def metadata_from_url(url, microdata=True, jsonld=True, rdfa=True):
resp = requests.get(url, timeout=30)
result = {
'url': url,
'status': '{} {}'.format(resp.status_code, resp.reason)
}
try:
resp.raise_for_status()
except requests.exceptions.HTTPError:
return result
parser = XmlDomHTMLParser(encoding=resp.encoding)
tree = lxml.html.fromstring(resp.content, parser=parser)
if microdata:
mde = MicrodataExtractor(nested=True)
result['microdata'] = mde.extract_items(tree, resp.url)
if jsonld:
jsonlde = JsonLdExtractor()
result['json-ld'] = jsonlde.extract_items(tree, resp.url)
if rdfa:
rdfae = RDFaExtractor()
result['rdfa'] = rdfae.extract_items(tree, resp.url)
return result
def __init__(self, response, microdata=False, jsonld=False, rdfa=False):
self.response = response
self.microdata = microdata
self.jsonld = jsonld
self.rdfa = rdfa
if rdfa:
try:
self.rdfae = RDFaExtractor()
self.rdfadata = self.rdfae.extract(self.response.text,
url=self.response.url)
except JSONDecodeError:
pass
if microdata:
try:
self.mde = MicrodataExtractor()
self.mdedata = self.mde.extract(self.response.text)
except JSONDecodeError:
pass
if jsonld:
try:
self.jlde = JsonLdExtractor()
self.jldata = self.jlde.extract(self.response.text)
except (JSONDecodeError, TypeError):
self.jldata = []
finally:
# Sometimes we get this in the meta dict from RISJExtractJSONLD
self.jldata.extend(self.response.meta.get('json-ld', []))
def test_expanded_opengraph_support(self):
body = get_testdata('misc','expanded_OG_support_test.html')
expected = json.loads(
get_testdata('misc','expanded_OG_support_test.json'
).decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, base_url='http://www.example.com/index.html')
self.assertJsonLDEqual(data,expected)
def test_wikipedia_xhtml_rdfa_no_prefix(self):
body = get_testdata('misc', 'Portfolio_Niels_Lubberman.html')
expected = json.loads(
get_testdata('misc',
'Portfolio_Niels_Lubberman.json').decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, base_url='http://nielslubberman.nl/drupal/')
self.assertJsonLDEqual(data, expected)
def test_w3c_rdf11primer(self):
for i in [14]:
fileprefix = 'w3c.rdf11primer.example{:03d}'.format(i)
body = get_testdata('w3crdfa', fileprefix + '.html').decode('UTF-8')
expected = json.loads(
get_testdata('w3crdfa', fileprefix + '.expanded.json'
).decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, url='http://www.exaple.com/index.html')
self.assertJsonLDEqual(data, expected)
def test_w3c_rdf11primer(self):
for i in [14]:
fileprefix = 'w3c.rdf11primer.example{:03d}'.format(i)
body = get_testdata('w3crdfa', fileprefix + '.html')
expected = json.loads(
get_testdata('w3crdfa',
fileprefix + '.expanded.json').decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, url='http://www.example.com/index.html')
self.assertJsonLDEqual(data, expected)
def test_wikipedia_xhtml_rdfa(self):
fileprefix = 'xhtml+rdfa'
body = get_testdata('wikipedia', fileprefix + '.html').decode('UTF-8')
expected = json.loads(
get_testdata('wikipedia', fileprefix + '.expanded.json'
).decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, url='http://www.exaple.com/index.html')
self.assertJsonLDEqual(data, expected)
def test_wikipedia_xhtml_rdfa(self):
fileprefix = 'xhtml+rdfa'
body = get_testdata('wikipedia', fileprefix + '.html').decode('UTF-8')
expected = json.loads(
get_testdata('wikipedia',
fileprefix + '.expanded.json').decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, url='http://www.exaple.com/index.html')
self.assertJsonLDEqual(data, expected)
def extract(self, html_text: str,
extract_title: bool = False,
extract_meta: bool = False,
extract_microdata: bool = False,
microdata_base_url: str = "",
extract_json_ld: bool = False,
extract_rdfa: bool = False,
rdfa_base_url: str = "") \
-> List[Extraction]:
"""
Args:
html_text (str): input html string to be extracted
extract_title (bool): True if string of 'title' tag needs to be extracted, return as { "title": "..." }
extract_meta (bool): True if string of 'meta' tags needs to be extracted, return as { "meta": { "author": "...", ...}}
extract_microdata (bool): True if microdata needs to be extracted, returns as { "microdata": [...] }
microdata_base_url (str): base namespace url for microdata, empty string if no base url is specified
extract_json_ld (bool): True if json-ld needs to be extracted, return as { "json-ld": [...] }
extract_rdfa (bool): True if rdfs needs to be extracted, returns as { "rdfa": [...] }
rdfa_base_url (str): base namespace url for rdfa, empty string if no base url is specified
Returns:
List[Extraction]: the list of extraction or the empty list if there are no matches.
