def history(self, app_id, mapped_entity, entity, projection, entity_id,
version):
""" gets an entity history """
query = Query(self.reference_date,
version=None,
session=self.session,
branch=self.branch)
query.set_query_context(app_id, mapped_entity, entity)
return query.history(mapped_entity, entity, projection, entity_id,
version)
def filter(self,
app_id,
mapped_entity,
entity,
projection,
page=None,
page_size=None):
""" execute a query on domain """
query = Query(self.reference_date,
self.version,
self.session,
branch=self.branch)
query.set_query_context(app_id, mapped_entity, entity)
return query.execute(projection, page=page, page_size=page_size)
def get_result(self, qr, method='Cosine', exp=False):
q = Query()
reader = CSV_Reader()
q_dict = q.split_query(qr)
all_d_dict = reader.get_dict("iif_weight.csv")
result = q.retrieve(q_dict, all_d_dict, method, exp)
if not result:
response = []
elif (result[0][1]) == 1 or m.isnan(result[0][1]):
response = []
else:
response = result
return response
def run(self):
self.queries = []
self.queries.append(Query(self.query_str, -1))
self.get_entity()
for q in self.queries:
if len(q.entities):
self.entities.extend(q.entities)
def run(self):
mentions = self.extract_mentions(self.comment)
for m in mentions:
self.queries.append(Query(m[0], m[1]))
# for q in self.queries:
# print (q.text+' %d' % q.index)
self.get_entity()
def run(self):
self.queries = []
if re.search(ur"[\u4e00-\u9fff]+", self.query_str): #Chinese
self.lan = "ch"
if len(self.query_str) > 5:
self.extract_zh_mentions(self.query_str)
else:
self.queries.append(Query(self.query_str, 0))
def split_querystr(self):
"""
Get sub query strings according to truncate the origin string
"""
ws = self.query_str.split()
for i in range(0, len(ws) - 2):
for j in range(2, len(ws)):
new_query = " ".join(ws[i:j])
self.queries.append(Query(new_query, -1))
def extract_mentions(self):
te = TermExtractor()
for t in te.get_terms(1, self.text):
try:
index = self.text.index(t)
q = Query(t.lower(), index)
self.queries.append(q)
except ValueError, e:
print "Substring:", t, "is not in string"
def extract_mentions(self):
"""
Extract terminology from query string
"""
te = TermExtractor()
terms = te.get_terms(1, self.query_str)
if len(terms) > 0:
for t in terms:
q = Query(t.lower(), -1)
self.queries.append(q)
print "%d term mentions" % len(self.queries)
def run(self):
self.queries = []
w_num = len(self.query_str.split())
if w_num < 3:
self.queries.append(Query(self.query_str, -1))
if w_num >= 5: # extract terminology
print "Extract terminology"
self.extract_mentions()
if self.no_entity(
) and w_num >= 3 and w_num <= 5: # split query string into short substring
print "Split Query string"
self.split_querystr()
if "-" in self.query_str:
for q in self.query_str.split():
if "-" in q:
self.queries.append(Query(q, -1))
break
self.get_entity()
for q in self.queries:
if len(q.entities):
self.entities.append(q.entities[0])
def extract_zh_mentions(self, s):
"""
Use jieba to extract mentions from Chinese string
"""
types = ['nt', 'nr', 'nz', 'ns']
seg_list = pseg.cut(s)
seg_index = []
last = 0
for seg in seg_list:
begin = s.index(seg.word, last)
last = begin + len(seg.word)
if seg.flag in types:
self.queries.append(Query(seg.word.encode('utf-8'), begin))
print "%d mentions" % len(self.queries)
def extract_en_mentions(self, s):
"""
Usr Standford Parser and NLTK to extract mentions from english string
"""
types = ['LOCATION', 'PERSON', 'ORGANIZATION']
if not self.en_parser:
self.en_parser = stanford_parser.Parser()
seg_list = self.en_parser.NER(s)
seg_index = []
begin = 0
last = 0
i = 0
if type(
seg_list[0]
) == list: #不同系统上的返回结果格式不一样,我们想要的是[('word','LOCATION'),('people','PERSON')],但有的是[[('word','LOCATION')]],这种情况要把里面的list提取出来
seg_list = seg_list[0]
while i < len(seg_list):
segs = []
word, tag = seg_list[i]
if tag in types:
j = i
# 属于指定类型，类型相同且连续的，合到一起
while j < len(seg_list) and seg_list[j][1] == tag:
segs.append(seg_list[j][0])
j += 1
i = j
query = " ".join(segs)
print "query:", query
try:
begin = s.index(query, last)
print "Can't find %s" % query
except:
continue
last = begin + len(query)
self.queries.append(Query(query, begin))
else:
#不属于指定类型，直接跳过
i += 1
#begin = s.index(word, last)
#last = begin + len(word)
print "%d mentions" % len(self.queries)
self.zh_parser = None
def run(self):
self.queries = []
if re.search(ur"[\u4e00-\u9fff]+", self.query_str): #Chinese
self.lan = "ch"
if len(self.query_str) > 5:
self.extract_zh_mentions(self.query_str)
else:
self.queries.append(Query(self.query_str, 0))
else: #English or others
if len(self.query_str.split()) > 3:
self.extract_en_mentions(self.query_str)
else:
self.queries.append(Query(self.query_str, 0))
self.get_entity()
def set_candb(self, db):
"""
Set Candidateset Mysql db
"""
self.candb = db
def set_graph(self, x):
"""
Set XLore db
"""
self.graph = x
from configuration.socket import SocketConfiguration
# TODO: class and def documentations
# TODO: write tests
app = Flask(__name__)
# version
version = 'v1'
config = SocketConfiguration(
cfg_file=os.path.join(os.path.dirname(__file__), 'lsapi.cfg'))
ls_accessor = Socket(config.connection_string, config.connection_type)
# init LS query class
ls_query = Query(ls_accessor)
@app.route('/%s/columns' % version, methods=['GET'])
def get_columns():
entity = 'columns'
columns = get_columns_from_get_parameter_or_use_defaults(
request.args, entity)
query_filter = get_filter_from_get_parameter(request.args)
with QueryTableCtx(ls_query, entity, query_filter, columns) as ls_ctx:
data, ls_return_code = ls_ctx.query()
with LivestatusActionCtx(data, ls_return_code) as task_ctx:
return task_ctx.return_table()
# COMMENTS