def fuzzy(qs, prefix=0, maxdist=2):
prefix = int(prefix)
maxdist = int(maxdist)
return query.FuzzyTerm(qs[0].fieldname,
qs[0].text,
prefixlength=prefix,
maxdist=maxdist)
def test_fuzzyterm():
schema = fields.Schema(id=fields.STORED, f=fields.TEXT)
ix = RamStorage().create_index(schema)
w = ix.writer()
w.add_document(id=1, f=u("alfa bravo charlie delta"))
w.add_document(id=2, f=u("bravo charlie delta echo"))
w.add_document(id=3, f=u("charlie delta echo foxtrot"))
w.add_document(id=4, f=u("delta echo foxtrot golf"))
w.commit()
with ix.searcher() as s:
q = query.FuzzyTerm("f", "brave")
assert_equal([d["id"] for d in s.search(q)], [1, 2])
def find(self, wordlist):
'''look up all the words in the wordlist.
If none are found return an empty dictionary
* more rules here
'''
wordlist = [
word for word in wordlist
if (self.minlength <= len(word) <= self.maxlength)
and not self.is_stopword(word.upper())
]
if not wordlist:
return {}
searcher = self._get_searcher()
q = query.And(
[query.FuzzyTerm("content", word.lower()) for word in wordlist])
results = searcher.search(q, limit=None)
return [tuple(result["identifier"].split(':')) for result in results]