def get_matches(diagram1, diagram2, threshold_score=3.0, consider_context=True):
# initVM for lucene
lucene.initVM()
# index creation
p1 = get_preprocessed_nodes(diagram1, consider_context)
p2 = get_preprocessed_nodes(diagram2, consider_context)
docs_from_p1 = transform_to_index(p1)
docs_from_p2 = transform_to_index(p2)
# set up in memory store for indexes
directory1 = store.RAMDirectory()
directory2 = store.RAMDirectory()
analyzer = analysis.standard.StandardAnalyzer()
# create separate indexes for each diagram
create_index(directory=directory1, analyzer=analyzer, documents_to_index=docs_from_p1)
create_index(directory=directory2, analyzer=analyzer, documents_to_index=docs_from_p2)
# SEARCH STEP: use indexes from one store to search on indexes from other store
res_1 = search_index(directory_to_search=directory2, entities_to_check=p1, analyzer=analyzer)
res_2 = search_index(directory_to_search=directory1, entities_to_check=p2, analyzer=analyzer)
# apply match search pruning rules
set_m = prune_results(res_1, res_2, threshold_score)
directory1.close()
directory2.close()
return set_m
def load(cls, directory, analyzer=None):
"""Open `IndexSearcher`_ with a lucene RAMDirectory, loading index into memory."""
with closing.store(directory) as directory:
directory = store.RAMDirectory(directory, store.IOContext.DEFAULT)
self = cls(directory, analyzer)
self.shared.add(self.directory)
return self
def directory(self, directory):
if directory is None:
directory = store.RAMDirectory()
self.add(directory)
elif isinstance(directory, string_types):
directory = store.FSDirectory.open(File(directory).toPath())
self.add(directory)
return directory
def test_grouping(tempdir, indexer, zipcodes):
field = indexer.fields['location'] = engine.NestedField(
'state.county.city', docValuesType='sorted')
for doc in zipcodes:
if doc['state'] in ('CA', 'AK', 'WY', 'PR'):
lat, lng = ('{0:08.3f}'.format(doc.pop(l))
for l in ['latitude', 'longitude'])
location = '.'.join(doc[name]
for name in ['state', 'county', 'city'])
indexer.add(doc, latitude=lat, longitude=lng, location=location)
indexer.commit()
states = list(indexer.terms('state'))
assert states[0] == 'AK' and states[-1] == 'WY'
counties = [
term.split('.')[-1] for term in indexer.terms('state.county', 'CA')
]
hits = indexer.search(field.prefix('CA'))
assert sorted({hit['county'] for hit in hits}) == counties
assert counties[0] == 'Alameda' and counties[-1] == 'Yuba'
cities = [
term.split('.')[-1]
for term in indexer.terms('state.county.city', 'CA.Los Angeles')
]
hits = indexer.search(field.prefix('CA.Los Angeles'))
assert sorted({hit['city'] for hit in hits}) == cities
assert cities[0] == 'Acton' and cities[-1] == 'Woodland Hills'
(hit, ) = indexer.search('zipcode:90210')
assert hit['state'] == 'CA' and hit['county'] == 'Los Angeles' and hit[
'city'] == 'Beverly Hills' and hit['longitude'] == '-118.406'
query = Q.prefix('zipcode', '90')
((field, facets), ) = indexer.facets(query, 'state.county').items()
assert field == 'state.county'
la, orange = sorted(filter(facets.get, facets))
assert la == 'CA.Los Angeles' and facets[la] > 100
assert orange == 'CA.Orange' and facets[orange] > 10
queries = {
term: Q.term(field, term)
for term in indexer.terms(field, 'CA.')
}
((field, facets), ) = indexer.facets(query, **{field: queries}).items()
assert all(value.startswith('CA.')
for value in facets) and set(facets) == set(queries)
assert facets['CA.Los Angeles'] == 264
groups = indexer.groupby(field, Q.term('state', 'CA'), count=1)
assert len(groups) == 1 < groups.count
(hits, ) = groups
assert hits.value == 'CA.Los Angeles' and len(
hits) == 1 and hits.count > 100
grouping = engine.documents.GroupingSearch(field,
sort=search.Sort(
indexer.sortfield(field)),
cache=False,
allGroups=True)
assert all(
grouping.search(indexer.indexSearcher, Q.alldocs()).facets.values())
assert len(grouping) == len(list(grouping)) > 100
assert set(grouping) > set(facets)
hits = indexer.search(query, timeout=-1)
assert not hits and not hits.count and math.isnan(hits.maxscore)
hits = indexer.search(query, timeout=10)
assert len(hits) == hits.count == indexer.count(
query) and hits.maxscore == 1.0
directory = store.RAMDirectory()
query = Q.term('state', 'CA')
size = indexer.copy(directory, query)
searcher = engine.IndexSearcher(directory)
assert len(searcher) == size and list(searcher.terms('state')) == ['CA']
path = os.path.join(tempdir, 'temp')
size = indexer.copy(path, exclude=query, merge=1)
assert len(searcher) + size == len(indexer)
searcher = engine.IndexSearcher(path)
assert len(searcher.segments) == 1 and 'CA' not in searcher.terms('state')
directory.close()