/watch?v=4iJBsjHMviQ'infer_redirection('https://test.com?url=http%3A%2F%2Flemonde.fr%3Fnext%3Dhttp%253A%252F%252Ftarget.fr')>>>'http://target.fr'infer_redirection('https://test.com?url=http%3A%2F%2Flemonde.fr%3Fnext%3Dhttp%253A%252F%252Ftarget.fr',recursive=False)>>>'http://lemonde.fr?next=http%3A%2F%2Ftarget.fr'
Function returning whether the given url isprobably a website's homepage, based on its path.
fromuralimportis_homepageis_homepage('http://lemonde.fr')>>>Trueis_homepage('http://lemonde.fr/index.html')>>>Trueis_homepage('http://lemonde.fr/business/article5.html')>>>False
Function returning whether the given url isprobably a shortened url. It works by matching the given url domain against most prominent shortener domains. So the result could be a false negative.
fromuralimportis_shortened_urlis_shortened_url('http://lemonde.fr')>>>Falseis_shortened_url('http://bit.ly/1sNZMwL')>>>True
Function returning whether the given hostname looks like a special host.
fromuralimportis_special_hostis_special_host('104.19.154.83')>>>Trueis_special_host('youtube.com')>>>False
Function returning whether the given string isprobably a typo error. This function doesn't test if the given string is a valid url. It works by matching the given url tld against most prominent typo-like tlds or by matching the given string against most prominent inclusive language terminations. So the result could be a false negative.
fromuralimportis_typo_urlis_typo_url('http://dirigeants.es')>>>Trueis_typo_url('https://www.instagram.com')>>>False
Function returning whether the given string is a valid url.
fromuralimportis_urlis_url('https://www2.lemonde.fr')>>>Trueis_url('lemonde.fr/economie/article.php',require_protocol=False)>>>Trueis_url('lemonde.falsetld/whatever.html',tld_aware=True)>>>False
Arguments
- stringstring: string to test.
- require_protocolbool [
True]: whether the argument has to have a protocol to be considered a url. - tld_awarebool [
False]: whether to check if the url's tld actually exists or not. - allow_spaces_in_pathbool [
False]: whether the allow spaces in URL paths. - only_http_httpsbool [
True]: whether to only allow thehttpandhttpsprotocols.
Function returning whether the given Top Level Domain (TLD) is valid as perIANA's list.
fromuralimportis_valid_tldis_valid_tld('.fr')>>>Trueis_valid_tld('com')>>>Trueis_valid_tld('.doesnotexist')>>>False
Function returning an iterator over the valid outgoing links present in given HTML text.
This is a variant ofurls_from_html suited to web crawlers. It can deduplicate the urls, canonicalize them, join them with a base url and filter out things that should not be followed such asmailto: or#"auto">Note this function is able to work both on string and bytes seamlessly.
fromuralimportlinks_from_htmlhtml=b"""<p> Hey! Check this site: <a href="https://medialab.sciencespo.fr/">médialab</a> And also this page: <a href="article.html">article</a> Or click on this: <a href="#">link</a></p>"""forlinkinlinks_from_html('http://lemonde.fr',html):print(link)>>>'https://medialab.sciencespo.fr/'>>>'http://lemonde.fr/article.html'
Arguments
- base_urlstring: the HTML's url.
- stringstring|bytes: html string or bytes.
- encoding?string [
utf-8]: if given binary, this encoding will be used to decode the found urls. - canonicalize?bool [
False]: whether to canonicalize the urls usingcanonicalize_url. - strip_fragment?bool [
False]: whether to strip the url fragments when usingcanonicalize. - unique?bool [
False]: whether to deduplicate the urls.
Function normalizing the given hostname, i.e. without usually irrelevant subdomains etc. Works a lot likenormalize_url.
For more details about this be sure to readthis section of the docs.
fromuralimportnormalize_hostnamenormalize_hostname('www.facebook.com')>>>'facebook.com'normalize_hostname('fr-FR.facebook.com')>>>'facebook.com'
Function normalizing the given url by stripping it of usually non-discriminant parts such as irrelevant query items or sub-domains etc.
This is a very useful utility when attempting to match similar urls written slightly differently when shared on social media etc.
For more details about this be sure to readthis section of the docs.
fromuralimportnormalize_urlnormalize_url('https://www2.lemonde.fr/index.php?utm_source=google')>>>'lemonde.fr'
Arguments
- urlstring: URL to normalize.
