add ignore_same_residue kwarg to InterRDF

package/CHANGELOG

@@ -13,7 +13,7 @@ The rules for this file:
   * release numbers follow "Semantic Versioning" http://semver.org
 
 ------------------------------------------------------------------------------
-??/??/?? IAlibay, jaclark5, MohitKumar020291
+??/??/?? IAlibay, jaclark5, MohitKumar020291, orionarcher
 
  * 2.6.0
 
@@ -23,6 +23,7 @@ Fixes
     (PR #4163, Issue #4159)
 
 Enhancements
+  * Added a `exclude_same` kwarg to InterRDF (PR #4161)
   * LAMMPSDump Reader use of `continuous` option to stitch trajectories
     (Issue #3546)
 

package/MDAnalysis/analysis/rdf.py

@@ -125,6 +125,10 @@ class InterRDF(AnalysisBase):
 
     exclusion_block : tuple
         A tuple representing the tile to exclude from the distance array.
+    exclude_same : str
+        Will exclude pairs of atoms that share the same "residue", "segment", or "chain".
+        Those are the only valid values. This is intended to remove atoms that are
+        spatially correlated due to direct bonded connections.
     verbose : bool
         Show detailed progress of the calculation if set to `True`
 
@@ -220,6 +224,7 @@ def __init__(self,
                  range=(0.0, 15.0),
                  norm="rdf",
                  exclusion_block=None,
+                 exclude_same=None,
                  **kwargs):
         super(InterRDF, self).__init__(g1.universe.trajectory, **kwargs)
         self.g1 = g1
@@ -229,6 +234,17 @@ def __init__(self,
         self.rdf_settings = {'bins': nbins,
                              'range': range}
         self._exclusion_block = exclusion_block
+        if exclude_same is not None and exclude_same not in ['residue', 'segment', 'chain']:
+            raise ValueError(
+                "The exclude_same argument to InterRDF must be None, 'residue', 'segment' "
+                "or 'chain'."
+            )
+        if exclude_same is not None and exclusion_block is not None:
+            raise ValueError(
+                "The exclude_same argument to InterRDF cannot be used with exclusion_block."
+            )
+        name_to_attr = {'residue': 'resindices', 'segment': 'segindices', 'chain': 'chainIDs'}
+        self.exclude_same = name_to_attr.get(exclude_same)
 
         if self.norm not in ['rdf', 'density', 'none']:
             raise ValueError(f"'{self.norm}' is an invalid norm. "
@@ -261,6 +277,13 @@ def _single_frame(self):
             mask = np.where(idxA != idxB)[0]
             dist = dist[mask]
 
+        if self.exclude_same is not None:
+            # Ignore distances between atoms in the same attribute
+            attr_ix_a = getattr(self.g1, self.exclude_same)[pairs[:, 0]]
+            attr_ix_b = getattr(self.g2, self.exclude_same)[pairs[:, 1]]
+            mask = np.where(attr_ix_a != attr_ix_b)[0]
+            dist = dist[mask]
+
         count, _ = np.histogram(dist, **self.rdf_settings)
         self.results.count += count
 

testsuite/MDAnalysisTests/analysis/test_rdf.py

@@ -32,7 +32,9 @@
 
 @pytest.fixture()
 def u():
-    return mda.Universe(two_water_gro, in_memory=True)
+    u = mda.Universe(two_water_gro, in_memory=True)
+    u.add_TopologyAttr('chainIDs', u.atoms.resids)
+    return u
 
 
 @pytest.fixture()
@@ -94,6 +96,27 @@ def test_exclusion(sels):
     assert rdf.results.count.sum() == 4
 
 
+@pytest.mark.parametrize("attr, count", [
+    ("residue", 8),
+    ("segment", 0),
+    ("chain", 8)])
+def test_ignore_same_residues(sels, attr, count):
+    # should see two distances with 4 counts each
+    s1, s2 = sels
+    rdf = InterRDF(s2, s2, exclude_same=attr).run()
+    assert rdf.rdf[0] == 0
+    assert rdf.results.count.sum() == count
+
+
+def test_ignore_same_residues_fails(sels):
+    s1, s2 = sels
+    with pytest.raises(ValueError, match="The exclude_same argument to InterRDF must be"):
+        InterRDF(s2, s2, exclude_same="unsupported").run()
+
+    with pytest.raises(ValueError, match="The exclude_same argument to InterRDF cannot be used with"):
+        InterRDF(s2, s2, exclude_same="residue", exclusion_block=tuple()).run()
+
+
 @pytest.mark.parametrize("attr", ("rdf", "bins", "edges", "count"))
 def test_rdf_attr_warning(sels, attr):
     s1, s2 = sels