import re def mapQNameToPyName(prefix, lname) : if prefix == 'skos' : v = 'SKOS[\''+lname+'\']' elif prefix == 'skosxl' : v = 'XL[\''+lname+'\']' elif prefix == 'rdf' : v = 'RDF.'+lname elif prefix == 'rdfs' : v = 'RDFS.'+lname elif prefix == 'owl' : v = 'OWL.'+lname else : v = 'FIXME' return v def renderTypeTriples(out, sid, s, t) : out.write('\t'+sid+'.append((XL[\''+s+'\'], RDF.type, OWL[\''+t+'\']))\n') if ( t == 'AnnotationProperty' or t == 'ObjectProperty' or t == 'DatatypeProperty' ) : out.write('\t'+sid+'.append((XL[\''+s+'\'], RDF.type, RDF.Property))\n') def generatePythonCodeFromStatements(textfile, outfile): outfile.write('from rdflib import Namespace, RDF, RDFS, ConjunctiveGraph\n\n') outfile.write('def statements() :\n\n') outfile.write('\tXL = Namespace(\'http://www.w3.org/2008/05/skos-xl#\')\n') outfile.write('\tSKOS = Namespace(\'http://www.w3.org/2004/02/skos/core#\')\n') outfile.write('\tOWL = Namespace(\'http://www.w3.org/2002/07/owl#\')\n\n') outfile.write('\tstatements = []\n\n') statementRe = re.compile('#\\s(\\S+)') typeRe = re.compile('# S\\d+ skosxl:(\\w+) is an instance of owl:(\w+)') typeTwoRe = re.compile('# S\\d+ skosxl:(\\w+) and skosxl:(\\w+) are each instances of owl:(\\w+)') typeMultiRe = re.compile('# S\\d+ ((?:skosxl:\\w+, )+)skosxl:(\\w+) and skosxl:(\\w+) are each instances of owl:(\\w+)') domainRangeRe = re.compile('# S\\d+ The rdfs:(range|domain) of skosxl:(\\w+) is the class (\\w+):(\\w+)') domainsRangesRe = re.compile('# S\\d+ The rdfs:(range|domain) of each of ((?:skosxl:\\w+, )*)skosxl:(\\w+) and skosxl:(\\w+) is the class (\\w+):(\\w+)') disjointSKOSClassRe = re.compile('# S\\d+ skosxl:(\\w+) is disjoint with skos:(\\w+)') subPropertiesRe = re.compile('# S\\d+ ((?:skosxl:\\w+, )+)skosxl:(\\w+) and skosxl:(\\w+) are each sub-properties of skosxl:(\\w+)') subPropertyRe = re.compile('skosxl:(\\w+) is a sub-property of skosxl:(\\w+)') inverseRe = re.compile('skosxl:(\\w+) is the owl:inverseOf the property skosxl:(\\w+)') disjointSKOSClassesRe = re.compile('# S\\d+ skosxl:(\\w+) is disjoint with each of ((?:skos:\\w+, )*)skos:(\\w+) and skos:(\\w+)') subClassRe = re.compile('skosxl:(\\w+) is a sub-class of skosxl:(\\w+)') for line in textfile : # output line as comment outfile.write('\t'+line) # output statement array initialisation sid = statementRe.search(line).group(1) outfile.write('\t'+sid+' = []\n') # output triples appended typeReMatch = typeRe.search(line) typeTwoReMatch = typeTwoRe.search(line) typeMultiReMatch = typeMultiRe.search(line) domainRangeReMatch = domainRangeRe.search(line) domainsRangesReMatch = domainsRangesRe.search(line) disjointSKOSClassReMatch = disjointSKOSClassRe.search(line) subPropertiesReMatch = subPropertiesRe.search(line) subPropertyReMatchGroups = subPropertyRe.findall(line) inverseReMatchGroups = inverseRe.findall(line) disjointSKOSClassesReMatch = disjointSKOSClassesRe.search(line) subClassReMatchGroups = subClassRe.findall(line) if typeReMatch != None : #typeRe = re.compile('# S\\d+ skos:(\\w+) is an instance of owl:(\w+)') renderTypeTriples(outfile, sid, typeReMatch.group(1), typeReMatch.group(2)) elif typeTwoReMatch != None : renderTypeTriples(outfile, sid, typeTwoReMatch.group(1), typeTwoReMatch.group(3)) renderTypeTriples(outfile, sid, typeTwoReMatch.group(2), typeTwoReMatch.group(3)) elif typeMultiReMatch != None : # comma values for ln in typeMultiReMatch.group(1).split(',') : ln = ln.replace(' ','') if len(ln) > 0 : renderTypeTriples(outfile, sid, ln.partition(':')[2], typeMultiReMatch.group(4)) renderTypeTriples(outfile, sid, typeMultiReMatch.group(2), typeMultiReMatch.group(4)) renderTypeTriples(outfile, sid, typeMultiReMatch.group(3), typeMultiReMatch.group(4)) elif domainRangeReMatch != None : v = mapQNameToPyName(domainRangeReMatch.group(3), domainRangeReMatch.group(4)) # if domainRangeReMatch.group(3) == 'xl' : # v = 'XL[\''+domainRangeReMatch.group(4)+'\']' # elif domainRangeReMatch.group(3) == 'rdf' : # v = 'RDF.'+domainRangeReMatch.group(4) # elif domainRangeReMatch.group(3) == 'rdfs' : # v = 'RDFS.'