extruct is a library for extracting embedded metadata from HTML markup.
Currently, extruct supports:
- W3C's HTML Microdata
- embedded JSON-LD
- Microformat via mf2py
- Facebook's Open Graph
- (experimental) RDFa via rdflib
The microdata algorithm is a revisit of this Scrapinghub blog post showing how to use EXSLT extensions.
pip install extruct
The simplest example how to use extruct is to call
extruct.extract(htmlstring, base_url=base_url)
with some HTML string and an optional base URL.
Let's try this on a webpage that uses all the syntaxes supported (RDFa with `ogp`_).
First fetch the HTML using python-requests and then feed the response body to extruct:
>>> import extruct
>>> import requests
>>> import pprint
>>> from w3lib.html import get_base_url
>>>
>>> pp = pprint.PrettyPrinter(indent=2)
>>> r = requests.get('https://www.optimizesmart.com/how-to-use-open-graph-protocol/')
>>> base_url = get_base_url(r.text, r.url)
>>> data = extruct.extract(r.text, base_url=base_url)
>>>
>>> pp.pprint(data)
{ 'json-ld': [ { '@context': 'https://schema.org',
'@id': '#organization',
'@type': 'Organization',
'logo': 'https://www.optimizesmart.com/wp-content/uploads/2016/03/optimize-smart-Twitter-logo.jpg',
'name': 'Optimize Smart',
'sameAs': [ 'https://www.facebook.com/optimizesmart/',
'https://uk.linkedin.com/in/analyticsnerd',
'https://www.youtube.com/user/optimizesmart',
'https://twitter.com/analyticsnerd'],
'url': 'https://www.optimizesmart.com/'}],
'microdata': [ { 'properties': {'headline': ''},
'type': 'http://schema.org/WPHeader'}],
'microformat': [ { 'children': [ { 'properties': { 'category': [ 'specialized-tracking'],
'name': [ 'Open Graph '
'Protocol for '
'Facebook '
'explained with '
'examples\n'
'\n'
'Specialized '
'Tracking\n'
'\n'
'\n'
(...)
'Follow '
'@analyticsnerd\n'
'!function(d,s,id){var '
"js,fjs=d.getElementsByTagName(s)[0],p=/^http:/.test(d.location)?'http':'https';if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src=p+'://platform.twitter.com/widgets.js';fjs.parentNode.insertBefore(js,fjs);}}(document, "
"'script', "
"'twitter-wjs');"]},
'type': ['h-entry']}],
'properties': { 'name': [ 'Open Graph Protocol for '
'Facebook explained with '
'examples\n'
(...)
'Follow @analyticsnerd\n'
'!function(d,s,id){var '
"js,fjs=d.getElementsByTagName(s)[0],p=/^http:/.test(d.location)?'http':'https';if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src=p+'://platform.twitter.com/widgets.js';fjs.parentNode.insertBefore(js,fjs);}}(document, "
"'script', 'twitter-wjs');"]},
'type': ['h-feed']}],
'opengraph': [ { 'namespace': {'og': 'http://ogp.me/ns#'},
'properties': [ ('og:locale', 'en_US'),
('og:type', 'article'),
( 'og:title',
'Open Graph Protocol for Facebook '
'explained with examples'),
( 'og:description',
'What is Open Graph Protocol and why you '
'need it? Learn to implement Open Graph '
'Protocol for Facebook on your website. '
'Open Graph Protocol Meta Tags.'),
( 'og:url',
'https://www.optimizesmart.com/how-to-use-open-graph-protocol/'),
('og:site_name', 'Optimize Smart'),
( 'og:updated_time',
'2018-03-09T16:26:35+00:00'),
( 'og:image',
