def test_search_query_compile(self):
f = DynamicDocument.fields
sq = SearchQuery()
self.assert_expression(sq, {})
self.assertEqual(collect_doc_classes(sq), set())
sq = SearchQuery(Term(f.user, 'kimchy')).limit(10).offset(0)
self.assert_expression(
sq,
{
"from": 0,
"size": 10,
"query": {
"term": {"user": "******"}
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = SearchQuery(Term(f.user, 'kimchy')).filter(f.age >= 16)
self.assert_expression(
sq,
{
"query": {
"filtered": {
"query": {
"term": {"user": "******"}
},
"filter": {
"range": {
"age": {"gte": 16}
}
}
}
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = SearchQuery(Term(f.user, 'kimchy'), _compiler=QueryCompiled20).filter(f.age >= 16)
self.assert_expression(
sq,
{
"query": {
"bool": {
"must": {
"term": {"user": "******"}
},
"filter": {
"range": {
"age": {"gte": 16}
}
}
}
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery(Term(f.user, 'kimchy'))
.query(f.user != 'kimchy')
)
self.assert_expression(
sq,
{
"query": {
"bool": {
"must_not": [
{
"term": {"user": "******"}
}
]
}
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery(Term(f.user, 'kimchy'))
.query(None)
)
self.assert_expression(sq, {})
self.assertEqual(collect_doc_classes(sq), set())
sq = (
SearchQuery(Term(f.user, 'kimchy'))
.filter(f.age >= 16)
.filter(f.lang == 'English')
)
self.assert_expression(
sq,
{
"query": {
"filtered": {
"query": {
"term": {"user": "******"}
},
"filter": {
"bool": {
"must": [
{
"range": {
"age": {"gte": 16}
}
},
{
"term": {
"lang": "English"
}
}
]
}
}
}
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery()
.order_by(
f.opinion_rating.desc(missing='_last'),
f.opinion_count.desc(),
f.id
)
)
self.assert_expression(
sq,
{
"sort": [
{
"opinion_rating": {
"order": "desc",
"missing": "_last"
}
},
{
"opinion_count": "desc"
},
"id"
]
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery()
.order_by(
f.opinion_rating.desc(missing='_last'),
f.opinion_count.desc(),
f.id
)
.order_by(None)
.order_by(None)
)
self.assert_expression(sq, {})
self.assertEqual(collect_doc_classes(sq), set())
sq = SearchQuery().source(f.name, f.company)
self.assert_expression(
sq,
{
"_source": ["name", "company"]
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = SearchQuery().source(exclude=[f.name, f.company])
self.assert_expression(
sq,
{
"_source": {
"exclude": ["name", "company"]
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery()
.source(
include=[f.obj1.wildcard('*'), f.obj2.wildcard('*')],
# FIXME: f.wildcard('*')
exclude=DynamicDocument.wildcard('*').description
)
)
self.assert_expression(
sq,
{
"_source": {
"include": ["obj1.*", "obj2.*"],
"exclude": "*.description"
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery()
.source(None)
.source(f.name, f.company)
.source(None)
)
self.assert_expression(sq, {})
self.assertEqual(collect_doc_classes(sq), set())
sq = (
SearchQuery()
.source(f.name, f.company)
.source(False)
)
self.assert_expression(
sq,
{
"_source": False
}
)
self.assertEqual(collect_doc_classes(sq), set())
sq = (
SearchQuery()
.source(True)
)
self.assert_expression(
sq,
{
"_source": True
}
)
self.assertEqual(collect_doc_classes(sq), set())
sq = SearchQuery().fields(f.name, f.company)
self.assert_expression(
sq,
{
"fields": ["name", "company"]
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery()
.fields(True)
)
self.assert_expression(
sq,
{
"fields": '*'
}
)
