def query(self, q, ranked=True):
"""Return a ranked list of matching `Document` instances."""
qq = Query.parse(q)
res = self.discodex_client.query(self.spec.invindex_name, qq)
res = map(TfIdf.undemux, res)
if not res:
return []
pageranks = None
if ranked:
scoredb = ScoreDB(self.spec.scoredb_path)
uris = [e[0] for e in res]
pageranks = dict(scoredb.rank(uris))
if not pageranks:
raise Exception("no ranks available")
docs = []
for uri,scores in res:
doc = self.docset.get(uri)
doc.score = Score(**scores)
if pageranks:
doc.score['pagerank'] = pageranks[uri]
doc.excerpt = doc.excerpt(qq)
docs.append(doc)
return docs
def test_excerpt_lowercases(self):
qq = Query.parse('welcome')
self.assertEquals('Welcome to example', fixtures.example.excerpt(qq, radius=20))
def test_excerpt(self):
qq = Query.parse('example')
self.assertEquals('Welcome to example', fixtures.example.excerpt(qq, radius=11))
self.assertEquals('... example', fixtures.example.excerpt(qq, radius=1))
def test_eliminates_stopwords_when_stemming(self):
qq = Query.parse('welcome & a')
self.assertEquals('welcom&a|~a', qq.format())
def test_stems(self):
self.assertEquals('welcom', Query.parse('welcome').format())
self.assertEquals('welcom&univers', Query.parse('welcome & university').format())
def test_non_negated_literals(self):
qq = Query.parse('abcd & ~wxyz & efgh')
self.assertEquals(set(['abcd', 'efgh']), set(qq.non_negated_literals()))