- infer_redirection?bool [
True]: whether to attempt resolving common redirects by leveraging well-known GET parameters. - fix_common_mistakes?bool [
True]: whether to attempt to fix common URL mistakes. - normalize_amp?bool [
True]: whether to attempt to normalize Google AMP urls. - sort_query?bool [
True]: whether to sort query items. - strip_authentication?bool [
True]: whether to strip authentication. - strip_fragment?bool|str [
'except-routing']: whether to strip the url's fragment. If set toexcept-routing, will only strip the fragment if the fragment is not deemed to be js routing (i.e. if it contains a/). - strip_index?bool [
True]: whether to strip trailing index. - strip_irrelevant_subdomains?bool [
False]: whether to strip irrelevant subdomains such aswwwetc. - strip_protocol?bool [
True]: whether to strip the url's protocol. - strip_trailing_slash?bool [
True]: whether to strip trailing slash. - quoted?bool [
False]: by default the function will unquote the url as much as possible all while keeping the url safe. If this kwarg is set toTrue, the function will instead quote the url as much as possible all while ensuring nothing will be double-quoted. - platform_aware?bool [
False]: whether to take some well-known platforms supported byuralsuch as facebook, youtube etc. into account when normalizing the url.
Function returning whether the given href should be followed (usually from a crawler's context). This means it will filter out anchors, and url having a scheme which is not http/https.
fromuralimportshould_follow_hrefshould_follow_href('#top')>>>Falseshould_follow_href('http://lemonde.fr')>>>Trueshould_follow_href('/article.html')>>>True
Function returning whether the given function looks like something you would want to resolve because the url willprobably lead to some redirection.
It is quite similar tois_shortened_url but covers more ground since it also deal with url patterns which are not shortened per se.
fromuralimportshould_resolveshould_resolve('http://lemonde.fr')>>>Falseshould_resolve('http://bit.ly/1sNZMwL')>>>Trueshould_resolve('https://doi.org/10.4000/vertigo.26405')>>>True
Function splitting a hostname or a url's hostname into the domain part and the suffix part (while respectingMozzila's Public Suffix List).
fromuralimportsplit_suffixsplit_suffix('http://www.bbc.co.uk/article.html')>>> ('www.bbc','co.uk')split_suffix('http://www.bbc.idontexist')>>>Nonesplit_suffix('lemonde.fr')>>> ('lemonde','fr')
Function removing the protocol from the url.
fromuralimportstrip_protocolstrip_protocol('https://www2.lemonde.fr/index.php')>>>'www2.lemonde.fr/index.php'
Arguments
- urlstring: URL to format.
Function taking a url and returning its path, tokenized as a list.
fromuralimporturlpathspliturlpathsplit('http://lemonde.fr/section/article.html')>>> ['section','article.html']urlpathsplit('http://lemonde.fr/')>>> []# If you want to split a path directlyfromuralimportpathsplitpathsplit('/section/articles/')>>> ['section','articles']
Function returning an iterator over the urls present in the links of given HTML text.
Note this function is able to work both on string and bytes seamlessly.
fromuralimporturls_from_htmlhtml="""<p>Hey! Check this site: <a href="https://medialab.sciencespo.fr/">médialab</a></p>"""forurlinurls_from_html(html):print(url)>>>'https://medialab.sciencespo.fr/'
Arguments
- stringstring|bytes: html string or bytes.
- encoding?string [
utf-8]: if given binary, this encoding will be used to decode the found urls. - errors?string [
strict]: policy on decode errors.
Function returning an iterator over the urls present in the string argument. Extracts only urls having a protocol.
Note that this function is somewhat markdown-aware, and punctuation-aware.
fromuralimporturls_from_texttext="Hey! Check this site: https://medialab.sciencespo.fr/, it looks really cool. They're developing many tools on https://github.com/"forurlinurls_from_text(text):print(url)>>>'https://medialab.sciencespo.fr/'>>>'https://github.com/'
Arguments
- stringstring: source string.
