Skip to content

Commit

Permalink
Updated ontology and taglifter
Browse files Browse the repository at this point in the history
  • Loading branch information
@timgdavies
timgdavies committed Oct 4, 2015
1 parent 4b3d6b0 commit 792e6da
Show file tree
Hide file tree
Showing 2 changed files with 124 additions and 84 deletions.
92 changes: 53 additions & 39 deletions modules/taglifter.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -99,7 +99,7 @@ def get_country(self,row,path="#country",return_default = True):
if path + "+identifier" in row.keys():
country = row[path + '+identifier']
if(country):
return country.lower()
return country.upper()

if (len(country) < 2) and (path in row.keys()):
if row.get(path,"xx") in self.country_cache.keys():
Expand All @@ -117,7 +117,7 @@ def get_country(self,row,path="#country",return_default = True):
else:
country = "unknown"

return country.lower()
return country.upper()

# ToDo - refactor in future
def get_country_code_from_name(self,name):
Expand All @@ -126,12 +126,12 @@ def get_country_code_from_name(self,name):
except KeyError:
country = countrycode(codes=[name],origin='country_name',target="iso2c")[0]
if(len(country)==2):
self.country_cache[name] = country
self.country_cache[name] = country.upper()
else:
country = random_string()
self.country_cache[name] = country
self.country_cache[name] = country.upper()

return country.lower()
return country.upper()

def get_language(self,row):
lang = row.get("#language",default=self.default_language)
Expand Down Expand Up @@ -392,7 +392,7 @@ def build_graph(self):

# Now check for other possible relationships: https://github.com/timgdavies/tag-lifter/issues/1
possible_relationships = self.check_available_relationships(tag)

###Implementing https://github.com/timgdavies/tag-lifter/issues/1 (but very inefficiently!)
# (2) Does a column exist at the next level of depth?
for rel in possible_relationships:
Expand Down Expand Up @@ -421,6 +421,7 @@ def build_graph(self):
possible_relationships.remove(rel)
if search in entity_cache:
relationship = self.seek_class_relationship(entity,entity_class,entity_cache[search]['entity'],entity_cache[search]['class'],row,country,lang,source_row)
inverse_relationship = self.seek_class_relationship(entity_cache[search]['entity'],entity_cache[search]['class'],entity,entity_class,row,country,lang,source_row)
else:
print( search + " has not been created yet, so we can't relate to it")
except ValueError:
Expand Down Expand Up @@ -459,6 +460,9 @@ def build_graph(self):
if "skos:" in tag: # We are dealing with a skos tag
tag = tag.split(":")[1]
self.graph.add((entity,SKOS[tag],Literal(row[key],lang=col_lang)))
elif "foaf:" in tag: # We are dealing with a skos tag
tag = tag.split(":")[1]
self.graph.add((entity,FOAF[tag],Literal(row[key],lang=col_lang)))
else:
if self.allow_extra_properties:
self.graph.add((entity,self.ontology["misc/"+tag],Literal(row[key])))
Expand All @@ -478,33 +482,44 @@ def check_available_relationships(self,source):
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
SELECT DISTINCT ?t1 ?t2
WHERE {
?o a owl:Class.
?o a owl:Class.
FILTER(?o = <%s>)
{
?o rdfs:subClassOf+ ?restriction.
?restriction a owl:Restriction.
?restriction owl:onProperty ?p1.
{
{
?restriction owl:someValuesFrom ?t1.
} UNION {
?restriction owl:someValuesFrom ?restrictionClass.
?t1 rdfs:subClassOf+ ?restrictionClass.
{
?t1 rdfs:subClassOf+ ?restriction.
?restriction a owl:Restriction.
{
?restriction owl:someValuesFrom ?o.
} UNION {
?restriction owl:someValuesFrom ?restrictionClass.
?o rdfs:subClassOf+ ?restrictionClass.
}
} UNION
{
{
?o rdfs:subClassOf+ ?restriction.
?restriction a owl:Restriction.
{
{
?restriction owl:someValuesFrom ?t1.
} UNION {
?restriction owl:someValuesFrom ?restrictionClass.
?t1 rdfs:subClassOf+ ?restrictionClass.
}
}
}
}
}
OPTIONAL {
?t1 rdfs:subClassOf+ ?restriction2.
?restriction2 a owl:Restriction.
?restriction2 owl:onProperty ?p2.
{
{
?restriction2 owl:someValuesFrom ?t2.
} UNION {
?restriction2 owl:someValuesFrom ?restriction2class.
?t2 rdfs:subClassOf+ ?restriction2class.
OPTIONAL {
?t1 rdfs:subClassOf+ ?restriction2.
?restriction2 a owl:Restriction.
{
{
?restriction2 owl:someValuesFrom ?t2.
} UNION {
?restriction2 owl:someValuesFrom ?restriction2class.
?t2 rdfs:subClassOf+ ?restriction2class.
}
}
}
}
}
}"""%(source_class)
rels = []
Expand Down Expand Up @@ -543,20 +558,18 @@ def add_data_property(self,subj,subj_class,predicate,value,lang):
We rely on Pandas to cast some values right now...
"""

value = value.strip()
if ((predicate,RDFS.domain,subj_class)) in self.onto:
predicate_range = self.onto.value(predicate,RDFS.range)
if predicate_range == XSD.dateTime:
value = pandas.to_datetime(value)
elif predicate_range ==XSD.boolean:
value = True if value else False ## ToDo - improve the handling of Booleans
elif predicate_range ==XSD.float:
# value = str(value).translate(None,['1','2','3','4','5','7','8','9','0','.']) # ToDo - set-up better replacemnt for errange % signs etc.
value = locale.atoi(re.sub("[^\d\.]", "", value))
#try:
# value = locale.atof(str(value))
#except ValueError:
# value = value
try:
value = locale.atoi(re.sub("[^\d\.]", "", value))
except ValueError:
value = value
elif predicate_range ==XSD.integer:
# value = str(value).translate(None,['1','2','3','4','5','7','8','9','0','.'])
try:
Expand Down Expand Up @@ -584,9 +597,10 @@ def seek_class_relationship(self,source,source_class,target,target_class,row,cou
At present, the sparql only handles for owl:someValuesFrom - (ToDo: Extend this to include owl:allValuesFrom)
"""
if self.debug: print(str(source_class) + " -----> " + str(target_class))
if self.debug: print(str(source) + " -----> " + str(target))
if not source == target:
if self.debug: print(str(source_class) + " -----> " + str(target_class))
if self.debug: print(str(source) + " -----> " + str(target))

cache_key = str(source) + "_"+ str(target)
if cache_key in self.relationship_cache.keys():
return self.relationship_cache[cache_key]
Expand Down
Loading

0 comments on commit 792e6da

Please sign in to comment.