def naive_has_similar(self, s, distance):
for t in self.strings:
if len(t) != len(s):
continue
if hamming_distance(t, s) <= distance:
return True
return False
def hamming_single_linkage(strings, mismatches, use_trie=False):
"""
Cluster a set of strings by their hamming distance: Strings with
a distance of at most 'mismatches' will be put into the same cluster.
Return a list of connected components (clusters).
"""
components = []
# First pre-cluster strings by length
string_lists = defaultdict(list)
for s in strings:
string_lists[len(s)].append(s)
for strings in string_lists.values():
graph = Graph(strings)
if use_trie:
trie = Trie()
for s in strings:
trie.add(s)
for s in strings:
for neighbor in trie.find_all_similar(s, mismatches):
if neighbor != s:
graph.add_edge(s, neighbor)
else:
for i, s in enumerate(strings):
for j, t in enumerate(strings[i + 1:]):
if hamming_distance(s, t) <= mismatches:
graph.add_edge(s, t)
components.extend(graph.connected_components())
return components
def naive_find_all_similar(self, s, distance):
for t in self.strings:
if len(t) != len(s):
continue
if hamming_distance(t, s) <= distance:
yield t
def naive_has_similar(t):
for s in strings:
if hamming_distance(s, t) <= dist:
return True
return False