If you want to upgrade the package's data wrt Mozilla suffixes and IANA TLDs, you can do so either by running the following command:
python -m ural upgrade
or directly in your python code:
fromural.tldimportupgradeupgrade()# Or if you want to patch runtime only this time, or regularly# (for long running programs or to avoid rights issues etc.):upgrade(transient=True)
Class implementing a hierarchic set of hostnames so you can efficiently query whether urls match hostnames in the set.
fromuralimportHostnameTrieSettrie=HostnameTrieSet()trie.add('lemonde.fr')trie.add('business.lefigaro.fr')trie.match('https://liberation.fr/article1.html')>>>Falsetrie.match('https://lemonde.fr/article1.html')>>>Truetrie.match('https://www.lemonde.fr/article1.html')>>>Truetrie.match('https://lefigaro.fr/article1.html')>>>Falsetrie.match('https://business.lefigaro.fr/article1.html')>>>True
Method add a single hostname to the set.
fromuralimportHostnameTrieSettrie=HostnameTrieSet()trie.add('lemonde.fr')
Arguments
- hostnamestring: hostname to add to the set.
Method returning whether the given url matches any of the set's hostnames.
fromuralimportHostnameTrieSettrie=HostnameTrieSet()trie.add('lemonde.fr')trie.match('https://liberation.fr/article1.html')>>>Falsetrie.match('https://lemonde.fr/article1.html')>>>True
Arguments
- urlstring|urllib.parse.SplitResult: url to match.
Function converting the given url to a serialized lru.
fromural.lruimporturl_to_lruurl_to_lru('http://www.lemonde.fr:8000/article/1234/index.html?field=value#2')>>>'s:http|t:8000|h:fr|h:lemonde|h:www|p:article|p:1234|p:index.html|q:field=value|f:2|'
Arguments
- urlstring: url to convert.
- suffix_aware?bool: whether to be mindful of suffixes when converting (e.g. considering "co.uk" as a single token).
Function converting the given serialized lru or lru stems to a proper url.
fromural.lruimportlru_to_urllru_to_url('s:http|t:8000|h:fr|h:lemonde|h:www|p:article|p:1234|p:index.html|')>>>'http://www.lemonde.fr:8000/article/1234/index.html'lru_to_url(['s:http','h:fr','h:lemonde','h:www','p:article','p:1234','p:index.html'])>>>'http://www.lemonde.fr:8000/article/1234/index.html'
Function returning url parts in hierarchical order.
fromural.lruimportlru_stemslru_stems('http://www.lemonde.fr:8000/article/1234/index.html?field=value#2')>>> ['s:http','t:8000','h:fr','h:lemonde','h:www','p:article','p:1234','p:index.html','q:field=value','f:2']
Arguments
- urlstring: URL to parse.
- suffix_aware?bool: whether to be mindful of suffixes when converting (e.g. considering "co.uk" as a single token).
Function canonicalizing the url and returning its parts in hierarchical order.
fromural.lruimportcanonicalized_lru_stemscanonicalized_lru_stems('http://www.lemonde.fr/article/1234/index.html?field=value#2')>>> ['s:http','h:fr','h:lemonde','p:article','p:1234','q:field=value','f:2']
Arguments
This function accepts the same arguments ascanonicalize_url.
Function normalizing the url and returning its parts in hierarchical order.
fromural.lruimportnormalized_lru_stemsnormalized_lru_stems('http://www.lemonde.fr/article/1234/index.html?field=value#2')>>> ['h:fr','h:lemonde','p:article','p:1234','q:field=value']
Arguments
This function accepts the same arguments asnormalize_url.
Function fingerprinting the url and returning its parts in hierarchical order.
fromural.lruimportfingerprinted_lru_stemsfingerprinted_lru_stems('http://www.lemonde.fr/article/1234/index.html?field=value#2',strip_suffix=True)>>> ['h:lemonde','p:article','p:1234','q:field=value']
Arguments
This function accepts the same arguments asfingerprint_url.
Function serializing lru stems to a string.
fromural.lruimportserialize_lruserialize_lru(['s:https','h:fr','h:lemonde'])>>>'s:https|h:fr|h:lemonde|'
Function unserializing stringified lru to a list of stems.
fromural.lruimportunserialize_lruunserialize_lru('s:https|h:fr|h:lemonde|')>>> ['s:https','h:fr','h:lemonde']
Class implementing a prefix tree (Trie) storing URLs hierarchically by storing them as LRUs along with some arbitrary metadata. It is very useful when needing to match URLs by longest common prefix.
Note that this class directly inherits from thephylactery library'sTrieDict so you can also use any of its methods.
fromural.lruimportLRUTrietrie=LRUTrie()# To respect suffixestrie=LRUTrie(suffix_aware=True)
Method storing a URL in a LRUTrie along with its metadata.
fromural.lruimportLRUTrietrie=LRUTrie()trie.set('http://www.lemonde.fr', {'type':'general press'})trie.match('http://www.lemonde.fr')>>> {'type':'general press'}
Arguments
- urlstring: url to store in the LRUTrie.