self.assertEqual(collect_doc_classes(sq), set())
sq = (
SearchQuery()
.fields(None)
.fields(f.name, f.company)
.fields(None)
)
self.assert_expression(sq, {})
self.assertEqual(collect_doc_classes(sq), set())
sq = (
SearchQuery()
.fields(f.name, f.company)
.fields(False)
)
self.assert_expression(
sq,
{
"fields": []
}
)
self.assertEqual(collect_doc_classes(sq), set())
self.assert_expression(
SearchQuery()
.function_score({'random_score': {"seed": 1234}}),
{
"query": {
"function_score": {
"functions": [
{
"random_score": {"seed": 1234}
}
],
}
}
}
)
sq = (
SearchQuery(MultiMatch('Iphone 6', fields=[f.name, f.description]))
.filter(f.status == 0)
.function_score(None)
.function_score({'_score': {"seed": 1234}})
.function_score(None)
.function_score({'field_value_factor': {'field': f.popularity,
'factor': 1.2,
'modifier': 'sqrt'}},
boost_mode='sum')
.function_score({'boost_factor': 3,
'filter': f.region == 12})
)
self.assert_expression(
sq,
{
"query": {
"filtered": {
"query": {
"function_score": {
"query": {
"multi_match": {
"query": "Iphone 6",
"fields": ["name", "description"]
}
},
"functions": [
{
"field_value_factor": {
"field": "popularity",
"factor": 1.2,
"modifier": "sqrt"
}
},
{
"filter": {
"term": {"region": 12}
},
"boost_factor": 3
}
],
"boost_mode": "sum"
}
},
"filter": {
"term": {"status": 0}
}
}
}
}
)
self.assertEqual(collect_doc_classes(sq), {DynamicDocument})
sq = (
SearchQuery()
.filter(f.status == 0)
.boost_score(
{'filter': f.discount_percent >= 10, 'weight': 1000},
{'filter': f.discount_percent >= 50, 'weight': 2000},
{'filter': f.presence == 'available', 'weight': 10000},
)
)
self.assert_expression(
sq,
{
"query": {
"filtered": {
"query": {
"function_score": {
"functions": [
{
"filter": {"range": {"discount_percent": {"gte": 10}}},
"weight": 1000
},
{
"filter": {"range": {"discount_percent": {"gte": 50}}},
"weight": 2000
},
{
"filter": {"term": {"presence": "available"}},
"weight": 10000
},
],
"score_mode": "sum",
"boost_mode": "sum"
}
},
"filter": {
"term": {"status": 0}
}
}
}
}
)
sq = (
SearchQuery(f.name.match('test'))
.filter(f.status == 0)
.function_score(
{'field_value_factor': {'field': f.popularity}},
)
.boost_score(
{'filter': f.discount_percent >= 10, 'weight': 100},
)
.boost_score(None)
.boost_score(
{'filter': f.discount_percent >= 10, 'weight': 1000},
{'filter': f.discount_percent >= 50, 'weight': 2000},
score_mode='max',
)
)
self.assert_expression(
sq,
{
"query": {
"filtered": {
"query": {
"function_score": {
"query": {
"function_score": {
"query": {
"match": {
"name": "test"
}
},
"functions": [
{
"field_value_factor": {
"field": "popularity"
}
}
]
}
},
"functions": [
{
"filter": {"range": {"discount_percent": {"gte": 10}}},
"weight": 1000
},
{
"filter": {"range": {"discount_percent": {"gte": 50}}},
"weight": 2000
},
],
"score_mode": "max",
"boost_mode": "sum"
}
},
"filter": {
"term": {"status": 0}
}
}
}
}
)
sq = (
SearchQuery()
.rescore(
QueryRescorer(
self.index.t.field1.match('the quick brown', type='phrase', slop=2)
)
)
.rescore(None)
.rescore(
QueryRescorer(
self.index.t.field1.match('the quick brown fox', type='phrase', slop=2),
query_weight=0.7,
rescore_query_weight=1.2
),
window_size=100,
)
.rescore(
QueryRescorer(
FunctionScore(script_score={'script': "log10(doc['numeric'].value + 2)"}),
score_mode='multiply'
),
window_size=10,
)
)
self.assert_expression(
sq,
{
"rescore": [
{
"window_size": 100,
"query": {
"rescore_query": {
