def resolve_keyword(keyword):
query = make_template("""
prefix anno: <http://www.eha.io/types/annotation_prop/>
prefix oboInOwl: <http://www.geneontology.org/formats/oboInOwl#>
prefix obo: <http://purl.obolibrary.org/obo/>
prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
SELECT ?entity
WHERE {
BIND (obo:DOID_4 AS ?disease)
?entity rdfs:subClassOf* ?disease .
?entity oboInOwl:hasNarrowSynonym|oboInOwl:hasRelatedSynonym|oboInOwl:hasExactSynonym|rdfs:label ?label
FILTER regex(?label, "^({{keyword | escape}})$", "i")
}
""").render(keyword=re.escape(keyword))
resp = sparql_utils.query(query)
bindings = resp.json()['results']['bindings']
if len(bindings) == 0:
print("no match for", keyword.encode('ascii', 'xmlcharrefreplace'))
elif len(bindings) > 1:
print("multiple matches for",
keyword.encode('ascii', 'xmlcharrefreplace'))
print(bindings)
return [binding['entity']['value'] for binding in bindings]
def print_result(result):
for binding in result.json()['results']['bindings']:
for key, value in binding.items():
raw_val = value['value']
print "[" + key + "]"
# Check for the delimiter used to combine results in a "group by" query group.
if ";;" in raw_val:
print raw_val.split(";;")
continue
# If the value references an annotation, query it and display
# the full text.
if raw_val.startswith('http://www.eha.io/types/annotation/'):
query = make_template("""
prefix anno: <http://www.eha.io/types/annotation_prop/>
prefix dep: <http://www.eha.io/types/annotation_prop/dep/>
prefix con: <http://www.eha.io/types/content/>
SELECT ?phraseStart ?phraseEnd ?prepStart ?sourceText
WHERE {
<{{annotation_uri}}> anno:start ?phraseStart
; anno:end ?phraseEnd
; anno:source_doc/con:text ?sourceText
}
""").render(annotation_uri=raw_val)
result = sparql_utils.query(query)
bindings = result.json()['results']['bindings']
if len(bindings) == 0:
print "Could not resolve source text for:"
print key, raw_val
for binding in bindings:
text = binding['sourceText']['value']
start = int(binding['phraseStart']['value'])
end = int(binding['phraseEnd']['value'])
print text[start:end]
else:
print raw_val
print ""
print "~~--~~--~~"
def print_result(result):
for binding in result.json()['results']['bindings']:
for key, value in binding.items():
raw_val = value['value']
print "[" + key + "]"
# Check for the delimiter used to combine results in a "group by" query group.
if ";;" in raw_val:
print raw_val.split(";;")
continue
# If the value references an annotation, query it and display
# the full text.
if raw_val.startswith('http://www.eha.io/types/annotation/'):
query = make_template("""
prefix anno: <http://www.eha.io/types/annotation_prop/>
prefix dep: <http://www.eha.io/types/annotation_prop/dep/>
prefix con: <http://www.eha.io/types/content/>
SELECT ?phraseStart ?phraseEnd ?prepStart ?sourceText
WHERE {
<{{annotation_uri}}> anno:start ?phraseStart
; anno:end ?phraseEnd
; anno:source_doc/con:text ?sourceText
}
""").render(annotation_uri=raw_val)
result = sparql_utils.query(query)
bindings = result.json()['results']['bindings']
if len(bindings) == 0:
print "Could not resolve source text for:"
print key, raw_val
for binding in bindings:
text = binding['sourceText']['value']
start = int(binding['phraseStart']['value'])
end = int(binding['phraseEnd']['value'])
print text[start:end]
else:
print raw_val
print ""
print "~~--~~--~~"
def resolve_keyword(keyword):
query = make_template("""
prefix anno: <http://www.eha.io/types/annotation_prop/>
prefix oboInOwl: <http://www.geneontology.org/formats/oboInOwl#>
prefix obo: <http://purl.obolibrary.org/obo/>
prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
SELECT ?entity
WHERE {
BIND (obo:DOID_4 AS ?disease)
?entity rdfs:subClassOf* ?disease .
