def get_agent_from_refs(db_refs):
"""Get an agent given its db_refs."""
name = get_standard_name(db_refs)
if not name:
name = db_refs.get('TEXT')
if name and db_refs:
return Agent(name, db_refs=db_refs)
def get_raw_statement_text_grounding_counts(stmts):
texts = []
ev_text_for_agent_text = {}
for stmt in stmts:
for agent in stmt.agent_list():
if agent is None:
continue
if stmt.evidence[0].source_api == 'eidos':
txt = agent.db_refs['TEXT_NORM']
else:
txt = agent.db_refs['TEXT']
ev_text_for_agent_text[txt] = (stmt.evidence[0].pmid,
stmt.evidence[0].text)
assert txt, agent.db_refs
gr = agent.get_grounding()
standard_name = get_standard_name(*gr) if gr[0] else ''
url = get_identifiers_url(*gr) if gr[0] is not None else ''
gilda_grounding = gilda.ground(txt, context=stmt.evidence[0].text)
gilda_grounding = '%s:%s' % (gilda_grounding[0].term.db,
gilda_grounding[0].term.id) \
if gilda_grounding else ''
texts.append((txt, ('%s:%s' % gr) if gr[0] else '', standard_name,
url, gilda_grounding))
cnt = Counter(texts)
return cnt, ev_text_for_agent_text
def get_eidos_gilda_grounding_counts(stmts):
"""Return normalized text counts (name in case of Eidos concepts)
and evidence texts corresponding to each agent text."""
texts = []
ev_text_for_agent_text = {}
for stmt in stmts:
for agent in stmt.agent_list():
txt = agent.name
matches = gilda.ground(txt)
if matches:
gr = matches[0].term.db, matches[0].term.id
else:
gr = None, None
standard_name = get_standard_name(*gr) \
if gr[0] is not None else ''
url = get_identifiers_url(*gr) if gr[0] is not None else ''
ev_text_for_agent_text[txt] = (stmt.evidence[0].pmid,
stmt.evidence[0].text)
texts.append(
(txt, ('%s:%s' % gr) if gr[0] else '', standard_name, url, ''))
# Count the unique text-grounding entries
cnt = Counter(texts)
return cnt, ev_text_for_agent_text
db_refs[db_ns] = db_id
db_refs = standardize_db_refs(db_refs)
return db_refs
def get_prioritized_db_refs_key(db_refs):
for db_ns in default_ns_order:
db_id = db_refs.get(db_ns)
if db_id:
return db_ns, db_id
return None, None
def get_unique_prioritized_keys(map_name=default_map_name):
valid_element_refs = get_all_valid_element_refs(map_name)
db_refs = [indra_db_refs_from_minerva_refs(refs)
for refs in valid_element_refs]
prioritized_keys = [get_prioritized_db_refs_key(db_ref)
for db_ref in db_refs]
unique_prioritized_keys = {key for key in prioritized_keys
if key[0] is not None}
return unique_prioritized_keys
if __name__ == '__main__':
keys = get_unique_prioritized_keys(default_map_name)
with open('minerva_disease_map_indra_ids.csv', 'w') as fh:
for db_ns, db_id in sorted(keys):
name = get_standard_name({db_ns: db_id})
fh.write('%s,%s,%s\n' % (db_ns, db_id, name))