Fixes error GAP assignment error in fasta2select (#3156)

Fixes #3124

## Work done in this PR
- Moves the assignment of the GAP variable in fasta2select in order to avoid cases where it is referred to but not assigned.

package/AUTHORS

@@ -164,6 +164,7 @@ Chronological list of authors
   - Paarth Thadani
   - Henry Kobin
   - Kosuke Kudo
+  - Sulay Shah
 
 External code
 -------------

package/CHANGELOG

@@ -18,11 +18,13 @@ The rules for this file:
          calcraven，xiki-tempula, mieczyslaw, manuel.nuno.melo, PicoCentauri,
          hanatok, rmeli, aditya-kamath, tirkarthi, LeonardoBarneschi, hejamu,
          biogen98, orioncohen, z3y50n, hp115, ojeda-e, thadanipaarth, HenryKobin,
-         1ut
+         1ut, sulays
 
   * 2.0.0
 
 Fixes
+  * Fixed issue with unassigned 'GAP' variable in fasta2algin function when
+    resids are provided in input (Issue #3124)
   * Improve diffusionmap coverage (Issue #3208)
   * Removed deprecated parameters `n_jobs` and `precompute_distances` of
     sklearn.cluster.KMeans (Issue #2986)

package/MDAnalysis/analysis/align.py

@@ -1073,12 +1073,12 @@ def fasta2select(fastafilename, is_aligned=False,
     # implict assertion that we only have two sequences in the alignment
     orig_resids = [ref_resids, target_resids]
     offsets = [ref_offset, target_offset]
+    GAP = "-"
     for iseq, a in enumerate(alignment):
         # need iseq index to change orig_resids
         if orig_resids[iseq] is None:
             # build default: assume consecutive numbering of all
             # residues in the alignment
-            GAP = "-"
             length = len(a.seq) - a.seq.count(GAP)
             orig_resids[iseq] = np.arange(1, length + 1)
         else:

testsuite/MDAnalysisTests/analysis/test_align.py

@@ -465,15 +465,14 @@ def test_average_structure_in_memory(self, universe):
 
 class TestAlignmentProcessing(object):
     seq = FASTA
+    error_msg = "selection string has unexpected length"
 
     def test_fasta2select_aligned(self):
         """test align.fasta2select() on aligned FASTA (Issue 112)"""
         sel = align.fasta2select(self.seq, is_aligned=True)
         # length of the output strings, not residues or anything real...
-        assert len(sel['reference']) == 30623, ("selection string has"
-                                                "unexpected length")
-        assert len(
-            sel['mobile']) == 30623, "selection string has unexpected length"
+        assert len(sel['reference']) == 30623, self.error_msg
+        assert len(sel['mobile']) == 30623, self.error_msg
 
     @pytest.mark.skipif(executable_not_found("clustalw2"),
                         reason="Test skipped because clustalw2 executable not found")
@@ -483,10 +482,8 @@ def test_fasta2select_file(self, tmpdir):
         with tmpdir.as_cwd():
             sel = align.fasta2select(self.seq, is_aligned=False,
                                      alnfilename=None, treefilename=None)
-            assert len(sel['reference']) == 23080, ("selection string has"
-                                                    "unexpected length")
-            assert len(sel['mobile']) == 23090, ("selection string has"
-                                                 "unexpected length")
+            assert len(sel['reference']) == 23080, self.error_msg
+            assert len(sel['mobile']) == 23090, self.error_msg
 
     @pytest.mark.skipif(executable_not_found("clustalw2"),
                         reason="Test skipped because clustalw2 executable not found")
@@ -500,11 +497,17 @@ def test_fasta2select_ClustalW(self, tmpdir):
         # numbers computed from alignment with clustalw 2.1 on Mac OS X
         # [orbeckst] length of the output strings, not residues or anything
         # real...
-        assert len(sel['reference']) == 23080, ("selection string has"
-                                                "unexpected length")
-        assert len(
-            sel['mobile']) == 23090, "selection string has unexpected length"
-
+        assert len(sel['reference']) == 23080, self.error_msg
+        assert len(sel['mobile']) == 23090, self.error_msg
+
+    def test_fasta2select_resids(self, tmpdir):
+        """test align.fasta2select() when resids provided (Issue #3124)"""
+        resids = [x for x in range(705)]
+        sel = align.fasta2select(self.seq, is_aligned=True,
+                                 ref_resids=resids, target_resids=resids)
+        # length of the output strings, not residues or anything real...
+        assert len(sel['reference']) == 30621, self.error_msg
+        assert len(sel['mobile']) == 30621, self.error_msg
 
 def test_sequence_alignment():
     u = mda.Universe(PSF)