'https://www.optimizesmart.com/wp-content/uploads/2010/07/open-graph-protocol.jpg'),
( 'og:image:secure_url',
'https://www.optimizesmart.com/wp-content/uploads/2010/07/open-graph-protocol.jpg')]}],
'rdfa': [ { '@id': 'https://www.optimizesmart.com/how-to-use-open-graph-protocol/#header',
'http://www.w3.org/1999/xhtml/vocab#role': [ { '@id': 'http://www.w3.org/1999/xhtml/vocab#banner'}]},
{ '@id': 'https://www.optimizesmart.com/how-to-use-open-graph-protocol/',
'article:modified_time': [ { '@value': '2018-03-09T16:26:35+00:00'}],
'article:published_time': [ { '@value': '2010-07-02T18:57:23+00:00'}],
'article:publisher': [ { '@value': 'https://www.facebook.com/optimizesmart/'}],
'article:section': [{'@value': 'Specialized Tracking'}],
'http://ogp.me/ns#description': [ { '@value': 'What is Open '
'Graph Protocol '
'and why you need '
'it? Learn to '
'implement Open '
'Graph Protocol '
'for Facebook on '
'your website. '
'Open Graph '
'Protocol Meta '
'Tags.'}],
'http://ogp.me/ns#image': [ { '@value': 'https://www.optimizesmart.com/wp-content/uploads/2010/07/open-graph-protocol.jpg'}],
'http://ogp.me/ns#image:secure_url': [ { '@value': 'https://www.optimizesmart.com/wp-content/uploads/2010/07/open-graph-protocol.jpg'}],
'http://ogp.me/ns#locale': [{'@value': 'en_US'}],
'http://ogp.me/ns#site_name': [{'@value': 'Optimize Smart'}],
'http://ogp.me/ns#title': [ { '@value': 'Open Graph Protocol for '
'Facebook explained with '
'examples'}],
'http://ogp.me/ns#type': [{'@value': 'article'}],
'http://ogp.me/ns#updated_time': [ { '@value': '2018-03-09T16:26:35+00:00'}],
'http://ogp.me/ns#url': [ { '@value': 'https://www.optimizesmart.com/how-to-use-open-graph-protocol/'}],
'https://api.w.org/': [ { '@id': 'https://www.optimizesmart.com/wp-json/'}]}]}
It is possible to select which syntaxes to extract by passing a list with the desired ones to extract. Valid values: 'microdata', 'json-ld', 'opengraph', 'microformat', 'rdfa'. If no list is passed all syntaxes will be extracted and returned:
>>> r = requests.get('http://www.songkick.com/artists/236156-elysian-fields')
>>> base_url = get_base_url(r.text, r.url)
>>> data = extruct.extract(r.text, base_url, syntaxes=['microdata', 'opengraph', 'rdfa'])
>>>
>>> pp.pprint(data)
{ 'microdata': [],
'opengraph': [ { 'namespace': { 'concerts': 'http://ogp.me/ns/fb/songkick-concerts#',
'fb': 'http://www.facebook.com/2008/fbml',
'og': 'http://ogp.me/ns#'},
'properties': [ ('fb:app_id', '308540029359'),
('og:site_name', 'Songkick'),
('og:type', 'songkick-concerts:artist'),
('og:title', 'Elysian Fields'),
( 'og:description',
'Find out when Elysian Fields is next '
'playing live near you. List of all '
'Elysian Fields tour dates and concerts.'),
( 'og:url',
'https://www.songkick.com/artists/236156-elysian-fields'),
( 'og:image',
'http://images.sk-static.com/images/media/img/col4/20100330-103600-169450.jpg')]}],
'rdfa': [ { '@id': 'https://www.songkick.com/artists/236156-elysian-fields',
'al:ios:app_name': [{'@value': 'Songkick Concerts'}],
'al:ios:app_store_id': [{'@value': '438690886'}],