"""
res = list()
soup = BeautifulSoup(html_text, 'html.parser')
if soup.title and extract_title:
title = self._wrap_data("title", soup.title.string.encode('utf-8').decode('utf-8'))
res.append(title)
if soup.title and extract_meta:
meta_content = self._wrap_meta_content(soup.find_all("meta"))
meta_data = self._wrap_data("meta", meta_content)
res.append(meta_data)
if extract_microdata:
mde = MicrodataExtractor()
mde_data = self._wrap_data("microdata", mde.extract(html_text, microdata_base_url))
res.append(mde_data)
if extract_json_ld:
jslde = JsonLdExtractor()
jslde_data = self._wrap_data("json-ld", jslde.extract(html_text))
res.append(jslde_data)
if extract_rdfa:
rdfae = RDFaExtractor()
rdfae_data = self._wrap_data("rdfa", rdfae.extract(html_text, rdfa_base_url))
res.append(rdfae_data)
return res
def test_w3c_rdfaprimer(self):
for i in [5, 6, 7, 8, 9, 10, 11, 15]:
fileprefix = 'w3c.rdfaprimer.example{:03d}'.format(i)
print(fileprefix)
body = get_testdata('w3crdfa', fileprefix + '.html').decode('UTF-8')
expected = json.loads(
get_testdata('w3crdfa', fileprefix + '.expanded.json'
).decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, url='http://www.example.com/index.html')
print("extracted:\n%s" % pformat(tupleize(data)))
print("expected:\n%s" % pformat(tupleize(expected)))
print("extracted:\n%s" % self.prettify(data))
print("expected:\n%s" % self.prettify(expected))
self.assertJsonLDEqual(data, expected)
def test_w3c_rdfaprimer(self):
for i in [5, 6, 7, 8, 9, 10, 11, 15]:
fileprefix = 'w3c.rdfaprimer.example{:03d}'.format(i)
print(fileprefix)
body = get_testdata('w3crdfa', fileprefix + '.html')
expected = json.loads(
get_testdata('w3crdfa',
fileprefix + '.expanded.json').decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, url='http://www.example.com/index.html')
print("extracted:\n%s" % pformat(tupleize(data)))
print("expected:\n%s" % pformat(tupleize(expected)))
print("extracted:\n%s" % self.prettify(data))
print("expected:\n%s" % self.prettify(expected))
self.assertJsonLDEqual(data, expected)
def extract(htmlstring, url='http://www.example.com/', encoding="UTF-8"):
domparser = XmlDomHTMLParser(encoding=encoding)
tree = fromstring(htmlstring, parser=domparser)
return {
name: extractor.extract_items(tree, url=url)
for name, extractor in (('json-ld', JsonLdExtractor()),
('microdata', MicrodataExtractor()),
('rdfa', RDFaExtractor()))
}
def test_w3c_rdfaprimer(self):
for i in [5, 6, 7, 8, 9, 10, 11, 15]:
fileprefix = 'w3c.rdfaprimer.example{:03d}'.format(i)
print(fileprefix)
body = get_testdata('w3crdfa', fileprefix + '.html')
expected = json.loads(
get_testdata('w3crdfa', fileprefix + '.expanded.json'
).decode('UTF-8'))
rdfae = RDFaExtractor()
data = rdfae.extract(body, base_url='http://www.example.com/index.html')
self.assertJsonLDEqual(data, expected)
# This is for testing that the fix to issue 116 does not affect
# severely rdfa output even in a presence of a bug in the code
def mocked_fix_order(x, y, z):
raise Exception()
rdfae._fix_order = mocked_fix_order
data = rdfae.extract(body, base_url='http://www.example.com/index.html')
self.assertJsonLDEqual(data, expected)
def extract(self, html_text: str,
extract_title: bool = False,
extract_meta: bool = False,
extract_microdata: bool = False,
extract_json_ld: bool = False,
extract_rdfa: bool = False) \
-> List[Extraction]:
res = list()
soup = BeautifulSoup(html_text, 'html.parser')
if soup.title and extract_title:
title = self.wrap_data(
"title",