- metadataany: metadata of the url.
Method storing a URL already represented as a LRU or LRU stems along with its metadata.
fromural.lruimportLRUTrietrie=LRUTrie()# Using stemstrie.set_lru(['s:http','h:fr','h:lemonde','h:www'], {'type':'general press'})# Using serialized lrutrie.set_lru('s:http|h:fr|h:lemonde|h:www|', {'type':'general_press'})
Arguments
- lrustring|list: lru to store in the Trie.
- metadataany: metadata to attach to the lru.
Method returning the metadata attached to the longest prefix match of your query URL. Will returnNone if no common prefix can be found.
fromural.lruimportLRUTrietrie=LRUTrie()trie.set('http://www.lemonde.fr', {'media':'lemonde'})trie.match('http://www.lemonde.fr')>>> {'media':'lemonde'}trie.match('http://www.lemonde.fr/politique')>>> {'media':'lemonde'}trie.match('http://www.lefigaro.fr')>>>None
Arguments
- urlstring: url to match in the LRUTrie.
Method returning the metadata attached to the longest prefix match of your query LRU. Will returnNone if no common prefix can be found.
fromural.lruimportLRUTrietrie=LRUTrie()trie.set(['s:http','h:fr','h:lemonde','h:www'], {'media':'lemonde'})trie.match(['s:http','h:fr','h:lemonde','h:www'])>>> {'media':'lemonde'}trie.match('s:http|h:fr|h:lemonde|h:www|p:politique|')>>> {'media':'lemonde'}trie.match(['s:http','h:fr','h:lefigaro','h:www'])>>>None
Arguments
- lrustring|list: lru to match in the LRUTrie.
TheCanonicalizedLRUTrie is nearly identical to the standardLRUTrie except that it canonicalizes given urls before attempting any operation using thecanonicalize_url function.
Its constructor therefore takes the same arguments as the beforementioned function.
fromural.lruimportCanonicalizedLRUTrietrie=CanonicalizedLRUTrie(strip_fragment=False)
TheNormalizedLRUTrie is nearly identical to the standardLRUTrie except that it normalizes given urls before attempting any operation using thenormalize_url function.
Its constructor therefore takes the same arguments as the beforementioned function.
fromural.lruimportNormalizedLRUTrietrie=NormalizedLRUTrie(normalize_amp=False)
TheFingerprintedLRUTrie is nearly identical to the standardLRUTrie except that it fingerprints given urls before attempting any operation using thefingerprint_url function.
Its constructor therefore takes the same arguments as the beforementioned function.
fromural.lruimportFingerprintedLRUTrietrie=FingerprintedLRUTrie(strip_suffix=False)
Function returning whether the given url is pointing to a Facebook resource potentially having comments (such as a post, photo or video for instance).
fromural.facebookimporthas_facebook_commentshas_facebook_comments('https://www.facebook.com/permalink.php?story_fbid=1354978971282622&id=598338556946671')>>>Truehas_facebook_comments('https://www.facebook.com/108824017345866/videos/311658803718223')>>>Truehas_facebook_comments('https://www.facebook.com/astucerie/')>>>Falsehas_facebook_comments('https://www.lemonde.fr')>>>Falsehas_facebook_comments('/permalink.php?story_fbid=1354978971282622&id=598338556946671',allow_relative_urls=True)>>>True
Function returning whether the given string is a valid Facebook id or not.
fromural.facebookimportis_facebook_idis_facebook_id('974583586343')>>>Trueis_facebook_id('whatever')>>>False
Function returning whether the given string is a valid Facebook full post id or not.
fromural.facebookimportis_facebook_full_idis_facebook_full_id('974583586343_9749757953')>>>Trueis_facebook_full_id('974583586343')>>>Falseis_facebook_full_id('whatever')>>>False
Function returning whether given url is from Facebook or not.
fromural.facebookimportis_facebook_urlis_facebook_url('http://www.facebook.com/post/974583586343')>>>Trueis_facebook_url('https://fb.me/846748464')>>>Trueis_facebook_url('https://www.lemonde.fr')>>>False
Function returning whether the given url is a Facebook post or not.
fromural.facebookimportis_facebook_post_urlis_facebook_post_url('http://www.facebook.com/post/974583586343')>>>Trueis_facebook_post_url('http://www.facebook.com')>>>Falseis_facebook_post_url('https://www.lemonde.fr')>>>False
Function returning whether the given url is a Facebook redirection link.