'al:ios:url': [ { '@value': 'songkick://artists/236156-elysian-fields'}],
'http://ogp.me/ns#description': [ { '@value': 'Find out when '
'Elysian Fields is '
'next playing live '
'near you. List of '
'all Elysian '
'Fields tour dates '
'and concerts.'}],
'http://ogp.me/ns#image': [ { '@value': 'http://images.sk-static.com/images/media/img/col4/20100330-103600-169450.jpg'}],
'http://ogp.me/ns#site_name': [{'@value': 'Songkick'}],
'http://ogp.me/ns#title': [{'@value': 'Elysian Fields'}],
'http://ogp.me/ns#type': [{'@value': 'songkick-concerts:artist'}],
'http://ogp.me/ns#url': [ { '@value': 'https://www.songkick.com/artists/236156-elysian-fields'}],
'http://www.facebook.com/2008/fbmlapp_id': [ { '@value': '308540029359'}]}]}
Another option is to uniform the output of microformat, opengraph, microdata and json-ld syntaxes to the following structure:
{'@context': 'http://example.com',
'@type': 'example_type',
/* All other the properties in keys here */
}
To do so set uniform=True when calling extract, it's false by default for backward compatibility. Here the same example as before but with uniform set to True:
>>> r = requests.get('http://www.songkick.com/artists/236156-elysian-fields')
>>> base_url = get_base_url(r.text, r.url)
>>> data = extruct.extract(r.text, base_url, syntaxes=['microdata', 'opengraph', 'rdfa'], uniform=True)
>>>
>>> pp.pprint(data)
{ 'microdata': [],
'opengraph': [ { '@context': { 'concerts': 'http://ogp.me/ns/fb/songkick-concerts#',
'fb': 'http://www.facebook.com/2008/fbml',
'og': 'http://ogp.me/ns#'},
'@type': 'songkick-concerts:artist',
'fb:app_id': '308540029359',
'og:description': 'Find out when Elysian Fields is next '
'playing live near you. List of all '
'Elysian Fields tour dates and concerts.',
'og:image': 'http://images.sk-static.com/images/media/img/col4/20100330-103600-169450.jpg',
'og:site_name': 'Songkick',
'og:title': 'Elysian Fields',
'og:url': 'https://www.songkick.com/artists/236156-elysian-fields'}],
'rdfa': [ { '@id': 'https://www.songkick.com/artists/236156-elysian-fields',
'al:ios:app_name': [{'@value': 'Songkick Concerts'}],
'al:ios:app_store_id': [{'@value': '438690886'}],
'al:ios:url': [ { '@value': 'songkick://artists/236156-elysian-fields'}],
'http://ogp.me/ns#description': [ { '@value': 'Find out when '
'Elysian Fields is '
'next playing live '
'near you. List of '
'all Elysian '
'Fields tour dates '
'and concerts.'}],
'http://ogp.me/ns#image': [ { '@value': 'http://images.sk-static.com/images/media/img/col4/20100330-103600-169450.jpg'}],
'http://ogp.me/ns#site_name': [{'@value': 'Songkick'}],
'http://ogp.me/ns#title': [{'@value': 'Elysian Fields'}],
'http://ogp.me/ns#type': [{'@value': 'songkick-concerts:artist'}],
'http://ogp.me/ns#url': [ { '@value': 'https://www.songkick.com/artists/236156-elysian-fields'}],
'http://www.facebook.com/2008/fbmlapp_id': [ { '@value': '308540029359'}]}]}
NB rdfa structure is not uniformed yet
It is also possible to get references to HTML node for every extracted metadata item. The feature is supported only by microdata syntax.
To use that, just set the return_html_node option of extract method to True.