soup.title.string.encode('utf-8').decode('utf-8'))
res.append(title)
if soup.title and extract_meta:
meta_content = self.wrap_meta_content(soup.find_all("meta"))
meta_data = self.wrap_data("meta", meta_content)
res.append(meta_data)
if extract_microdata:
mde = MicrodataExtractor()
mde_data = self.wrap_data("microdata", mde.extract(html_text))
res.append(mde_data)
if extract_json_ld:
jslde = JsonLdExtractor()
jslde_data = self.wrap_data("json-ld", jslde.extract(html_text))
res.append(jslde_data)
if extract_rdfa:
rdfae = RDFaExtractor()
rdfae_data = self.wrap_data("rdfa", rdfae.extract(html_text))
res.append(rdfae_data)
return res
def get_rdfa_from_warc(warc_file_no, path):
global iteration_count
global report_at_every_nth_step
rdfaFileID = 1
htmlURL = ''
data = ''
append = False
rdfaExtractor = RDFaExtractor()
if not os.path.exists('RDFa Files\\WARC_{0}'.format(warc_file_no)):
os.makedirs('RDFa Files\\WARC_{0}'.format(warc_file_no))
if not os.path.exists('XML Files\\WARC_{0}'.format(warc_file_no)):
os.makedirs('XML Files\\WARC_{0}'.format(warc_file_no))
print('[INFO/PROGRESS] The file being processed: {0}'.format(path))
with open(path, encoding='utf-8', errors='replace') as file:
for line in file:
if debug and iteration_count % report_at_every_nth_step == 0:
print("[DEBUG/PROGRESS] Processing line #{0:n}".format(
iteration_count))
if 'WARC/1.0' in line and append:
append = False
try:
rdfaData = rdfaExtractor.extract(data, base_url=htmlURL)
if rdfaData != []:
with open('RDFa Files\\WARC_{0}\\RDFa_{1}.txt'.format(
warc_file_no, rdfaFileID),
'w',
encoding='utf-8') as f:
f.write('URL: {0}\n\n'.format(htmlURL))
f.write(str(rdfaData))
f.close()
ConvertToXML.convertInstant(
str(rdfaData),
"XML Files\\WARC_{0}\\RDFa_{1}.xml".format(
warc_file_no, rdfaFileID))
if debug:
print(
"[DEBUG/PROGRESS] Processed file #{0} at URI {1} successfully"
.format(rdfaFileID, htmlURL))
except json.decoder.JSONDecodeError as jde:
print(
'[ERROR] Current file (#{0}) could not be converted to RDF/XML: {1} | This JSON-LD may be invalid.'
.format(rdfaFileID, str(jde)))
except lxml.etree.ParserError as pe:
print(
'[ERROR] Current file (#{0}) could not be converted to RDF/XML: {1} | This file may not have a valid RDFa representation.'
.format(rdfaFileID, str(pe)))
except Exception as exc:
if str(exc).startswith('Can\'t split'):
print(
'[ERROR] Current file (#{0}) could not be converted to RDF/XML: {1} | This file may be containing invalid XML namespaces.'
.format(rdfaFileID, str(exc)))
else:
print(
'[ERROR] An error has occurred while processing current file (#{0}): {1}'
.format(rdfaFileID, str(exc)))
finally:
rdfaFileID += 1
data = ''
htmlURL = ''
if 'WARC-Target-URI:' in line:
htmlURL = line.replace('WARC-Target-URI: ',
'').replace('\r', '').replace('\n', '')
if '<!DOCTYPE html' in line or '<!doctype html' in line or '<html' in line:
append = True
if append:
data = data + line + '\n'
iteration_count = iteration_count + 1
return iteration_count
def extract(htmlstring,
base_url=None,
encoding="UTF-8",
syntaxes=SYNTAXES,
errors='strict',
uniform=False,
schema_context='http://schema.org',
**kwargs):
"""htmlstring: string with valid html document;
base_url: base url of the html document
encoding: encoding of the html document
syntaxes: list of syntaxes to extract, default SYNTAXES
errors: set to 'log' to log the exceptions, 'ignore' to ignore them
or 'strict'(default) to raise them
uniform: if True uniform output format of all syntaxes to a list of dicts.