"match": {
"field1": {
"query": "the quick brown fox",
"type": "phrase",
"slop": 2
}
}
},
"query_weight": 0.7,
"rescore_query_weight": 1.2
}
},
{
"window_size": 10,
"query": {
"score_mode": "multiply",
"rescore_query": {
"function_score": {
"script_score": {
"script": "log10(doc['numeric'].value + 2)"
}
}
}
}
}
]
}
)
self.assertEqual(collect_doc_classes(sq), {self.index.t})
sq = SearchQuery().post_filter(self.index.shirt.color == 'red')
self.assert_expression(
sq,
{
"post_filter": {
"term": {"color": "red"}
}
}
)
self.assertEqual(collect_doc_classes(sq), {self.index.shirt})
sq = (
SearchQuery()
.filter(self.index.shirt.brand == 'gucci')
.post_filter(self.index.shirt.color == 'red')
.post_filter(self.index.shirt.model == 't-shirt')
)
self.assert_expression(
sq,
{
"query": {
"filtered": {
"filter": {
"term": {"brand": "gucci"}
}
}
},
"post_filter": {
"bool": {
"must": [
{"term": {"color": "red"}},
{"term": {"model": "t-shirt"}}
]
}
}
}
)
self.assertEqual(collect_doc_classes(sq), {self.index.shirt})
def test_expression(self):
f = DynamicDocument.fields
e = Params({'foo': 'bar'})
self.assert_expression(
e,
{"foo": "bar"}
)
self.assertEqual(e['foo'], 'bar')
self.assertTrue('foo' in e)
self.assert_expression(
Match(f.message, 'this is a test'),
{
"match": {
"message": "this is a test",
}
}
)
self.assert_expression(
Match(
f.message, 'this is a test',
minimum_should_match='100%',
cutoff_frequency=0.001,
boost=2.1
),
{
"match": {
"message": {
"query": "this is a test",
"minimum_should_match": "100%",
"cutoff_frequency": 0.001,
"boost": 2.1,
}
}
}
)
self.assert_expression(
Term(f.user, 'kimchy'),
{
"term": {"user": "******"}
}
)
self.assert_expression(
Term(f.user, 'kimchy', boost=1.2),
{
"term": {"user": {"value": "kimchy", "boost": 1.2}}
}
)
self.assert_expression(
Term('user.login', 'kimchy'),
{
"term": {"user.login": "******"}
}
)
self.assert_expression(
Terms(f.status, [0]),
{
"terms": {
"status": [0]
}
}
)
self.assert_expression(
Terms(f.tags, ['blue', 'pill'], minimum_should_match=1),
{
"terms": {
"tags": ["blue", "pill"],
"minimum_should_match": 1
}
}
)
self.assert_expression(
Exists(f.tags),
{
"exists": {"field": "tags"}
}
)
self.assert_expression(
Missing(f.tags, _cache=True),
{
"missing": {
"field": "tags",
"_cache": True
}
}
)
self.assert_expression(
Bool(
must=Term(f.user, 'kimchy'),
filter=Term(f.tag, 'tech'),
must_not=Range(f.age, from_=10, to=20),
should=[Term(f.tag, 'wow'), Term(f.tag, 'elasticsearch', boost=2.1)],
minimum_should_match=1,
boost=1.0,
),
{
"bool": {
"must": {
"term": {"user": "******"}
},
"filter": {
"term": {"tag": "tech"}
},
"must_not": {
"range": {
"age": {"from": 10, "to": 20}
}
},
"should": [
{
"term": {"tag": "wow"}
},
{
"term": {"tag": {"value": "elasticsearch", "boost": 2.1}}
}
],
"minimum_should_match": 1,
"boost": 1.0
}
}
)
e = MultiMatch(
"Will Smith",
[self.index.star.title.boost(4), self.index.star.wildcard('*_name').boost(2)],
minimum_should_match='100%'
)
self.assert_expression(
e,
{
"multi_match": {
"query": "Will Smith",
"fields": ["title^4", "*_name^2"],
"minimum_should_match": "100%"
}
}
)
self.assertEqual(
e._collect_doc_classes(),
{self.index.star}
)
self.assert_expression(
Range(self.index.product.price, lte=100, boost=2.2, execution='index', _cache=False),
{
"range": {
"price": {"lte": 100, "boost": 2.2},
"execution": "index",
"_cache": False
}
}
)
self.assert_expression(
Boosting(
positive=Term(f.field1, 'value1'),
negative=Term(f.field2, 'value2'),
negative_boost=0.2
),
{
"boosting": {
"positive": {
"term": {
"field1": "value1"
}
},
"negative": {
"term": {
"field2": "value2"
}
},
"negative_boost": 0.2
}
}
)
self.assert_expression(
Common(
f.body, 'nelly the elephant not as a cartoon',
cutoff_frequency=0.001,
minimum_should_match=dict(low_freq=2, high_freq=3),
),
{
"common": {
"body": {
"query": "nelly the elephant not as a cartoon",
"cutoff_frequency": 0.001,
"minimum_should_match": {
"low_freq": 2,
"high_freq": 3
}
}
}
}
)
self.assert_expression(
ConstantScore(filter=Term(f.user, 'kimchy'), boost=1.2),
{
"constant_score": {
"filter": {
"term": { "user": "******"}
},
"boost": 1.2
}
}
)
self.assert_expression(
FunctionScore(
query=MatchAll(),
field_value_factor={
'field': f.popularity,
'factor': 1.2,
'modifier': 'sqrt',
}
),
{
"function_score": {
"query": {"match_all": {}},
"field_value_factor": {
"field": "popularity",
"factor": 1.2,
"modifier": "sqrt"
}
}
}
)
self.assert_expression(
DisMax([Term(f.age, 34), Term(f.age, 35)], boost=1.2, tie_breaker=0.7),
{
"dis_max": {
"tie_breaker": 0.7,
"boost": 1.2,
"queries": [
{
"term" : { "age" : 34 }
},
{
"term" : { "age" : 35 }
}
]
}
}
)
self.assert_expression(
Filtered(
filter=Range(f.created, gte='now - 1d / d'),
query=Match(f.tweet, 'full text search')
),
{
"filtered": {
"query": {
"match": { "tweet": "full text search" }
},
"filter": {
"range": { "created": { "gte": "now - 1d / d" }}
}
}
}
)
self.assert_expression(
Ids(['123456']),
{
"ids": {
"values": ["123456"]
}
}
)
self.assert_expression(
Ids(['1', '4', '100'], type="my_type"),
{
"ids": {
"type": "my_type",
"values": ["1", "4", "100"]
}
}
)
self.assert_expression(
Prefix(f.user, 'ki', boost=2.0),
{
"prefix": { "user": { "value": "ki", "boost": 2.0 } }
}
)
self.assert_expression(
MatchAll(),
{"match_all": {}}
)
self.assert_expression(
MatchAll(boost=1.2),
{
"match_all": { "boost" : 1.2 }
}
)
self.assert_expression(
Query(Match(f.title, 'this that thus')),
{
"query": {
"match": {
"title": "this that thus"
}
}
}
)
self.assert_expression(
Query(Match(f.title, 'this that thus'), _cache=True),
{
"fquery": {
"query": {
"match": {
"title": "this that thus"
}
},
"_cache": True
}
}
)
self.assertRaises(NotImplementedError, BooleanExpression)
self.assert_expression(
And(
Range(f.post_date, from_='2010-03-01', to='2010-04-01'),
Prefix(f.name.second, 'ba')
),
{
"and": [
{
"range": {
"post_date": {
"from": "2010-03-01",
"to": "2010-04-01"
}
}
},
{
"prefix" : { "name.second" : "ba" }
}
]
}
)
self.assert_expression(
And(
Range(f.post_date, from_='2010-03-01', to='2010-04-01'),
Prefix(f.name.second, 'ba'),
_cache=True
),
{
"and": {
"filters": [
{
"range": {
"post_date": {
"from": "2010-03-01",
"to": "2010-04-01"
}
}
},
{
"prefix" : { "name.second" : "ba" }
}
],
"_cache": True
}
}
)
self.assert_expression(
Or(Term(f.name.second, 'banon'), Term(f.name.nick, 'kimchy')),
{
"or": [
{
"term": {"name.second": "banon"}
},
{
"term": {"name.nick": "kimchy"}
}
]
}
)
self.assert_expression(
And(Or(Term(f.name.nick, 'kimchy'))),
{
"term": {"name.nick": "kimchy"}
}
)
self.assert_expression(
Not(
Range(f.post_date, from_='2010-03-01', to='2010-04-01'),
),
{
"not": {
"range": {
"post_date": {
"from": "2010-03-01",
"to": "2010-04-01"
}
}
}
}
)
self.assert_expression(
Not(
Range(f.post_date, from_='2010-03-01', to='2010-04-01'),
_cache=True,
),
{
"not": {
"filter": {
"range": {
"post_date": {
"from": "2010-03-01",
"to": "2010-04-01"
}
}
},
"_cache": True
}
}
)
self.assert_expression(
Sort(f.post_date),
"post_date"
)
self.assert_expression(
Sort(f.age, 'desc'),
{
"age": "desc"
}
)
self.assert_expression(
Sort(f.price, 'asc', mode='avg'),
{
"price": {
"order": "asc",
"mode": "avg"
}
}
)
self.assert_expression(
Sort(
f.offer.price.sort, 'asc', mode='avg',
nested_filter=Term(f.offer.color, 'blue')
),
{
"offer.price.sort": {
"order": "asc",
"mode": "avg",
"nested_filter": {
"term": {"offer.color": "blue"}
}
}
}
)
self.assert_expression(
SpanFirst(SpanTerm(f.user, 'kimchy'), end=3),
{
"span_first": {
"match": {
"span_term": {"user": "******"}
},
"end": 3
}
}
)
self.assert_expression(
SpanMulti(Prefix(f.user, 'ki', boost=1.08)),
{
"span_multi": {
"match": {
"prefix": {
"user": {"value": "ki", "boost": 1.08}
}
}
}
}
)
self.assert_expression(
SpanNear(
[SpanTerm(f.field, 'value1'),
SpanTerm(f.field, 'value2'),
SpanTerm(f.field, 'value3')],
slop=12,
in_order=False,
collect_payloads=False,
),
{
"span_near": {
"clauses": [
{"span_term": {"field": "value1"}},
{"span_term": {"field": "value2"}},
{"span_term": {"field": "value3"}}
],
"slop": 12,
"in_order": False,
"collect_payloads": False
}
}
)
self.assert_expression(
SpanNot(
SpanTerm(f.field1, 'hoya'),
SpanNear([SpanTerm(f.field1, 'la'), SpanTerm(f.field1, 'hoya')], slop=0, in_order=True),
),
{
"span_not": {
"include": {
"span_term": {"field1": "hoya"}
},
"exclude": {
"span_near": {
"clauses": [
{"span_term": {"field1": "la"}},
{"span_term": {"field1": "hoya"}}
],
"slop": 0,
"in_order": True
}
}
}
}
)
self.assert_expression(
SpanOr(
[
SpanTerm(f.field, 'value1'),
SpanTerm(f.field, 'value2'),
SpanTerm(f.field, 'value3')
],
boost=2,
),
{
"span_or": {
"clauses": [
{"span_term": {"field": "value1"}},
{"span_term": {"field": "value2"}},
{"span_term": {"field": "value3"}}
],
"boost": 2
}
}
)
self.assert_expression(
Limit(1000),
{
"limit": {
"value": 1000
}
}
)
e = Nested(
self.index.movie.stars,
Match(self.index.movie.stars.full_name, 'Will Smith'),
score_mode='max',
)
self.assert_expression(
e,
{
"nested": {
"path": "stars",
"query": {
"match": {
"stars.full_name": "Will Smith"
}
},
"score_mode": "max"
}
}
)
self.assertEqual(
e._collect_doc_classes(),
{self.index.movie}
)
e = HasParent(
self.index.blog.tag == 'something',
parent_type=self.index.blog,
score_mode='score',
)
self.assert_expression(
e,
{
"has_parent": {
"parent_type": "blog",
"query": {
"term": {
"tag": "something"
}
},
"score_mode": "score"
}
}
)
self.assertEqual(
e._collect_doc_classes(),
set()
)
e = HasParent(
self.index.blog.tag == 'something',
score_mode='score',
)
self.assert_expression(
e,
{
"has_parent": {
"parent_type": "blog",
"query": {
"term": {
"tag": "something"
}
},
"score_mode": "score"
}
}
)
self.assertEqual(
e._collect_doc_classes(),
set()
)
e = HasChild(
self.index.blog_tag.tag == 'something',
type=self.index.blog_tag,
score_mode='sum',
)
self.assert_expression(
e,
{
"has_child": {
"type": "blog_tag",
"query": {
"term": {
"tag": "something"
}
},
"score_mode": "sum"
}
}
)
self.assertEqual(
e._collect_doc_classes(),
set()
)
e = HasChild(
self.index.blog_tag.tag == 'something',
score_mode='sum',
)
self.assert_expression(
e,
{
"has_child": {
"type": "blog_tag",
"query": {
"term": {
"tag": "something"
}
},
"score_mode": "sum"
}
}
)
self.assertEqual(
e._collect_doc_classes(),
set()
)