As the result, an additional key "nodeHtml" will be included in the result for every
item. Each node is of lxml.etree.Element type:
>>> r = requests.get('http://www.rugpadcorner.com/shop/no-muv/')
>>> base_url = get_base_url(r.text, r.url)
>>> data = extruct.extract(r.text, base_url, syntaxes=['microdata'], return_html_node=True)
>>>
>>> pp.pprint(data)
{ 'microdata': [ { 'htmlNode': <Element div at 0x7f10f8e6d3b8>,
'properties': { 'description': 'KEEP RUGS FLAT ON CARPET!\n'
'Not your thin sticky pad, '
'No-Muv is truly the best!',
'image': ['', ''],
'name': ['No-Muv', 'No-Muv'],
'offers': [ { 'htmlNode': <Element div at 0x7f10f8e6d138>,
'properties': { 'availability': 'http://schema.org/InStock',
'price': 'Price: '
'$45'},
'type': 'http://schema.org/Offer'},
{ 'htmlNode': <Element div at 0x7f10f8e60f48>,
'properties': { 'availability': 'http://schema.org/InStock',
'price': '(Select '
'Size/Shape '
'for '
'Pricing)'},
'type': 'http://schema.org/Offer'}],
'ratingValue': ['5.00', '5.00']},
'type': 'http://schema.org/Product'}]}
You can also use each extractor individually. See below.
>>> import pprint
>>> pp = pprint.PrettyPrinter(indent=2)
>>>
>>> from extruct.w3cmicrodata import MicrodataExtractor
>>>
>>> # example from http://www.w3.org/TR/microdata/#associating-names-with-items
>>> html = """<!DOCTYPE HTML>
... <html>
... <head>
... <title>Photo gallery</title>
... </head>
... <body>
... <h1>My photos</h1>
... <figure itemscope itemtype="http://n.whatwg.org/work" itemref="licenses">
... <img itemprop="work" src="images/house.jpeg" alt="A white house, boarded up, sits in a forest.">
... <figcaption itemprop="title">The house I found.</figcaption>
... </figure>
... <figure itemscope itemtype="http://n.whatwg.org/work" itemref="licenses">
... <img itemprop="work" src="images/mailbox.jpeg" alt="Outside the house is a mailbox. It has a leaflet inside.">
... <figcaption itemprop="title">The mailbox.</figcaption>
... </figure>
... <footer>
... <p id="licenses">All images licensed under the <a itemprop="license"
... href="http://www.opensource.org/licenses/mit-license.php">MIT
... license</a>.</p>
... </footer>
... </body>
... </html>"""
>>>
>>> mde = MicrodataExtractor()
>>> data = mde.extract(html)
>>> pp.pprint(data)
[{'properties': {'license': 'http://www.opensource.org/licenses/mit-license.php',
'title': 'The house I found.',
'work': 'http://www.example.com/images/house.jpeg'},
'type': 'http://n.whatwg.org/work'},
{'properties': {'license': 'http://www.opensource.org/licenses/mit-license.php',
'title': 'The mailbox.',
'work': 'http://www.example.com/images/mailbox.jpeg'},
'type': 'http://n.whatwg.org/work'}]
>>> import pprint
>>> pp = pprint.PrettyPrinter(indent=2)
>>>
>>> from extruct.jsonld import JsonLdExtractor
>>>
>>> html = """<!DOCTYPE HTML>
... <html>
... <head>
... <title>Some Person Page</title>
... </head>
... <body>
... <h1>This guys</h1>
... <script type="application/ld+json">
... {
... "@context": "http://schema.org",
... "@type": "Person",
... "name": "John Doe",
... "jobTitle": "Graduate research assistant",
... "affiliation": "University of Dreams",
... "additionalName": "Johnny",
... "url": "http://www.example.com",
... "address": {
... "@type": "PostalAddress",
... "streetAddress": "1234 Peach Drive",
... "addressLocality": "Wonderland",
... "addressRegion": "Georgia"
... }
... }
... </script>
... </body>
... </html>"""
>>>
>>> jslde = JsonLdExtractor()
>>>
>>> data = jslde.extract(html)
>>> pp.pprint(data)
[{'@context': 'http://schema.org',
'@type': 'Person',
'additionalName': 'Johnny',
'address': {'@type': 'PostalAddress',
'addressLocality': 'Wonderland',
'addressRegion': 'Georgia',
'streetAddress': '1234 Peach Drive'},
'affiliation': 'University of Dreams',
'jobTitle': 'Graduate research assistant',
'name': 'John Doe',
'url': 'http://www.example.com'}]
>>> import pprint
>>> pp = pprint.PrettyPrinter(indent=2)
>>> from extruct.rdfa import RDFaExtractor # you can ignore the warning about html5lib not being available
INFO:rdflib:RDFLib Version: 4.2.1
/home/paul/.virtualenvs/extruct.wheel.test/lib/python3.5/site-packages/rdflib/plugins/parsers/structureddata.py:30: UserWarning: html5lib not found! RDFa and Microdata parsers will not be available.
