# file is a local file when we suppose it to be a utf-8 dom = parser.parse(input, encoding=self.charset) else : # No charset set. The HTMLLib parser tries to sniff into the # the file to find a meta header for the charset; if that # works, fine, otherwise it falls back on window-... dom = parser.parse(input) try : if isinstance(name, basestring) : input.close() input = self._get_input(name) else : input.seek(0) from pyRdfa.host import adjust_html_version self.rdfa_version = adjust_html_version(input, self.rdfa_version) except : # if anyting goes wrong, it is not really important; rdfa version stays what it was... pass else : # in other cases an XML parser has to be used from pyRdfa.host import adjust_xhtml_and_version parse = xml.dom.minidom.parse dom = parse(input) (adjusted_host_language, version) = adjust_xhtml_and_version(dom, self.options.host_language, self.rdfa_version) self.options.host_language = adjusted_host_language self.rdfa_version = version except Exception, e : # These are various parsing exception. Per spec, this is a case when # error triples MUST be returned, ie, the usage of rdfOutput (which switches between an HTML formatted
def graph_from_source(self, name, graph = None, rdfOutput = False, pgraph = None) :
"""
Extract an RDF graph from an RDFa source. The source is parsed, the RDF extracted, and the RDFa Graph is
returned. This is a front-end to the L{pyRdfa.graph_from_DOM} method.
@param name: a URI, a file name, or a file-like object
@param graph: rdflib Graph instance. If None, a new one is created.
@param pgraph: rdflib Graph instance for the processor graph. If None, and the error/warning triples are to be generated, they will be added to the returned graph. Otherwise they are stored in this graph.
@param rdfOutput: whether runtime exceptions should be turned into RDF and returned as part of the processor graph
@return: an RDF Graph
@rtype: rdflib Graph instance
"""
def copyErrors(tog, options) :
if tog == None :
tog = Graph()
if options.output_processor_graph :
for t in options.processor_graph.graph :
tog.add(t)
for k,ns in options.processor_graph.graph.namespaces() :
tog.bind(k,ns)
options.reset_processor_graph()
return tog
# Separating this for a forward Python 3 compatibility
try :
# Python 2 branch
isstring = isinstance(name, basestring)
except :
# Python 3 branch
isstring = isinstance(name, str)
try :
# First, open the source... Possible HTTP errors are returned as error triples
input = None
try :
input = self._get_input(name)
except FailedSource :
f = sys.exc_info()[1]
self.http_status = 400
if not rdfOutput : raise f
err = self.options.add_error(f.msg, FileReferenceError, name)
self.options.processor_graph.add_http_context(err, 400)
return copyErrors(graph, self.options)
except HTTPError :
h = sys.exc_info()[1]
self.http_status = h.http_code
if not rdfOutput : raise h
err = self.options.add_error("HTTP Error: %s (%s)" % (h.http_code,h.msg), HTError, name)
self.options.processor_graph.add_http_context(err, h.http_code)
return copyErrors(graph, self.options)
except Exception :
e = sys.exc_info()[1]
self.http_status = 500
# Something nasty happened:-(
if not rdfOutput : raise e
err = self.options.add_error(str(e), context = name)
self.options.processor_graph.add_http_context(err, 500)
return copyErrors(graph, self.options)
dom = None
try :
msg = ""
parser = None
if self.options.host_language == HostLanguage.html5 :
import warnings
warnings.filterwarnings("ignore", category=DeprecationWarning)
import html5lib
parser = html5lib.HTMLParser(tree=html5lib.treebuilders.getTreeBuilder("dom"))
if self.charset :
# This means the HTTP header has provided a charset, or the
# file is a local file when we suppose it to be a utf-8
dom = parser.parse(input, encoding=self.charset)
else :
# No charset set. The HTMLLib parser tries to sniff into the
# the file to find a meta header for the charset; if that
# works, fine, otherwise it falls back on window-...
dom = parser.parse(input)
try :
if isstring :
input.close()
input = self._get_input(name)
else :
input.seek(0)
from pyRdfa.host import adjust_html_version
self.rdfa_version = adjust_html_version(input, self.rdfa_version)
except :
# if anyting goes wrong, it is not really important; rdfa version stays what it was...
pass
else :
# in other cases an XML parser has to be used
from pyRdfa.host import adjust_xhtml_and_version
parse = xml.dom.minidom.parse
dom = parse(input)
(adjusted_host_language, version) = adjust_xhtml_and_version(dom, self.options.host_language, self.rdfa_version)
self.options.host_language = adjusted_host_language
self.rdfa_version = version
except ImportError :
msg = "HTML5 parser not available. Try installing html5lib <http://code.google.com/p/html5lib>"
raise ImportError(msg)
except Exception :
e = sys.exc_info()[1]
# These are various parsing exception. Per spec, this is a case when
# error triples MUST be returned, ie, the usage of rdfOutput (which switches between an HTML formatted
# return page or a graph with error triples) does not apply
err = self.options.add_error(str(e), context = name)
self.http_status = 400
self.options.processor_graph.add_http_context(err, 400)
return copyErrors(graph, self.options)
# If we got here, we have a DOM tree to operate on...
return self.graph_from_DOM(dom, graph, pgraph)
except Exception :
# Something nasty happened during the generation of the graph...
(a,b,c) = sys.exc_info()
sys.excepthook(a,b,c)
if isinstance(b, ImportError) :
self.http_status = None
else :
self.http_status = 500
if not rdfOutput : raise b
err = self.options.add_error(str(b), context = name)
self.options.processor_graph.add_http_context(err, 500)
return copyErrors(graph, self.options)