?entity oboInOwl:hasNarrowSynonym|oboInOwl:hasRelatedSynonym|oboInOwl:hasExactSynonym|rdfs:label ?label
FILTER regex(?label, "^({{keyword | escape}})$", "i")
}
""").render(
keyword=re.escape(keyword)
)
resp = sparql_utils.query(query)
bindings = resp.json()['results']['bindings']
if len(bindings) == 0:
print("no match for", keyword.encode('ascii', 'xmlcharrefreplace'))
elif len(bindings) > 1:
print("multiple matches for", keyword.encode('ascii', 'xmlcharrefreplace'))
print(bindings)
return [binding['entity']['value'] for binding in bindings]
?descriptor anno:start ?d_start
; anno:end ?d_end
; anno:root/anno:pos ?pos
.
FILTER (?pos NOT IN ("X", "PUNCT"))
?target anno:category "diseases"
; anno:start ?t_start
; anno:end ?t_end
; ^dc:relation ?rel
.
?rel rdf:label "malaria" .
# The descriptor is outside of the target
FILTER ( ?d_end <= ?t_start || ?t_end <= ?d_start )
} LIMIT 100
"""
result = sparql_utils.query(query)
print_result(result)
print "Pathogens and the sentences they appear in"
query = prefixes+"""
SELECT ?phrase ?target
WHERE {
?phrase anno:start ?p_start
; anno:end ?p_end
; dep:ROOT ?noop
.
?target anno:category "pathogens"
; anno:start ?t_start
; anno:end ?t_end
.
?phrase anno:source_doc ?same_source .
?target anno:source_doc ?same_source .
max_items = int(args.max_items)
article_query_template = make_template("""
prefix con: <http://www.eha.io/types/content/>
prefix anno: <http://www.eha.io/types/annotation_prop/>
prefix eha: <http://www.eha.io/types/>
SELECT ?item_uri ?content
WHERE {
?item_uri con:text ?content
FILTER NOT EXISTS {
?item_uri anno:annotated_by eha:spacy_0
}
}
ORDER BY asc(?item_uri)
LIMIT 100
""")
items_processed = 0
while max_items < 0 or items_processed < max_items:
print("Items processed: ", str(items_processed))
result = sparql_utils.query(article_query_template.render())
bindings = result.json()['results']['bindings']
if len(bindings) == 0:
print("No more results")
break
else:
items_processed += len(bindings)
for binding in bindings:
item_uri = binding['item_uri']['value']
content = binding['content']['value']
print("Parsing ", item_uri)
create_annotations(item_uri, content)
?descriptor anno:start ?d_start
; anno:end ?d_end
; anno:root/anno:pos ?pos
.
FILTER (?pos NOT IN ("X", "PUNCT"))
?target anno:category "diseases"
; anno:start ?t_start
; anno:end ?t_end
; ^dc:relation ?rel
.
?rel rdf:label "malaria" .
# The descriptor is outside of the target
FILTER ( ?d_end <= ?t_start || ?t_end <= ?d_start )
} LIMIT 100
"""
result = sparql_utils.query(query)
print_result(result)
print "Pathogens and the sentences they appear in"
query = prefixes + """
SELECT ?phrase ?target
WHERE {
?phrase anno:start ?p_start
; anno:end ?p_end
; dep:ROOT ?noop
.
?target anno:category "pathogens"
; anno:start ?t_start
; anno:end ?t_end
.
?phrase anno:source_doc ?same_source .
