def compileQuadToParams(self,quadSlots):
subjSlot,predSlot,objSlot,conSlot = quadSlots
return (subjSlot.md5Int,
term2Letter(subjSlot.term),
objSlot.md5Int,
term2Letter(objSlot.term),
conSlot.md5Int,
term2Letter(conSlot.term))
def __init__(self, position, term, useSignedInts=False):
assert position in POSITION_LIST, "Unknown quad position: %s" % \
position
self.position = position
self.term = term
self.termType = term2Letter(term)
self.useSignedInts = useSignedInts
self.md5Int = normalizeValue(term, term2Letter(term), useSignedInts)
def __init__(self, position, term, useSignedInts=False):
assert position in POSITION_LIST, "Unknown quad position: %s" % \
position
self.position = position
self.term = term
self.termType = term2Letter(term)
self.useSignedInts = useSignedInts
self.md5Int = normalizeValue(term, term2Letter(term), useSignedInts)
def compileQuadToParams(self, quadSlots):
subjSlot, predSlot, objSlot, conSlot = quadSlots
dTypeParam = objSlot.term.datatype and normalizeValue(
objSlot.term.datatype, 'U', self.useSignedInts) or None
langParam = objSlot.term.language and objSlot.term.language or None
rtList = [
subjSlot.md5Int,
term2Letter(subjSlot.term), predSlot.md5Int,
term2Letter(predSlot.term), objSlot.md5Int, conSlot.md5Int,
term2Letter(conSlot.term)
]
for item in [dTypeParam, langParam]:
if item:
rtList.append(item)
return tuple(rtList)
def extractIdentifiers(self,quadSlots):
"""
Test literal data type extraction
>>> from rdflib.namespace import RDF
>>> from rdfextras.store.FOPLRelationalModel.QuadSlot import genQuadSlots
>>> class DummyClass:
... def __init__(self,test=False):
... self.test = test
... def updateIdentifierQueue(self,stuff):
... if self.test:
... term,termType = stuff[-1]
... assert termType == 'U',"Datatype's are URIs!"
>>> class Tester(NamedLiteralProperties):
... def __init__(self):
... self.idHash = DummyClass(True)
... self.valueHash = DummyClass()
>>> c = Tester()
>>> slots = genQuadSlots([BNode(),RDF.first,Literal(1),BNode()])
>>> c.extractIdentifiers(slots)
"""
subjSlot,predSlot,objSlot,conSlot = quadSlots
idTerms = [
(subjSlot.term,subjSlot.termType),
(predSlot.term,predSlot.termType),
(conSlot.term,conSlot.termType)]
if objSlot.term.datatype:
idTerms.append(
(objSlot.term.datatype,term2Letter(objSlot.term.datatype)))
self.idHash.updateIdentifierQueue(idTerms)
self.valueHash.updateIdentifierQueue(
[(objSlot.term,objSlot.termType)])
def extractIdentifiers(self, quadSlots):
"""
Test literal data type extraction
>>> from rdflib.namespace import RDF
>>> from rdfextras.store.FOPLRelationalModel.QuadSlot import genQuadSlots
>>> class DummyClass:
... def __init__(self,test=False):
... self.test = test
... def updateIdentifierQueue(self,stuff):
... if self.test:
... term,termType = stuff[-1]
... assert termType == 'U',"Datatype's are URIs!"
>>> class Tester(NamedLiteralProperties):
... def __init__(self):
... self.idHash = DummyClass(True)
... self.valueHash = DummyClass()
>>> c = Tester()
>>> slots = genQuadSlots([BNode(),RDF.first,Literal(1),BNode()])
>>> c.extractIdentifiers(slots)
"""
subjSlot, predSlot, objSlot, conSlot = quadSlots
idTerms = [(subjSlot.term, subjSlot.termType),
(predSlot.term, predSlot.termType),
(conSlot.term, conSlot.termType)]
if objSlot.term.datatype:
idTerms.append(
(objSlot.term.datatype, term2Letter(objSlot.term.datatype)))
self.idHash.updateIdentifierQueue(idTerms)
self.valueHash.updateIdentifierQueue([(objSlot.term, objSlot.termType)
])
def compileQuadToParams(self,quadSlots):
subjSlot,predSlot,objSlot,conSlot = quadSlots
dTypeParam = objSlot.term.datatype and normalizeValue(
objSlot.term.datatype, 'U', self.useSignedInts) or None
langParam = objSlot.term.language and objSlot.term.language or None
rtList = [
subjSlot.md5Int,
term2Letter(subjSlot.term),
predSlot.md5Int,
term2Letter(predSlot.term),
objSlot.md5Int,
conSlot.md5Int,
term2Letter(conSlot.term)]
for item in [dTypeParam,langParam]:
if item:
rtList.append(item)
return tuple(rtList)
def generateWhereClause(self,queryPattern):
"""
Takes a query pattern (a list of quad terms -
subject,predicate,object,context) and generates a SQL WHERE clauses
which works in conjunction to the intersections to filter the result
set by partial matching (by REGEX), full matching (by integer
half-hash), and term types. For maximally efficient SELECT queries
"""
whereClauses = []
whereParameters = []
asserted = dereferenceQuad(CONTEXT,queryPattern) is None
