from urllib2 import urlopen
import logging
import os
+import sys
import types
+from pkg_resources import resource_listdir, resource_string
import lxml.html
import lxml.html.clean
return node
-def toTypedNode(value):
+def toTypedNode(value, language="en"):
"""Convert a python variable to a RDF Node with its closest xsd type
"""
if type(value) == types.BooleanType:
if value_type is not None:
node = RDF.Node(literal=value, datatype=value_type)
else:
- node = RDF.Node(literal=unicode(value).encode('utf-8'))
+ node = RDF.Node(literal=unicode(value).encode('utf-8'), language=language)
return node
def fromTypedNode(node):
"""Convert a typed RDF Node to its closest python equivalent
"""
- if node is None:
- return None
+ if not isinstance(node, RDF.Node):
+ return node
+ if node.is_resource():
+ return node
value_type = get_node_type(node)
literal = node.literal_value['string']
statements = []
retries = 3
+ succeeded = False
while retries > 0:
try:
retries -= 1
statements = rdf_parser.parse_as_stream(url, ns)
retries = 0
+ succeeded = True
except RDF.RedlandError, e:
errmsg = "RDF.RedlandError: {0} {1} tries remaining"
logger.error(errmsg.format(str(e), retries))
-
+
+ if not succeeded:
+ logger.warn("Unable to download %s", url)
+
for s in statements:
conditionally_add_statement(model, s, ns)
or in the list of directories provided in schema_path
"""
- if schema_path is None:
- path, _ = os.path.split(__file__)
- schema_path = [os.path.join(path, 'schemas')]
- elif type(schema_path) in types.StringTypes:
- schema_path = [schema_path]
+ schemas = resource_listdir(__name__, 'schemas')
+ for s in schemas:
+ schema = resource_string(__name__, 'schemas/' + s)
+ namespace = 'file://localhost/htsworkflow/schemas/'+s
+ add_schema(model, schema, namespace)
- for p in schema_path:
- for f in glob(os.path.join(p, '*.turtle')):
- add_schema(model, f)
+ if schema_path:
+ if type(schema_path) in types.StringTypes:
+ schema_path = [schema_path]
-def add_schema(model, filename):
+ for path in schema_path:
+ for pathname in glob(os.path.join(path, '*.turtle')):
+ url = 'file://' + os.path.splitext(pathname)[0]
+ stream = open(pathname, 'r')
+ add_schema(model, stream, url)
+ stream.close()
+
+def add_schema(model, schema, url):
"""Add a schema to a model.
Main difference from 'load_into_model' is it tags it with
"""
parser = RDF.Parser(name='turtle')
context = RDF.Node(RDF.Uri(SCHEMAS_URL))
- url = 'file://' + filename
- for s in parser.parse_as_stream(url):
+ for s in parser.parse_string_as_stream(schema, url):
try:
model.append(s, context)
except RDF.RedlandError as e:
writer.set_namespace('wot', wotNS._prefix)
# should these be here, kind of specific to an application
- writer.set_namespace('libraryOntology', libraryOntology._prefix)
+ writer.set_namespace('htswlib', libraryOntology._prefix)
writer.set_namespace('ucscSubmission', submissionOntology._prefix)
writer.set_namespace('ucscDaf', dafTermOntology._prefix)
+ writer.set_namespace('geoSoft', geoSoftNS._prefix)
+ writer.set_namespace('encode3', encode3NS._prefix)
return writer
+def get_turtle_header():
+ """Return a turtle header with our typical namespaces
+ """
+ serializer = get_serializer()
+ empty = get_model()
+ return serializer.serialize_model_to_string(empty)
def dump_model(model, destination=None):
if destination is None: