def get_count(self):
try:
return self.es.count(self.get_query(), self.get_filters())
except queryparser.QueryParseError:
# try queries one by one
for i, q in enumerate(self.get_queries()):
queryparser.parse_to_terms(q.query, context=(q.declared_label or i+1))
# if error wasn't raised yet, re-raise original
raise
def get_count(self):
try:
return self.es.count(self.get_query(), self.get_filters())
except queryparser.QueryParseError:
# try queries one by one
for i, q in enumerate(self.get_queries()):
queryparser.parse_to_terms(q.query, context=(q.declared_label or i+1))
# if error wasn't raised yet, re-raise original
raise
def filter_query(self, query: str):
"""
Filter a queryset using a querystring. This is different from query(), as it will
not be able to generate highlights or scores. It is quicker, however.
"""
query_term = queryparser.parse_to_terms(query)
return self._copy(filters=self.filters + (query_term.get_dsl(),))
def filter_query(self, query: str):
"""
Filter a queryset using a querystring. This is different from query(), as it will
not be able to generate highlights or scores. It is quicker, however.
"""
query_term = queryparser.parse_to_terms(query)
return self._copy(filters=self.filters + (query_term.get_dsl(),))
def highlight_article(self, aid, query):
query = queryparser.parse_to_terms(query).get_dsl()
highlight_opts = {
"highlight_query": query,
"number_of_fragments": 0
}
body = {
'filter': build_filter(ids=aid),
'highlight': {
"fields": {
"text": highlight_opts,
"headline": highlight_opts,
"byline": highlight_opts
}
}
}
r = self.search(body, fields=[])
try:
hl = r['hits']['hits'][0]['highlight']
return {f: hl[f][0] for f in hl}
except KeyError:
log.exception("Could not get highlights from {r!r}".format(**locals()))
def query(self, query: str) -> "ESQuerySet":
"""
Use a query string to query the final result. If you do not need scoring, use filter_query.
@param query: query in Lucene syntax
"""
query_dsl = queryparser.parse_to_terms(query)
return self._copy(_query=query_dsl)
def query(self, query: str) -> "ESQuerySet":
"""
Use a query string to query the final result. If you do not need scoring, use filter_query.
@param query: query in Lucene syntax
"""
query_dsl = queryparser.parse_to_terms(query)
return self._copy(_query=query_dsl)
def build_body(query=None, filters={}, query_as_filter=False):
"""
Construct the query body from the query and/or filter(s)
(call with dict(build_body)
@param query: a elastic query string (i.e. lucene syntax, e.g. 'piet AND (ja* OR klaas)')
@param filter: field filter DSL query dict, defaults to build_filter(**filters)
@param query_as_filter: if True, use the query as a filter (faster but not score/relevance)
"""
filters = list(get_filter_clauses(**filters))
if query:
terms = queryparser.parse_to_terms(query)
if query_as_filter:
filters.append(terms.get_filter_dsl())
else:
yield ('query', terms.get_dsl())
if filters:
yield ('filter', combine_filters(filters))
def build_body(query=None, filters=EMPTY_RO_DICT, query_as_filter=False):
"""
Construct the query body from the query and/or filter(s)
(call with dict(build_body)
@param query: an elastic query string (i.e. lucene syntax, e.g. 'piet AND (ja* OR klaas)')
@param filters: field filter DSL query dict, defaults to build_filter(**filters)
@param query_as_filter: if True, use the query as a filter (faster but not score/relevance)
"""
filters = list(get_filter_clauses(**filters))
if query:
terms = queryparser.parse_to_terms(query)
if query_as_filter:
filters.append(terms.get_filter_dsl())
else:
yield ('query', terms.get_dsl())
if filters:
yield ('filter', combine_filters(filters))
def test_parse(self):
q = lambda s: str(parse_to_terms(s))
self.assertEqual(q('a'), '_all::a')
self.assertEqual(q('a b'), 'OR[_all::a _all::b]')
self.assertEqual(q('a OR b'), 'OR[_all::a _all::b]')
self.assertEqual(q('a AND b'), 'AND[_all::a _all::b]')
self.assertEqual(q('a NOT b'), 'NOT[_all::a _all::b]')
self.assertEqual(q('a AND (b c)'), 'AND[_all::a OR[_all::b _all::c]]')
self.assertEqual(q('(a AND b) c'), 'OR[AND[_all::a _all::b] _all::c]')