fromural.facebookimportis_facebook_linkis_facebook_link('https://l.facebook.com/l.php?u=http%3A%2F%2Fwww.chaos-controle.com%2Farchives%2F2013%2F10%2F14%2F28176300.html&h=AT0iUqJpUTMzHAH8HAXwZ11p8P3Z-SrY90wIXZhcjMnxBTHMiau8Fv1hvz00ZezRegqmF86SczyUXx3Gzdt_MdFH-I4CwHIXKKU9L6w522xwOqkOvLAylxojGEwrp341uC-GlVyGE2N7XwTPK9cpP0mQ8PIrWh8Qj2gHIIR08Js0mUr7G8Qe9fx66uYcfnNfTTF1xi0Us8gTo4fOZxAgidGWXsdgtU_OdvQqyEm97oHzKbWfXjkhsrzbtb8ZNMDwCP5099IMcKRD8Hi5H7W3vwh9hd_JlRgm5Z074epD_mGAeoEATE_QUVNTxO0SHO4XNn2Z7LgBamvevu1ENBcuyuSOYA0BsY2cx8mPWJ9t44tQcnmyQhBlYm_YmszDaQx9IfVP26PRqhsTLz-kZzv0DGMiJFU78LVWVPc9QSw2f9mA5JYWr29w12xJJ5XGQ6DhJxDMWRnLdG8Tnd7gZKCaRdqDER1jkO72u75-o4YuV3CLh4j-_4u0fnHSzHdVD8mxr9pNEgu8rvJF1E2H3-XbzA6F2wMQtFCejH8MBakzYtTGNvHSexSiKphE04Ci1Z23nBjCZFsgNXwL3wbIXWfHjh2LCKyihQauYsnvxp6fyioStJSGgyA9GGEswizHa20lucQF0S0F8H9-')>>>Trueis_facebook_link('https://lemonde.fr')>>>False
Function returning the mobile version of the given Facebook url. Will raise an exception if a non-Facebook url is given.
fromural.facebookimportconvert_facebook_url_to_mobileconvert_facebook_url_to_mobile('http://www.facebook.com/post/974583586343')>>>'http://m.facebook.com/post/974583586343'
Function parsing the given Facebook url.
fromural.facebookimportparse_facebook_url# Importing related classes if you need to perform testsfromural.facebookimport (FacebookHandle,FacebookUser,FacebookGroup,FacebookPost,FacebookPhoto,FacebookVideo)parse_facebook_url('https://www.facebook.com/people/Sophia-Aman/102016783928989')>>>FacebookUser(id='102016783928989')parse_facebook_url('https://www.facebook.com/groups/159674260452951')>>>FacebookGroup(id='159674260452951')parse_facebook_url('https://www.facebook.com/groups/159674260852951/permalink/1786992671454427/')>>>FacebookPost(id='1786992671454427',group_id='159674260852951')parse_facebook_url('https://www.facebook.com/108824017345866/videos/311658803718223')>>>FacebookVideo(id='311658803718223',parent_id='108824017345866')parse_facebook_url('https://www.facebook.com/photo.php?fbid=10222721681573727')>>>FacebookPhoto(id='10222721681573727')parse_facebook_url('/annelaure.rivolu?rc=p&__tn__=R',allow_relative_urls=True)>>>FacebookHandle(handle='annelaure.rivolu')parse_facebook_url('https://lemonde.fr')>>>None
Function extracting target url from a Facebook redirection link.
