def test_min_match_not_satisfied(self):
'''Tests returned values correct if there are some matching characters,
but fewer than the minimum required
'''
matches = spm.get_matches('PNF', 'SPLNFRT', 2)
self.assertEqual(matches, [])
matches = spm.get_matches('PNF', 'PTTP', 2)
self.assertEqual(matches, [])
def test_min_match_satisfied(self):
'''Tests returned value correct when minMatch > 1
'''
matches = spm.get_matches('PNF', 'PNF', 2)
self.assertEqual(matches, [(0, 3)])
matches = spm.get_matches('AB', 'ABTABAA', 2)
self.assertEqual(matches, [(0, 2), (3, 2)])
matches = spm.get_matches('AB', 'ATTABTTAJ', 2)
self.assertEqual(matches, [(3, 2)])
matches = spm.get_matches('ABC', 'ATTABTTAJABC', 2)
self.assertEqual(matches, [(3, 2), (9, 3)])
def score(self, query, matchString):
'''Scores a text-based match between strings, giving matching beginnings high scores.
Overrides ScoringMethod.score()
Arguments:
query -- the query string, e.g. 'lond'
matchString -- UTF-8 string on which the query matched; e.g. 'london'.
Returns:
A score in range [0, 1] where 0 indicates no match at all (as determined by
get_matches method in simpleprefixmatch.py), and 1 indicates
a character-for-character exact match between query and matchString.
Preconditions:
-- All alphabetical characters in query and matchString are in the
same case (all upper or all lower).
'''
# Get location and number of character matches
match = simpleprefixmatch.get_matches(query, matchString, 1)
baseShift = self.baseShift
substringBonus = self.subStringBonus
exactMatchBonus = self.exactMatchBonus
startMatchBonus = self.startMatchBonus
# No match at all, return the lowest possible score.
if (len(match) == 0):
return 0.0
# Full, character-for-character match returns 1.0
if (matchString == query):
return 1.0
# If the match is partial, the more characters match the better.
# Single-character matches result in tonnes of results, so don't count them.
matchLen = 0.0
fullSubString = False
for tup in match:
if (tup[1] > 1):
matchLen = matchLen + tup[1]
if (tup[1] >= len(query)):
fullSubString = True
# The greater portion of query matches the matchString, the better:
baseScore = matchLen / len(matchString)
# Since a raw score of 0 will result in a score of 0.5 once sigmoid applied,
# shift the function to the right by a bit.
# For reference, sigmoid(-2) is approx 0.12
baseScore = baseScore - baseShift
# Give an additional bonus if the query string is a substring of matchString.
# Note the query may occur multiple times within the matchString;
# e.g. query = 'lon' and matchString = 'lonlon'
if (fullSubString):
baseScore = baseScore + substringBonus
# Additional bonus if strings match exactly
if (len(query) == len(matchString)):
baseScore = baseScore + exactMatchBonus
# We give bonus points for matching at the start of the matchString,
# and for having a longer matching sequence
if (match[0][0] == 0):
baseScore = baseScore + startMatchBonus * match[0][1]
return sigmoid(baseScore)
def test_empty_text(self):
matches = spm.get_matches('adj', '')
self.assertEqual(matches, [])
def test_empty_pattern(self):
matches = spm.get_matches('', 'adj')
self.assertEqual(matches, [])
def test_mixed(self):
'''Multiple partial pattern matches in text'''
matches = spm.get_matches('jam', 'ajajamkaj', 1)
self.assertEqual(matches, [(1,2), (3, 3), (8, 1)])
def test_sequential(self):
matches = spm.get_matches('lond', 'lonlon')
self.assertEqual(matches, [(0, 3), (3, 3)])
def test_match_prefix(self):
'''Pattern is a prefix of text'''
matches = spm.get_matches('lond', 'londonderry')
self.assertEqual(matches, [(0, 4)])
def test_match_suffix(self):
'''Pattern is a suffix of text'''
matches = spm.get_matches('shire', 'oxfordshire')
self.assertEqual(matches, [(6, 5)])
def test_no_match(self):
matches = spm.get_matches('london', 'sydney')
self.assertEqual(matches, [])
def test_pattern_longer_none(self):
matches = spm.get_matches('baroomfoom', 'zim')
self.assertEqual(matches, [])
def test_pattern_longer_full(self):
matches = spm.get_matches('sydn', 'syd')
self.assertEqual(matches, [(0, 3)])
def test_pattern_longer_partial(self):
matches = spm.get_matches('baroomfoom', 'cabar')
self.assertEqual(matches, [(2, 3)])
def test_pattern_longer(self):
'''Case in which pattern is longer than text, full match.'''
matches = spm.get_matches('bambambam', 'bam')
self.assertEqual(matches,[(0, 3)])
