def get_predicates():
"""
get_predicates
Get a list of predicates used in statements issued by the knowledge source
:rtype: List[Predicate]
"""
# Possible types: {'CommonsMedia', 'Time', 'Quantity', 'WikibaseProperty', 'WikibaseItem', 'GlobeCoordinate',
# 'String', 'ExternalId', 'Math', 'Monolingualtext', 'TabularData', 'Url', 'GeoShape'}
query = """SELECT ?p ?pt ?pLabel ?d ?aliases WHERE {
{
SELECT ?p ?pt ?d (GROUP_CONCAT(DISTINCT ?alias; separator="|") as ?aliases) WHERE {
?p wikibase:propertyType ?pt .
OPTIONAL {?p skos:altLabel ?alias FILTER (LANG (?alias) = "en")}
OPTIONAL {?p schema:description ?d FILTER (LANG (?d) = "en") .}
} GROUP BY ?p ?pt ?d
}
SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}"""
results = execute_sparql_query(query)['results']['bindings']
results = [{k: v['value'] for k, v in item.items()} for item in results]
print(results[0])
items = [{
'id': "wd:" + x['p'].split("/")[-1],
'name': x['pLabel'],
'definition': x.get("d", ""),
'aliases': x['aliases'].split("|") if x['aliases'] else [],
'ptype': x['pt'].replace("http://wikiba.se/ontology#", "")
} for x in results]
# note: 'aliases' and 'ptype' are not in the official spec!
return items
def get_equiv_item(curie):
"""
From a curie, get the wikidata item
get_equiv_item("PMID:10028264")
:param curie:
:return:
"""
try:
pid, value = cu.parse_curie(curie)
except ValueError as e:
print(e)
return []
prop_direct = "<http://www.wikidata.org/prop/direct/{}>".format(
pid.split("/")[-1])
query_str = "SELECT ?item WHERE {{ ?item {} '{}' }}".format(
prop_direct, value)
d = execute_sparql_query(query_str)['results']['bindings']
equiv_qids = list(
set(chain(*[{v['value']
for k, v in x.items()} for x in d])))
equiv_qids = [
"wd:" + x.replace("http://www.wikidata.org/entity/", "")
for x in equiv_qids
]
return equiv_qids
def get_all_types():
"""
Get all semantic group types, and their counts.
:return: {"id": [], "frequency": xx} for all entity types in garbanzo
"""
agg = {}
for (entity_id, group_name) in qid_semgroup.items():
if isinstance(group_name, str):
group_name = [group_name]
for group in group_name:
if entity_id != 'Q5': # Q5 = human, can't do a count
query_str = """SELECT (COUNT (DISTINCT ?type) AS ?count) WHERE {{?type wdt:P31 wd:{0}}}""".format(
entity_id)
agg[entity_id] = {
'sum':
int(
execute_sparql_query(query_str)['results']['bindings']
[0]['count']['value']),
'group':
group
}
# ret = [{'id': '{} wd:{}'.format(v['group'], k), 'frequency': v['sum']} for k,v in agg.items()]
ret = [{
'id': 'wd:{}'.format(k),
'frequency': v['sum']
} for k, v in agg.items()]
return ret
def get_concept_details(qid):
"""
This will form the "details" for the GET /translator/concepts/{conceptId} endpoint
:param qid:
:return:
"""
query_str = """SELECT distinct ?prop ?propLabel ?value ?valueLabel
WHERE
{{
{} ?p ?value .
?prop wikibase:directClaim ?p .
SERVICE wikibase:label {{ bd:serviceParam wikibase:language "en" . }}
}}""".format(always_curie(qid))
d = execute_sparql_query(query_str)['results']['bindings']
results = [{k: v['value'] for k, v in item.items()} for item in d]
for result in results:
result['prop'] = result['prop'].replace(
"http://www.wikidata.org/entity/", "wd:")
result['tag'] = result['prop']
result['value'] = result['value'].replace(
"http://www.wikidata.org/entity/", "wd:")
return results
def _query_statements(s, t=None, relations=None, direction="f"):
"""
if direction = f (forward), s is source, t is target
if direction = r (reverse), s and t are reversed
if t is not given, target is unconstrained
if relations is not given, relations are unconstrained
"""
assert direction in {"f", "r"}, "direction must be 'f' or 'r'"
s = set(map(always_curie, s))
t = set(map(always_curie, t)) if t else set()
relations = set(map(always_curie, relations)) if relations else set()
s_str = " ".join(s)
t_str = " ".join(t)
r_str = " ".join(relations)
if direction == "r":
s_str, t_str = t_str, s_str
query_str = """
SELECT ?s ?sLabel ?r ?rLabel ?t ?tLabel ?id (GROUP_CONCAT(?stype) as ?stypes) (GROUP_CONCAT(?ttype) as ?ttypes) WHERE {{
{source_filter}
{target_filter}
{relation_filter}
?s ?propertyclaim ?id .
?r wikibase:claim ?propertyclaim .
?id ?b ?t .
OPTIONAL {{?s wdt:P31 ?stype}}
OPTIONAL {{?t wdt:P31 ?ttype}}
FILTER(regex(str(?b), "http://www.wikidata.org/prop/statement" ))
SERVICE wikibase:label {{ bd:serviceParam wikibase:language "en" }}
}} GROUP BY ?s ?sLabel ?r ?rLabel ?t ?tLabel ?id"""
query_str = query_str.format(
source_filter="values ?s {" + s_str + "}" if s_str else "",
target_filter="values ?t {" + t_str + "}" if t_str else "",
relation_filter="values ?r {" + r_str + "}" if r_str else "")
d = execute_sparql_query(query_str)['results']['bindings']
results = [{k: v['value'] for k, v in item.items()} for item in d]
# remove non item statements
results = [
x for x in results if "http://www.wikidata.org/entity/" in x['s']
and "http://www.wikidata.org/entity/" in x['t']
]
for result in results:
result['s'] = result['s'].replace("http://www.wikidata.org/entity/",
"wd:")
result['r'] = result['r'].replace("http://www.wikidata.org/entity/",
"wd:")
result['t'] = result['t'].replace("http://www.wikidata.org/entity/",
"wd:")
result['id'] = result['id'].replace(
"http://www.wikidata.org/entity/statement/",
"wds:").replace("-", "$", 1)
sType = [
x.replace("http://www.wikidata.org/entity/", "wd:")
for x in result['stypes'].split(" ")
] if result['stypes'] else []
tType = [
x.replace("http://www.wikidata.org/entity/", "wd:")
for x in result['ttypes'].split(" ")
] if result['ttypes'] else []
result['sSemanticGroup'] = " ".join(
get_semgroups_from_qids(sType)) if sType else ""
result['tSemanticGroup'] = " ".join(
get_semgroups_from_qids(tType)) if tType else ""
results = [x for x in results if x['id'].startswith("wds:Q")]
data = [{
'id': s['id'],
'subject': {
'id': s['s'],
'name': s['sLabel'],
'semanticGroup': s['sSemanticGroup']
},
'predicate': {
'id': s['r'],
'name': s['rLabel']
},
'object': {
'id': s['t'],
'name': s['tLabel'],
'semanticGroup': s['tSemanticGroup']
},
} for s in results]
#if direction == "r":
# for d in data:
# d['subject'], d['object'] = d['object'], d['subject']
return data