fromural.facebookimportextract_url_from_facebook_linkextract_url_from_facebook_link('https://l.facebook.com/l.php?u=http%3A%2F%2Fwww.chaos-controle.com%2Farchives%2F2013%2F10%2F14%2F28176300.html&h=AT0iUqJpUTMzHAH8HAXwZ11p8P3Z-SrY90wIXZhcjMnxBTHMiau8Fv1hvz00ZezRegqmF86SczyUXx3Gzdt_MdFH-I4CwHIXKKU9L6w522xwOqkOvLAylxojGEwrp341uC-GlVyGE2N7XwTPK9cpP0mQ8PIrWh8Qj2gHIIR08Js0mUr7G8Qe9fx66uYcfnNfTTF1xi0Us8gTo4fOZxAgidGWXsdgtU_OdvQqyEm97oHzKbWfXjkhsrzbtb8ZNMDwCP5099IMcKRD8Hi5H7W3vwh9hd_JlRgm5Z074epD_mGAeoEATE_QUVNTxO0SHO4XNn2Z7LgBamvevu1ENBcuyuSOYA0BsY2cx8mPWJ9t44tQcnmyQhBlYm_YmszDaQx9IfVP26PRqhsTLz-kZzv0DGMiJFU78LVWVPc9QSw2f9mA5JYWr29w12xJJ5XGQ6DhJxDMWRnLdG8Tnd7gZKCaRdqDER1jkO72u75-o4YuV3CLh4j-_4u0fnHSzHdVD8mxr9pNEgu8rvJF1E2H3-XbzA6F2wMQtFCejH8MBakzYtTGNvHSexSiKphE04Ci1Z23nBjCZFsgNXwL3wbIXWfHjh2LCKyihQauYsnvxp6fyioStJSGgyA9GGEswizHa20lucQF0S0F8H9-')>>>'http://www.chaos-controle.com/archives/2013/10/14/28176300.html'extract_url_from_facebook_link('http://lemonde.fr')>>>None
Returns whether the given url is probably a Google AMP url.
fromural.googleimportis_amp_urlis_amp_url('http://www.europe1.fr/sante/les-onze-vaccins.amp')>>>Trueis_amp_url('https://www.lemonde.fr')>>>False
Returns whether the given url is a Google search link.
fromural.googleimportis_google_linkis_google_link('https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&cad=rja&uact=8&ved=2ahUKEwjp8Lih_LnmAhWQlxQKHVTmCJYQFjADegQIARAB&url=http%3A%2F%2Fwww.mon-ip.com%2F&usg=AOvVaw0sfeZJyVtUS2smoyMlJmes')>>>Trueis_google_link('https://www.lemonde.fr')>>>False
Extracts the url from the given Google search link. This is useful to "resolve" the links scraped from Google's search results. ReturnsNone if given url is not valid nor relevant.
fromural.googleimportextract_url_from_google_linkextract_url_from_google_link('https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=2ahUKEwicu4K-rZzmAhWOEBQKHRNWA08QFjAAegQIARAB&url=https%3A%2F%2Fwww.facebook.com%2Fieff.ogbeide&usg=AOvVaw0vrBVCiIHUr5pncjeLpPUp')>>>'https://www.facebook.com/ieff.ogbeide'extract_url_from_google_link('https://www.lemonde.fr')>>>None
Extracts a file id from the given Google drive url. ReturnsNone if given url is not valid nor relevant.
fromural.googleimportextract_id_from_google_drive_urlextract_id_from_google_drive_url('https://docs.google.com/spreadsheets/d/1Q9sJtAb1BZhUMjxCLMrVASx3AoNDp5iV3VkbPjlg/edit#gid=0')>>>'1Q9sJtAb1BZhUMjxCLMrVASx3AoNDp5iV3VkbPjlg'extract_id_from_google_drive_url('https://www.lemonde.fr')>>>None
Parse the given Google drive url. ReturnsNone if given is not valid nor relevant.
fromural.googleimport (parse_google_drive_url,GoogleDriveFile,GoogleDrivePublicLink)parse_google_drive_url('https://docs.google.com/spreadsheets/d/1Q9sJtAb1BZhUMjxCLMrVASx3AoNDp5iV3VkbPjlg/edit#gid=0')>>>GoogleDriveFile('spreadsheets','1Q9sJtAb1BZhUMjxCLMrVASx3AoNDp5iV3VkbPjlg')parse_google_drive_url('https://www.lemonde.fr')>>>None
Function returning whether the given string is a valid Instagram post shortcode or not.
fromural.instagramimportis_instagram_post_shortcodeis_instagram_post_shortcode('974583By-5_86343')>>>Trueis_instagram_post_shortcode('whatever!!')>>>False
Function returning whether the given string is a valid Instagram username or not.
fromural.instagramimportis_instagram_usernameis_instagram_username('97458.3By-5_86343')>>>Trueis_instagram_username('whatever!!')>>>False
Returns whether the given url is from Instagram.
fromural.instagramimportis_instagram_urlis_instagram_url('https://lemonde.fr')>>>Falseis_instagram_url('https://www.instagram.com/guillaumelatorre')>>>True
Return a username from the given Instagram url orNone if we could not find one.