?target anno:source_doc ?same_source .
query_template = make_template("""
prefix con: <http://www.eha.io/types/content/>
prefix anno: <http://www.eha.io/types/annotation_prop/>
prefix eha: <http://www.eha.io/types/>
SELECT ?item_uri ?content
WHERE {
?item_uri con:text ?content
# FILTER(strstarts(str(?item_uri), "http://t11.tater.io/documents/"))
FILTER NOT EXISTS {
?item_uri anno:annotated_by eha:annie_1
}
}
ORDER BY rand()
LIMIT 100
""")
items_processed = 0
while max_items < 0 or items_processed < max_items:
print("Items processed: ", str(items_processed))
result = sparql_utils.query(query_template.render())
bindings = result.json()['results']['bindings']
if len(bindings) == 0:
print("No more results")
break
else:
items_processed += len(bindings)
for binding in bindings:
item_uri = binding['item_uri']['value']
content = binding['content']['value']
print("Annotating ", item_uri)
create_annotations(item_uri, content)
query = prefixes+"""
SELECT ?p1 ?p2
WHERE {
?p1 anno:start ?p1start
; anno:end ?p1end
; anno:source_doc ?same_source
.
?dep_rel rdf:type anno:dependency_relation .
?parent ?dep_rel ?p1 .
?p2 anno:start ?p2start
; anno:end ?p2end
; anno:source_doc ?same_source
; anno:category "diseases"
.
FILTER ( ?p1start <= ?p2start && ?p1end >= ?p2end )
FILTER (?p1 != ?p2)
}
"""
resp = sparql_utils.query(query)
print("Finished in", datetime.datetime.now() - start)
start = datetime.datetime.now()
print("Testing query speed with containment predicate...")
query = prefixes+"""
SELECT ?p1 ?p2
WHERE {
?p1 anno:contains ?p2
}
"""
resp = sparql_utils.query(query)
print("Finished in", datetime.datetime.now() - start)
import requests
from templater import make_template
import sparql_utils
result = sparql_utils.query("""
prefix pro: <http://www.eha.io/types/promed/>
prefix eha: <http://www.eha.io/types/>
prefix anno: <http://www.eha.io/types/annotation_prop/>
SELECT
?annotator
(count(?article) AS ?articles)
WHERE {
?article pro:post ?post .
OPTIONAL {
?article anno:annotated_by ?annotator
}
}
GROUP BY ?annotator
""")
print result.content
query_template = make_template("""
prefix con: <http://www.eha.io/types/content/>
prefix anno: <http://www.eha.io/types/annotation_prop/>
prefix eha: <http://www.eha.io/types/>
SELECT ?item_uri ?content
WHERE {
?item_uri con:text ?content
# FILTER(strstarts(str(?item_uri), "http://t11.tater.io/documents/"))
FILTER NOT EXISTS {
?item_uri anno:annotated_by eha:annie_1
}
}
ORDER BY rand()
LIMIT 100
""")
items_processed = 0
while max_items < 0 or items_processed < max_items:
print("Items processed: ", str(items_processed))
result = sparql_utils.query(query_template.render())
bindings = result.json()['results']['bindings']
if len(bindings) == 0:
print("No more results")
break
else:
items_processed += len(bindings)
for binding in bindings:
item_uri = binding['item_uri']['value']
content = binding['content']['value']
print("Annotating ", item_uri)
create_annotations(item_uri, content)
query = prefixes + """
SELECT ?p1 ?p2
WHERE {
?p1 anno:start ?p1start
; anno:end ?p1end
; anno:source_doc ?same_source
.
?dep_rel rdf:type anno:dependency_relation .
?parent ?dep_rel ?p1 .
?p2 anno:start ?p2start
; anno:end ?p2end
; anno:source_doc ?same_source
; anno:category "diseases"
.
FILTER ( ?p1start <= ?p2start && ?p1end >= ?p2end )
FILTER (?p1 != ?p2)
}
"""
resp = sparql_utils.query(query)
print("Finished in", datetime.datetime.now() - start)
start = datetime.datetime.now()
print("Testing query speed with containment predicate...")
query = prefixes + """
SELECT ?p1 ?p2
WHERE {
?p1 anno:contains ?p2
}
"""
resp = sparql_utils.query(query)
print("Finished in", datetime.datetime.now() - start)