Fix issue with masking of trailing nucleotides (#55)

althonos · May 18, 2024 · 1562471 · 1562471
1 parent 2a6eca0
commit 1562471
Show file tree

Hide file tree

Showing 4 changed files with 8 additions and 7 deletions.
diff --git a/pyrodigal/lib.pyi b/pyrodigal/lib.pyi
@@ -82,7 +82,8 @@ class Sequence(typing.Sized):
     def __init__(
         self,
         sequence: Union[str, bytes, bytearray, Sequence],
-        mask: bool = False
+        mask: bool = False,
+        mask_size; int = 50,
     ) -> None: ...
     def __len__(self) -> int: ...
     def __sizeof__(self) -> int: ...

diff --git a/pyrodigal/lib.pyx b/pyrodigal/lib.pyx
@@ -723,6 +723,8 @@ cdef class Sequence:
                     if i >= mask_size + mask_begin:
                         self.masks._add_mask(mask_begin, i)
                     mask_begin = -1
+        if mask_begin != -1:
+            self.masks._add_mask(mask_begin, self.slen)
         return 0
 
     cdef int _allocate(self, int slen) except 1:

diff --git a/pyrodigal/tests/test_gene.py b/pyrodigal/tests/test_gene.py
@@ -60,12 +60,6 @@ def test_translate_strictness(self):
         self.assertEqual(strict[1:9], "XXXXXXXX")
         self.assertEqual(non_strict[1:9], "PRLTVAGS")
 
-        # strict = gene.translate(strict=True, translation_table=3)
-        # non_strict = gene.translate(strict=False, translation_table=3)
-
-        # self.assertEqual(strict[1:9], "XXXXXXXX")
-        # self.assertEqual(non_strict[1:9], "PRTTVAGS")
-
         strict = gene.translate(strict=True, translation_table=12)
         non_strict = gene.translate(strict=False, translation_table=12)
 

diff --git a/pyrodigal/tests/test_sequence.py b/pyrodigal/tests/test_sequence.py
@@ -73,3 +73,7 @@ def test_shine_dalgarno_mismatch(self):
             # AGGxGG if i == 0 else None
             expected = 19 if i == 0 else 0
             self.assertEqual(seq.shine_dalgarno(i, 15, tinf, exact=False), expected, i)
+
+    def test_mask_trailing(self):
+        seq = Sequence("AGCGGGCTACTGNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN", mask=True, mask_size=10)
+        self.assertEqual(len(seq.masks), 1)