'parsers will not be available.')
>>>
>>> html = """<html>
... <head>
... ...
... </head>
... <body prefix="dc: http://purl.org/dc/terms/ schema: http://schema.org/">
... <div resource="/alice/posts/trouble_with_bob" typeof="schema:BlogPosting">
... <h2 property="dc:title">The trouble with Bob</h2>
... ...
... <h3 property="dc:creator schema:creator" resource="#me">Alice</h3>
... <div property="schema:articleBody">
... <p>The trouble with Bob is that he takes much better photos than I do:</p>
... </div>
... ...
... </div>
... </body>
... </html>
... """
>>>
>>> rdfae = RDFaExtractor()
>>> pp.pprint(rdfae.extract(html, base_url='http://www.example.com/index.html'))
[{'@id': 'http://www.example.com/alice/posts/trouble_with_bob',
'@type': ['http://schema.org/BlogPosting'],
'http://purl.org/dc/terms/creator': [{'@id': 'http://www.example.com/index.html#me'}],
'http://purl.org/dc/terms/title': [{'@value': 'The trouble with Bob'}],
'http://schema.org/articleBody': [{'@value': '\n'
' The trouble with Bob '
'is that he takes much better '
'photos than I do:\n'
' '}],
'http://schema.org/creator': [{'@id': 'http://www.example.com/index.html#me'}]}]
You'll get a list of expanded JSON-LD nodes.
>>> import pprint
>>> pp = pprint.PrettyPrinter(indent=2)
>>>
>>> from extruct.opengraph import OpenGraphExtractor
>>>
>>> html = """<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "https://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
... <html xmlns="https://www.w3.org/1999/xhtml" xmlns:og="https://ogp.me/ns#" xmlns:fb="https://www.facebook.com/2008/fbml">
... <head>
... <title>Himanshu's Open Graph Protocol</title>
... <meta http-equiv="Content-Type" content="text/html;charset=WINDOWS-1252" />
... <meta http-equiv="Content-Language" content="en-us" />
... <link rel="stylesheet" type="text/css" href="event-education.css" />
... <meta name="verify-v1" content="so4y/3aLT7/7bUUB9f6iVXN0tv8upRwaccek7JKB1gs=" >
... <meta property="og:title" content="Himanshu's Open Graph Protocol"/>
... <meta property="og:type" content="article"/>
... <meta property="og:url" content="https://www.eventeducation.com/test.php"/>
... <meta property="og:image" content="https://www.eventeducation.com/images/982336_wedding_dayandouan_th.jpg"/>
... <meta property="fb:admins" content="himanshu160"/>
... <meta property="og:site_name" content="Event Education"/>
... <meta property="og:description" content="Event Education provides free courses on event planning and management to event professionals worldwide."/>
... </head>
... <body>
... <div id="fb-root"></div>
... <script>(function(d, s, id) {
... var js, fjs = d.getElementsByTagName(s)[0];
... if (d.getElementById(id)) return;
... js = d.createElement(s); js.id = id;
... js.src = "//connect.facebook.net/en_US/all.js#xfbml=1&appId=501839739845103";
... fjs.parentNode.insertBefore(js, fjs);
... }(document, 'script', 'facebook-jssdk'));</script>
... </body>
... </html>"""
>>>
>>> opengraphe = OpenGraphExtractor()
>>> pp.pprint(opengraphe.extract(html))
[{"namespace": {
"og": "http://ogp.me/ns#"
},
"properties": [
[
"og:title",
"Himanshu's Open Graph Protocol"
],
[
"og:type",
"article"
],
[
"og:url",
"https://www.eventeducation.com/test.php"
],
[
"og:image",
"https://www.eventeducation.com/images/982336_wedding_dayandouan_th.jpg"
],
[
"og:site_name",
"Event Education"
],
[
"og:description",
"Event Education provides free courses on event planning and management to event professionals worldwide."