# starting with NOT
self.assertEqual(q('NOT a'), 'NOT[_all::a]')
# quotes
self.assertEqual(q('"a b"'), '_all::QUOTE[a b]')
self.assertEqual(q('"a b" c'), 'OR[_all::QUOTE[a b] _all::c]')
# proximity queries
self.assertEqual(q('a W/10 b'), '_all::PROX/10[a b]')
self.assertEqual(q('(a b) W/10 c'), '_all::PROX/10[OR[a b] c]')
self.assertEqual(q('(x:a b) W/10 c'), 'x::PROX/10[OR[a b] c]')
# proximity queries must have unique field and can only contain wildcards and disjunctions
self.assertRaises(QueryParseError, q, 'a W/10 (b OR (c OR d))')
self.assertRaises(QueryParseError, q, 'x:a W/10 y:b')
self.assertRaises(QueryParseError, q, 'a W/10 (b AND c)')
self.assertRaises(QueryParseError, q, 'a W/10 (b W/5 c)')
# lucene notation
self.assertEqual(q('"a b"~5'), '_all::PROX/5[a b]')
self.assertEqual(q('x:"a b"~5'), 'x::PROX/5[a b]')
self.assertEqual(q('"a (b c)"~5'), '_all::PROX/5[a OR[b c]]')
# disallow AND in lucene notation
self.assertRaises(QueryParseError, q, '"a (b AND c)"~5')
# disallow matchall wildcards anywhere but at the beginning
self.assertRaises(QueryParseError, q, '"a * b"~5')
self.assertRaises(QueryParseError, q, 'a AND *')
self.assertEqual(q('* NOT a'), 'NOT[_all::* _all::a]')
self.assertEqual(q('*'), '_all::*')
def test_parse(self):
q = lambda s: str(parse_to_terms(s))
self.assertEqual(q('a'), '_all::a')
self.assertEqual(q('a b'), 'OR[_all::a _all::b]')
self.assertEqual(q('a OR b'), 'OR[_all::a _all::b]')
self.assertEqual(q('a AND b'), 'AND[_all::a _all::b]')
self.assertEqual(q('a NOT b'), 'NOT[_all::a _all::b]')
self.assertEqual(q('a AND (b c)'), 'AND[_all::a OR[_all::b _all::c]]')
self.assertEqual(q('(a AND b) c'), 'OR[AND[_all::a _all::b] _all::c]')
# starting with NOT
self.assertEqual(q('NOT a'), 'NOT[_all::a]')
# quotes
self.assertEqual(q('"a b"'), '_all::QUOTE[a b]')
self.assertEqual(q('"a b" c'), 'OR[_all::QUOTE[a b] _all::c]')
# proximity queries
self.assertEqual(q('a W/10 b'), '_all::PROX/10[a b]')
self.assertEqual(q('(a b) W/10 c'), '_all::PROX/10[OR[a b] c]')
self.assertEqual(q('(x:a b) W/10 c'), 'x::PROX/10[OR[a b] c]')
# proximity queries must have unique field and can only contain wildcards and disjunctions
self.assertRaises(QueryParseError, q, 'a W/10 (b OR (c OR d))')
self.assertRaises(QueryParseError, q, 'x:a W/10 y:b')
self.assertRaises(QueryParseError, q, 'a W/10 (b AND c)')
self.assertRaises(QueryParseError, q, 'a W/10 (b W/5 c)')
# lucene notation
self.assertEqual(q('"a b"~5'), '_all::PROX/5[a b]')
self.assertEqual(q('x:"a b"~5'), 'x::PROX/5[a b]')
self.assertEqual(q('"a (b c)"~5'), '_all::PROX/5[a OR[b c]]')
# disallow AND in lucene notation
self.assertRaises(QueryParseError, q, '"a (b AND c)"~5')
# disallow matchall wildcards anywhere but at the beginning
self.assertRaises(QueryParseError, q, '"a * b"~5')
self.assertRaises(QueryParseError, q, 'a AND *')
self.assertEqual(q('* NOT a'), 'NOT[_all::* _all::a]')
self.assertEqual(q('*'), '_all::*')
def build_body(query=None, filters=EMPTY_RO_DICT, query_as_filter=False):
"""
Construct the query body from the query and/or filter(s)
(call with dict(build_body)
@param query: an elastic query string (i.e. lucene syntax, e.g. 'piet AND (ja* OR klaas)')
@param filters: field filter DSL query dict, defaults to build_filter(**filters)
@param query_as_filter: if True, use the query as a filter (faster but not score/relevance)
"""
query_body = {"bool": {}}
filters = list(get_filter_clauses(**filters))
if query:
terms = queryparser.parse_to_terms(query)
if query_as_filter:
filters.append(terms.get_filter_dsl())
else:
query_body["bool"]["must"] = terms.get_dsl()
if filters:
query_body["bool"]["filter"] = filters
return {"query": query_body}
def __init__(self, fields: Sequence[str], query: str, mark: str):
self.query_raw = query
self.query = queryparser.parse_to_terms(query) # type: Term
self.fields = fields
self.mark = mark
def __init__(self, fields: Sequence[str], query: str, mark: str):
self.query_raw = query
self.query = queryparser.parse_to_terms(query) # type: Term
self.fields = fields
self.mark = mark