Allow masking character

This is a minor change that only applies when there is a masking character (not involved in any analysis of the SHARK-Dive manuscript).

src/shark_atplab/core/sequence_pair_similarity.py

@@ -45,7 +45,10 @@ class SeqPairSimilarity:
         kmer_count_map = {}
         for i in range(len(sequence) - k_value + 1):
             kmer = sequence[i: i + k_value]
-            if kmer in kmer_count_map:
+            # residues can be masked with the '*' character. K-mers that contain masked residues should be ignored
+            if '*' in kmer:
+                continue
+            elif kmer in kmer_count_map:
                 kmer_count_map[kmer] += 1
             else:
                 kmer_count_map[kmer] = 1
@@ -211,7 +214,7 @@ class SeqPairSimilarity:
         column_scores = self.get_axis_score(self.kmer_count_map_2, self.kmer_count_map_1, self.similarity_matrix,
                                             reverse=True)
         final_score = (sum(row_scores) + sum(column_scores)) / (
-                len(self.sequence1) + len(self.sequence2) - 2 * (self.k - 1))
+                sum(self.kmer_count_map_1.values()) + sum(self.kmer_count_map_2.values()))
         return final_score
 
     def get_similarity_score__sparse(self):