fromural.instagramimportextract_username_from_instagram_urlextract_username_from_instagram_url('https://www.instagram.com/martin_dupont/p/BxKRx5CHn5i/')>>>'martin_dupont'extract_username_from_instagram_url('https://lemonde.fr')>>>None
Returns parsed information about the given Instagram url: either about the post, the user or the reel. If the url is an invalid Instagram url or if not an Instagram url, the function returnsNone.
fromural.instagramimport (parse_instagram_url,# You can also import the named tuples if you need themInstagramPost,InstagramUser,InstagramReel)parse_instagram_url('https://www.instagram.com/martin_dupont/p/BxKRx5CHn5i/')>>>InstagramPost(id='BxKRx5CHn5i',name='martin_dupont')parse_instagram_url('https://lemonde.fr')>>>Noneparse_instagram_url('https://www.instagram.com/p/BxKRx5-Hn5i/')>>>InstagramPost(id='BxKRx5-Hn5i',name=None)parse_instagram_url('https://www.instagram.com/martin_dupont')>>>InstagramUser(name='martin_dupont')parse_instagram_url('https://www.instagram.com/reels/BxKRx5-Hn5i')>>>InstagramReel(id='BxKRx5-Hn5i')
Arguments
- urlstr: Instagram url to parse.
Function returning whether the given string is a valid Telegram message id or not.
fromural.telegramimportis_telegram_message_idis_telegram_message_id('974583586343')>>>Trueis_telegram_message_id('whatever')>>>False
Returns whether the given url is from Telegram.
fromural.telegramimportis_telegram_urlis_telegram_url('https://lemonde.fr')>>>Falseis_telegram_url('https://telegram.me/guillaumelatorre')>>>Trueis_telegram_url('https://t.me/s/jesstern')>>>True
Function returning the public version of the given Telegram url. Will raise an exception if a non-Telegram url is given.
fromural.teglegramimportconvert_telegram_url_to_publicconvert_telegram_url_to_public('https://t.me/jesstern')>>>'https://t.me/s/jesstern'
Return a channel from the given Telegram url orNone if we could not find one.
fromural.telegramimportextract_channel_name_from_telegram_urlextract_channel_name_from_telegram_url('https://t.me/s/jesstern/345')>>>'jesstern'extract_channel_name_from_telegram_url('https://lemonde.fr')>>>None
Returns parsed information about the given telegram url: either about the channel, message or user. If the url is an invalid Telegram url or if not a Telegram url, the function returnsNone.
fromural.telegramimport (parse_telegram_url,# You can also import the named tuples if you need themTelegramMessage,TelegramChannel,TelegramGroup)parse_telegram_url('https://t.me/s/jesstern/76')>>>TelegramMessage(name='jesstern',id='76')parse_telegram_url('https://lemonde.fr')>>>Noneparse_telegram_url('https://telegram.me/rapsocialclub')>>>TelegramChannel(name='rapsocialclub')parse_telegram_url('https://t.me/joinchat/AAAAAE9B8u_wO9d4NiJp3w')>>>TelegramGroup(id='AAAAAE9B8u_wO9d4NiJp3w')
Arguments
- urlstr: Telegram url to parse.
Returns whether the given url is from Twitter.
fromural.twitterimportis_twitter_urlis_twitter_url('https://lemonde.fr')>>>Falseis_twitter_url('https://www.twitter.com/Yomguithereal')>>>Trueis_twitter_url('https://twitter.com')>>>True
Extracts a normalized user's screen name from a Twitter url. If given an irrelevant url, the function will returnNone.
fromural.twitterimportextract_screen_name_from_twitter_urlextract_screen_name_from_twitter_url('https://www.twitter.com/Yomguithereal')>>>'yomguithereal'extract_screen_name_from_twitter_url('https://twitter.fr')>>>None
Takes a Twitter url and returns either aTwitterUser namedtuple (contains a screen_name) if the given url is a link to a twitter user, aTwitterTweet namedtuple (contains a user_screen_name and an id) if the given url is a tweet's url, aTwitterList namedtuple (contains an id) orNone if the given url is irrelevant.
fromural.twitterimportparse_twitter_urlparse_twitter_url('https://twitter.com/Yomguithereal')>>>TwitterUser(screen_name='yomguithereal')parse_twitter_url('https://twitter.com/medialab_ScPo/status/1284154793376784385')>>>TwitterTweet(user_screen_name='medialab_scpo',id='1284154793376784385')parse_twitter_url('https://twitter.com/i/lists/15512656222798157826')>>>TwitterList(id='15512656222798157826')parse_twitter_url('https://twitter.com/home')>>>None
Returns whether the given url is from Youtube.