]
]
}]
>>> import pprint
>>> pp = pprint.PrettyPrinter(indent=2)
>>>
>>> from extruct.microformat import MicroformatExtractor
>>>
>>> html = """<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "https://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
... <html xmlns="https://www.w3.org/1999/xhtml" xmlns:og="https://ogp.me/ns#" xmlns:fb="https://www.facebook.com/2008/fbml">
... <head>
... <title>Himanshu's Open Graph Protocol</title>
... <meta http-equiv="Content-Type" content="text/html;charset=WINDOWS-1252" />
... <meta http-equiv="Content-Language" content="en-us" />
... <link rel="stylesheet" type="text/css" href="event-education.css" />
... <meta name="verify-v1" content="so4y/3aLT7/7bUUB9f6iVXN0tv8upRwaccek7JKB1gs=" >
... <meta property="og:title" content="Himanshu's Open Graph Protocol"/>
... <article class="h-entry">
... <h1 class="p-name">Microformats are amazing</h1>
... <p>Published by <a class="p-author h-card" href="http://example.com">W. Developer</a>
... on <time class="dt-published" datetime="2013-06-13 12:00:00">13<sup>th</sup> June 2013</time></p>
... <p class="p-summary">In which I extoll the virtues of using microformats.</p>
... <div class="e-content">
... <p>Blah blah blah</p>
... </div>
... </article>
... </head>
... <body></body>
... </html>"""
>>>
>>> microformate = MicroformatExtractor()
>>> data = microformate.extract(html)
>>> pp.pprint(data)
[{"type": [
"h-entry"
],
"properties": {
"name": [
"Microformats are amazing"
],
"author": [
{
"type": [
"h-card"
],
"properties": {
"name": [
"W. Developer"
],
"url": [
"http://example.com"
]
},
"value": "W. Developer"
}
],
"published": [
"2013-06-13 12:00:00"
],
"summary": [
"In which I extoll the virtues of using microformats."
],
"content": [
{
"html": "\n<p>Blah blah blah</p>\n",
"value": "\nBlah blah blah\n"
}
]
}
}]
extruct provides a command line tool that allows you to fetch a page and extract the metadata from it directly from the command line.
The command line tool depends on requests_, which is not installed by default when you install extruct. In order to use the command line tool, you can install extruct with the cli extra requirements:
pip install extruct[cli]
extruct "http://example.com"
Downloads "http://example.com" and outputs the Microdata, JSON-LD and RDFa, Open Graph and Microformat metadata to stdout.
By default, the command line tool will try to extract all the supported metadata formats from the page (currently Microdata, JSON-LD, RDFa, Open Graph and Microformat). If you want to restrict the output to just one or a subset of those, you can pass their individual names collected in a list through 'syntaxes' argument.
For example, this command extracts only Microdata and JSON-LD metadata from "http://example.com":
extruct "http://example.com" --syntaxes microdata json-ld
NB syntaxes names passed must correspond to these: microdata, json-ld, rdfa, opengraph, microformat
mkvirtualenv extruct pip install -r requirements-dev.txt
Run tests in current environment:
py.test tests
Use tox to run tests with different Python versions:
tox
Use bumpversion to conveniently change project version:
bumpversion patch # 0.0.0 -> 0.0.1 bumpversion minor # 0.0.1 -> 0.1.0 bumpversion major # 0.1.0 -> 1.0.0