Returned dicts structure:
{'@context': 'http://example.com',
'@type': 'example_type',
/* All other the properties in keys here */
}
schema_context: schema's context for current page"""
if base_url is None and 'url' in kwargs:
warnings.warn('"url" argument is deprecated, please use "base_url"',
DeprecationWarning)
base_url = kwargs.pop('url')
if kwargs:
raise TypeError('Unexpected keyword arguments')
if not (isinstance(syntaxes, list) and all(v in SYNTAXES
for v in syntaxes)):
raise ValueError("syntaxes must be a list with any or all (default) of"
"these values: {}".format(SYNTAXES))
if errors not in ['log', 'ignore', 'strict']:
raise ValueError('Invalid error command, valid values are either "log"'
', "ignore" or "strict"')
domparser = XmlDomHTMLParser(encoding=encoding)
tree = fromstring(htmlstring, parser=domparser)
processors = []
if 'microdata' in syntaxes:
processors.append(
('microdata', MicrodataExtractor().extract_items, tree))
if 'json-ld' in syntaxes:
processors.append(('json-ld', JsonLdExtractor().extract_items, tree))
if 'opengraph' in syntaxes:
processors.append(
('opengraph', OpenGraphExtractor().extract_items, tree))
if 'microformat' in syntaxes:
processors.append(
('microformat', MicroformatExtractor().extract_items, htmlstring))
if 'rdfa' in syntaxes:
processors.append(('rdfa', RDFaExtractor().extract_items, tree))
output = {}
for label, extract, document in processors:
try:
output[label] = list(extract(document, base_url=base_url))
except Exception:
if errors == 'log':
logger.exception('Failed to extract {}'.format(label))
if errors == 'ignore':
pass
if errors == 'strict':
raise
if uniform:
if 'microdata' in syntaxes:
output['microdata'] = _umicrodata_microformat(
output['microdata'], schema_context=schema_context)
if 'microformat' in syntaxes:
output['microformat'] = _umicrodata_microformat(
output['microformat'],
schema_context='http://microformats.org/wiki/')
if 'opengraph' in syntaxes:
output['opengraph'] = _uopengraph(output['opengraph'])
return output
def extract_all_rdfa(response):
rdfa_extractor = RDFaExtractor()
return rdfa_extractor.extract(response.text, url=response.url)
} for s in subset[0].get('article:section') or []]
json_ld_as_map = {
'@context': 'http://schema.org',
'@id': url,
'author': authors,
'section': sections
}
return json.dumps(json_ld_as_map)
def load_json_ld(tx, json_ld_data):
cypher_neosemantics = " CALL semantics.importRDFSnippet($payload,'JSON-LD');"
import_summary = tx.run(cypher_neosemantics, payload=json_ld_data)
print(import_summary)
uri = "bolt://localhost:7687"
driver = GraphDatabase.driver(uri, auth=("neo4j", "neo"))
rss_entries_as_json_ld, entry_url_list = get_rss(
'https://www.theguardian.com/uk/rss')
with driver.session() as session:
session.write_transaction(load_json_ld, rss_entries_as_json_ld)
rdfa_ext = RDFaExtractor()
for url in entry_url_list:
session.write_transaction(load_json_ld,
get_article_additional_details(url))
driver.close()
def extract(htmlstring,
base_url=None,
encoding="UTF-8",
syntaxes=SYNTAXES,
errors='strict',
uniform=False,
return_html_node=False,
schema_context='http://schema.org',
with_og_array=False,
**kwargs):
"""htmlstring: string with valid html document;
base_url: base url of the html document
encoding: encoding of the html document
syntaxes: list of syntaxes to extract, default SYNTAXES
errors: set to 'log' to log the exceptions, 'ignore' to ignore them
or 'strict'(default) to raise them
uniform: if True uniform output format of all syntaxes to a list of dicts.
Returned dicts structure:
{'@context': 'http://example.com',
'@type': 'example_type',
/* All other the properties in keys here */
}
return_html_node: if True, it includes into the result a HTML node of
respective embedded metadata under 'htmlNode' key.
The feature is supported only by microdata syntax.
Each node is of `lxml.etree.Element` type.