fromural.youtubeimportis_youtube_urlis_youtube_url('https://lemonde.fr')>>>Falseis_youtube_url('https://www.youtube.com/watch?v=otRTOE9i51o')>>>Trueis_youtube_url('https://youtu.be/otRTOE9i51o)>>>True
Returns whether the given string is a formally valid Youtube channel id. Note that it won't validate the fact that this id actually refers to an existing channel or not. You will need to call YouTube servers for that.
fromural.youtubeimportis_youtube_channel_idis_youtube_channel_id('UCCCPCZNChQdGa9EkATeye4g')>>>Trueis_youtube_channel_id('@France24')>>>False
Returns whether the given string is a formally valid YouTube video id. Note that it won't validate the fact that this id actually refers to an existing video or not. You will need to call YouTube servers for that.
fromural.youtubeimportis_youtube_video_idis_youtube_video_id('otRTOE9i51o')>>>Trueis_youtube_video_id('bDYTYET')>>>False
Returns parsed information about the given youtube url: either about the linked video, user or channel. If the url is an invalid Youtube url or if not a Youtube url, the function returnsNone.
fromural.youtubeimport (parse_youtube_url,# You can also import the named tuples if you need themYoutubeVideo,YoutubeUser,YoutubeChannel,YoutubeShort,)parse_youtube_url('https://www.youtube.com/watch?v=otRTOE9i51o')>>>YoutubeVideo(id='otRTOE9i51o')parse_youtube_url('https://www.youtube.com/shorts/GINlKobb41w')>>>YoutubeShort(id='GINlKobb41w')parse_youtube_url('https://lemonde.fr')>>>Noneparse_youtube_url('http://www.youtube.com/channel/UCWvUxN9LAjJ-sTc5JJ3gEyA/videos')>>>YoutubeChannel(id='UCWvUxN9LAjJ-sTc5JJ3gEyA',name=None)parse_youtube_url('http://www.youtube.com/user/ojimfrance')>>>YoutubeUser(id=None,name='ojimfrance')parse_youtube_url('https://www.youtube.com/taranisnews')>>>YoutubeChannel(id=None,name='taranisnews')
Arguments
- urlstr: Youtube url to parse.
- fix_common_mistakesbool [
True]: Whether to fix common mistakes that can be found in Youtube urls as you can find them when crawling the web.
Return a video id from the given Youtube url orNone if we could not find one. Note that this will also work with Youtube shorts.
fromural.youtubeimportextract_video_id_from_youtube_urlextract_video_id_from_youtube_url('https://www.youtube.com/watch?v=otRTOE9i51o')>>>'otRTOE9i51o'extract_video_id_from_youtube_url('https://lemonde.fr')>>>Noneextract_video_id_from_youtube_url('http://youtu.be/afa-5HQHiAs')>>>'afa-5HQHiAs'
Returns a normalized version of the given Youtube url. It will normalize video, user and channel urls so you can easily match them.
fromural.youtubeimportnormalize_youtube_urlnormalize_youtube_url('https://www.youtube.com/watch?v=otRTOE9i51o')>>>'https://www.youtube.com/watch?v=otRTOE9i51o'normalize_youtube_url('http://youtu.be/afa-5HQHiAs')>>>'https://www.youtube.com/watch?v=afa-5HQHiAs'
TL;DR: a LRU is a hierarchical reordering of a URL so that one can perform meaningful prefix queries on URLs.
If you observe many URLs, you will quickly notice that they are not written in sound hierarchical order. In this URL, for instance:
http://business.lemonde.fr/articles/money.html?id=34#contentSome parts, such as the subdomain, are written in an "incorrect order". And this is fine, really, this is how URLs always worked.
But if what you really want is to match URLs, you will need to reorder them so that their order closely reflects the hierarchy of their targeted content. And this is exactly what LRUs are (that and also a bad pun on URL, since a LRU is basically a "reversed" URL).
Now look how the beforementioned URL could be splitted into LRU stems:
['s:http','h:fr','h:lemonde','h:business','p:articles','p:money.html','q:id=34','f:content']
And typically, this list of stems will be serialized thusly:
s:http|h:fr|h:lemonde|h:business|p:articles|p:money.html|q:id=34|f:content|The trailing slash is added so that serialized LRUs can beprefix-free.
About
A helper library full of URL-related heuristics.