+domainRangeReMatch.group(4) # elif domainRangeReMatch.group(3) == 'owl' : # v = 'OWL.'+domainRangeReMatch.group(4) # else : # v = 'FIXME' outfile.write('\t'+sid+'.append((XL[\''+domainRangeReMatch.group(2)+'\'], RDFS.'+domainRangeReMatch.group(1)+', '+v+'))\n') elif domainsRangesReMatch != None: v = mapQNameToPyName(domainsRangesReMatch.group(5), domainsRangesReMatch.group(6)) # if domainsRangesReMatch.group(5) == 'xl' : # v = 'XL[\''+domainsRangesReMatch.group(6)+'\']' # elif domainsRangesReMatch.group(5) == 'rdf' : # v = 'RDF.'+domainsRangesReMatch.group(6) # elif domainsRangesReMatch.group(5) == 'rdfs' : # v = 'RDFS.'+domainsRangesReMatch.group(6) # elif domainsRangesReMatch.group(5) == 'owl' : # v = 'OWL.'+domainsRangesReMatch.group(6) # else : # v = 'FIXME' # comma values for ln in domainsRangesReMatch.group(2).split(',') : ln = ln.replace(' ','') if len(ln) > 0 : outfile.write('\t'+sid+'.append((XL[\''+ln.partition(':')[2]+'\'], RDFS.'+domainsRangesReMatch.group(1)+', '+v+'))\n') outfile.write('\t'+sid+'.append((XL[\''+domainsRangesReMatch.group(3)+'\'], RDFS.'+domainsRangesReMatch.group(1)+', '+v+'))\n') outfile.write('\t'+sid+'.append((XL[\''+domainsRangesReMatch.group(4)+'\'], RDFS.'+domainsRangesReMatch.group(1)+', '+v+'))\n') elif disjointSKOSClassReMatch != None : outfile.write('\t'+sid+'.append((XL[\''+disjointSKOSClassReMatch.group(1)+'\'], OWL[\'disjointWith\'], SKOS[\''+disjointSKOSClassReMatch.group(2)+'\']))\n') elif subPropertiesReMatch != None : # outfile.write('\t# groups: '+str(subPropertiesReMatch.groups())+'\n') # comma values for ln in subPropertiesReMatch.group(1).split(',') : ln = ln.replace(' ','') if len(ln) > 0 : outfile.write('\t'+sid+'.append((XL[\''+ln.partition(':')[2]+'\'], RDFS.subPropertyOf, XL[\''+subPropertiesReMatch.group(4)+'\']))\n') outfile.write('\t'+sid+'.append((XL[\''+subPropertiesReMatch.group(2)+'\'], RDFS.subPropertyOf, XL[\''+subPropertiesReMatch.group(4)+'\']))\n') outfile.write('\t'+sid+'.append((XL[\''+subPropertiesReMatch.group(3)+'\'], RDFS.subPropertyOf, XL[\''+subPropertiesReMatch.group(4)+'\']))\n') elif len(subPropertyReMatchGroups)>0 : for group in subPropertyReMatchGroups : outfile.write('\t'+sid+'.append((XL[\''+group[0]+'\'], RDFS.subPropertyOf, XL[\''+group[1]+'\']))\n') elif len(inverseReMatchGroups)>0 : for group in inverseReMatchGroups : outfile.write('\t'+sid+'.append((XL[\''+group[0]+'\'], OWL[\'inverseOf\'], XL[\''+group[1]+'\']))\n') outfile.write('\t'+sid+'.append((XL[\''+group[1]+'\'], OWL[\'inverseOf\'], XL[\''+group[0]+'\']))\n') elif disjointSKOSClassesReMatch != None : # outfile.write('\t# groups: '+str(disjointClassesReMatch.groups())+'\n') # comma values for ln in disjointSKOSClassesReMatch.group(2).split(',') : ln = ln.replace(' ','') if len(ln) > 0 : outfile.write('\t'+sid+'.append((XL[\''+disjointSKOSClassesReMatch.group(1)+'\'], OWL[\'disjointWith\'], SKOS[\''+ln.partition(':')[2]+'\']))\n') outfile.write('\t'+sid+'.append((XL[\''+disjointSKOSClassesReMatch.group(1)+'\'], OWL[\'disjointWith\'], SKOS[\''+disjointSKOSClassesReMatch.group(3)+'\']))\n') outfile.write('\t'+sid+'.append((XL[\''+disjointSKOSClassesReMatch.group(1)+'\'], OWL[\'disjointWith\'], SKOS[\''+disjointSKOSClassesReMatch.group(4)+'\']))\n') elif len(subClassReMatchGroups)>0 : for group in subClassReMatchGroups : outfile.write('\t'+sid+'.append((XL[\''+group[0]+'\'], RDFS.subClassOf, XL[\''+group[1]+'\']))\n') else : outfile.write('\t# no triples to add\n') # output statements appended outfile.write('\tstatements.append('+sid+')\n\n') outfile.write('\treturn statements\n\n') outfile.write('def schema() :\n\n') outfile.write('\tschema = ConjunctiveGraph()\n') outfile.write('\tfor statement in statements() :\n') outfile.write('\t for triple in statement :\n') outfile.write('\t schema.add(triple)\n\n') outfile.write('\treturn schema\n') outfile.write('# all done') textfile = open('./skos.txt', 'r') outfile = open('./skosxl.py', 'w') generatePythonCodeFromStatements(textfile, outfile) textfile.close() outfile.close()