schema_context: schema's context for current page"""
if base_url is None and 'url' in kwargs:
warnings.warn('"url" argument is deprecated, please use "base_url"',
DeprecationWarning,
stacklevel=2)
base_url = kwargs.pop('url')
if kwargs:
raise TypeError('Unexpected keyword arguments')
if not (isinstance(syntaxes, list) and all(v in SYNTAXES
for v in syntaxes)):
raise ValueError("syntaxes must be a list with any or all (default) of"
"these values: {}".format(SYNTAXES))
if errors not in ['log', 'ignore', 'strict']:
raise ValueError('Invalid error command, valid values are either "log"'
', "ignore" or "strict"')
try:
tree = parse_xmldom_html(htmlstring, encoding=encoding)
except Exception as e:
if errors == 'ignore':
return {}
if errors == 'log':
logger.exception('Failed to parse html, raises {}'.format(e))
return {}
if errors == 'strict':
raise
processors = []
if 'microdata' in syntaxes:
processors.append(
('microdata',
MicrodataExtractor(add_html_node=return_html_node).extract_items,
tree))
if 'json-ld' in syntaxes:
processors.append((
'json-ld',
JsonLdExtractor().extract_items,
tree,
))
if 'opengraph' in syntaxes:
processors.append(
('opengraph', OpenGraphExtractor().extract_items, tree))
if 'microformat' in syntaxes:
processors.append(
('microformat', MicroformatExtractor().extract_items, htmlstring))
if 'rdfa' in syntaxes:
processors.append((
'rdfa',
RDFaExtractor().extract_items,
tree,
))
if 'dublincore' in syntaxes:
processors.append((
'dublincore',
DublinCoreExtractor().extract_items,
tree,
))
output = {}
for syntax, extract, document in processors:
try:
output[syntax] = list(extract(document, base_url=base_url))
except Exception as e:
if errors == 'log':
logger.exception('Failed to extract {}, raises {}'.format(
syntax, e))
if errors == 'ignore':
pass
if errors == 'strict':
raise
if uniform:
uniform_processors = []
if 'microdata' in syntaxes:
uniform_processors.append((
'microdata',
_umicrodata_microformat,
output['microdata'],
schema_context,
))
if 'microformat' in syntaxes:
uniform_processors.append((
'microformat',
_umicrodata_microformat,
output['microformat'],
'http://microformats.org/wiki/',
))
if 'opengraph' in syntaxes:
uniform_processors.append((
'opengraph',
_uopengraph,
output['opengraph'],
None,
))
if 'dublincore' in syntaxes:
uniform_processors.append((
'dublincore',
_udublincore,
output['dublincore'],
None,
))
for syntax, uniform, raw, schema_context in uniform_processors:
try:
if syntax == 'opengraph':
output[syntax] = uniform(raw, with_og_array=with_og_array)
elif syntax == 'dublincore':
output[syntax] = uniform(raw)
else:
output[syntax] = uniform(raw, schema_context)
except Exception as e:
if errors == 'ignore':
output[syntax] = []
if errors == 'log':
output[syntax] = []
logger.exception(
'Failed to uniform extracted for {}, raises {}'.format(
syntax, e))
if errors == 'strict':
raise
return output
class RISJMetadataExtractor(object):
"""An extruct-based metadata extractor"""
# TODO: Extend to microdata and RDFa, replacing bespoke xpath code. Then
# test on body of crawlers!
def __init__(self, response, microdata=False, jsonld=False, rdfa=False):
self.response = response
self.microdata = microdata
self.jsonld = jsonld
self.rdfa = rdfa
if rdfa:
try:
self.rdfae = RDFaExtractor()
self.rdfadata = self.rdfae.extract(self.response.text,
url=self.response.url)
except JSONDecodeError:
pass
if microdata:
try:
self.mde = MicrodataExtractor()
self.mdedata = self.mde.extract(self.response.text)
except JSONDecodeError:
pass
if jsonld:
try:
self.jlde = JsonLdExtractor()
self.jldata = self.jlde.extract(self.response.text)
except (JSONDecodeError, TypeError):
self.jldata = []
finally:
# Sometimes we get this in the meta dict from RISJExtractJSONLD
self.jldata.extend(self.response.meta.get('json-ld', []))
def extract_newsarticle_schemaorg(self,
microdata=None,
jsonld=None,
rdfa=None):
"""Extract schema.org NewsArticle metadata, encoded using any
supported metadata format. Note that we only try to extract the
*first* block of NewsArticle data for each method (which is then
combined with the first extracted from other methods if more than
one is selected."""
if microdata is None:
microdata = self.microdata
if jsonld is None:
jsonld = self.jsonld
if rdfa is None:
rdfa = self.rdfa
outd = {}
if jsonld:
for d in self.jldata:
# logger.debug('Analysing JSON-LD data: '+pformat(d))
try:
if (re.match(r'https?://schema.org/?', d['@context'])
and d['@type'] == 'NewsArticle'):
outd.update(d)
except (KeyError, TypeError):
continue
if microdata:
for d in self.mdedata:
logger.debug('Analysing W3C microdata: ' + pformat(d))
if re.match(r'https?://schema.org/NewsArticle/?',
d.get('type', '')):
outd.update(d)
if rdfa:
raise NotImplementedError
# logger.debug('Returning schema.org NewsArticle: '+pformat(outd))
return outd
