fix: protein hashing

socialgene/base/molbio.py

@@ -149,6 +149,8 @@ def _standardize_sequence(self):
         amino acids, raising an error if an unknown character is found.
         """
         self.sequence = self.sequence.upper()
+        self.sequence = self.sequence.replace(" ", "")
+        self.sequence = self.sequence.strip("*")
         if not all([i in self._amino_acids for i in set(self.sequence)]):
             log.error(self.sequence)
             raise ValueError("Unknown character/letter in protein sequence")

socialgene/hashing/hashing.py

@@ -26,7 +26,7 @@ def hash_aminos(input, **kwargs):
     # make sure everything is uppercase before hashing
     cleaned = input.upper()
     # remove all whitespace
-    cleaned = "".join(cleaned.split())
+    cleaned = cleaned.replace(" ", "")
     # remove "*" if it's at the beginning or end of the string but not in the middle
     cleaned = cleaned.strip("*")
     return hasher(input=cleaned, **kwargs)