for idx in SlotPrefixes.keys():
queryTerm = dereferenceQuad(idx,queryPattern)
lookupAlias = 'rt_'+SlotPrefixes[idx]
if idx == CONTEXT and asserted:
whereClauses.append("%s.%s_term != 'F'" % \
(self,self.columnNames[idx]))
if idx < len(POSITION_LIST) and isinstance(queryTerm,REGEXTerm):
whereClauses.append("%s.lexical REGEXP "%lookupAlias+"%s")
whereParameters.append(queryTerm)
elif idx == CONTEXT \
and isinstance(queryTerm,Graph) \
and isinstance(queryTerm.identifier,REGEXTerm):
whereClauses.append("%s.lexical REGEXP "%lookupAlias+"%s")
whereParameters.append(queryTerm.identifier)
elif idx < len(POSITION_LIST) and queryTerm is not Any:
if self.columnNames[idx]:
if isinstance(queryTerm,list):
whereClauses.append("%s.%s" % \
(self,self.columnNames[idx])+" in (%s)" % \
','.join([
'%s' for item in range(len(queryTerm))
]))
whereParameters.extend(
[normalizeValue(item, term2Letter(item),
self.useSignedInts)
for item in queryTerm])
else:
whereClauses.append("%s.%s" % \
(self,self.columnNames[idx])+" = %s")
whereParameters.append(normalizeValue(
queryTerm, term2Letter(queryTerm),
self.useSignedInts))
if not idx in self.hardCodedResultTermsTypes \
and self.termEnumerations[idx] \
and not isinstance(queryTerm,list):
whereClauses.append("%s.%s_term" % \
(self,self.columnNames[idx])+" = %s")
whereParameters.append(term2Letter(queryTerm))
elif idx >= len(POSITION_LIST) \
and len(self.columnNames) > len(POSITION_LIST) \
and queryTerm is not None:
compVal = idx == DATATYPE_INDEX and normalizeValue(
queryTerm, term2Letter(queryTerm),
self.useSignedInts) or queryTerm
whereClauses.append("%s.%s" % \
(self,self.columnNames[idx][0])+" = %s")
whereParameters.append(compVal)
return ' AND '.join(whereClauses),whereParameters
def normalizeNode(node, useSignedInts=False):
return normalizeValue(node, term2Letter(node), useSignedInts)
def compileQuadToParams(self, quadSlots):
subjSlot, predSlot, objSlot, conSlot = quadSlots
return (subjSlot.md5Int, term2Letter(subjSlot.term), predSlot.md5Int,
term2Letter(predSlot.term), objSlot.md5Int,
term2Letter(objSlot.term), conSlot.md5Int,
term2Letter(conSlot.term))
def generateWhereClause(self, queryPattern):
"""
Takes a query pattern (a list of quad terms -
subject,predicate,object,context) and generates a SQL WHERE clauses
which works in conjunction to the intersections to filter the result
set by partial matching (by REGEX), full matching (by integer
half-hash), and term types. For maximally efficient SELECT queries
"""
whereClauses = []
whereParameters = []
asserted = dereferenceQuad(CONTEXT, queryPattern) is None
for idx in SlotPrefixes.keys():
queryTerm = dereferenceQuad(idx, queryPattern)
lookupAlias = 'rt_' + SlotPrefixes[idx]
if idx == CONTEXT and asserted:
whereClauses.append("%s.%s_term != 'F'" % \
(self,self.columnNames[idx]))
if idx < len(POSITION_LIST) and isinstance(queryTerm, REGEXTerm):
whereClauses.append("%s.lexical REGEXP " % lookupAlias + "%s")
whereParameters.append(queryTerm)
elif idx == CONTEXT \
and isinstance(queryTerm,Graph) \
and isinstance(queryTerm.identifier,REGEXTerm):
whereClauses.append("%s.lexical REGEXP " % lookupAlias + "%s")
whereParameters.append(queryTerm.identifier)
elif idx < len(POSITION_LIST) and queryTerm is not Any:
if self.columnNames[idx]:
if isinstance(queryTerm, list):
whereClauses.append("%s.%s" % \
(self,self.columnNames[idx])+" in (%s)" % \
','.join([
'%s' for item in range(len(queryTerm))
]))
whereParameters.extend([
normalizeValue(item, term2Letter(item),
self.useSignedInts)
for item in queryTerm
])
else:
whereClauses.append("%s.%s" % \
(self,self.columnNames[idx])+" = %s")
whereParameters.append(
normalizeValue(queryTerm, term2Letter(queryTerm),
self.useSignedInts))
if not idx in self.hardCodedResultTermsTypes \
and self.termEnumerations[idx] \
and not isinstance(queryTerm,list):
whereClauses.append("%s.%s_term" % \
(self,self.columnNames[idx])+" = %s")
whereParameters.append(term2Letter(queryTerm))
elif idx >= len(POSITION_LIST) \
and len(self.columnNames) > len(POSITION_LIST) \
and queryTerm is not None:
compVal = idx == DATATYPE_INDEX and normalizeValue(
queryTerm, term2Letter(queryTerm),
self.useSignedInts) or queryTerm
whereClauses.append("%s.%s" % \
(self,self.columnNames[idx][0])+" = %s")
whereParameters.append(compVal)
return ' AND '.join(whereClauses), whereParameters
def normalizeNode(node, useSignedInts=False):
return normalizeValue(node, term2Letter(node), useSignedInts)
