#1252 SMARTS loader load grouped components as separate molecules (#…

…1255)

api/tests/integration/ref/rsmiles/rsmiles_smarts.py.out

@@ -0,0 +1 @@
+SMARTS component-level grouping load ok

api/tests/integration/tests/rsmiles/rsmiles_smarts.py

@@ -0,0 +1,17 @@
+import os
+import sys
+
+sys.path.append(
+    os.path.normpath(
+        os.path.join(os.path.abspath(__file__), "..", "..", "..", "common")
+    )
+)
+from env_indigo import *
+
+indigo = Indigo()
+
+
+rxn1 = indigo.loadReactionSmarts("([#8:1].[#6:2])>>([#8:1].[#6:2])")
+assert rxn1.countReactants() == 1
+assert rxn1.countProducts() == 1
+print("SMARTS component-level grouping load ok")

core/indigo-core/graph/src/graph_decomposer.cpp

@@ -72,7 +72,7 @@ int GraphDecomposer::decompose(const Filter* filter, const Filter* edge_filter,
         int top = 1, bottom = 0;
 
         queue[0] = vertex_idx;
-        while (top != bottom)
+        while (top != bottom) // while queue not empty
         {
             auto v_bottom_id = queue[bottom];
             const Vertex& vertex = _graph.getVertex(v_bottom_id);
@@ -92,10 +92,10 @@ int GraphDecomposer::decompose(const Filter* filter, const Filter* edge_filter,
                 if (edge_filter != 0 && !edge_filter->valid(vertex.neiEdge(i)))
                     continue;
 
-                if (_component_ids[other] == -1)
+                if (_component_ids[other] == -1) // if other unused
                 {
-                    queue[top++] = other;
-                    _component_ids[other] = -2;
+                    queue[top++] = other;       // queue.push(other)
+                    _component_ids[other] = -2; // mark as in_queue
                 }
 
                 if (_component_ids[other] == -2)
@@ -129,7 +129,7 @@ int GraphDecomposer::decompose(const Filter* filter, const Filter* edge_filter,
                     }
                 }
             }
-            bottom++;
+            bottom++; // queue.pop()
         }
 
         n_comp++;

core/indigo-core/reaction/src/rsmiles_loader.cpp

@@ -75,6 +75,25 @@ void RSmilesLoader::loadQueryReaction(QueryReaction& rxn)
     _loadReaction();
 }
 
+static void _createComponenetsExternalNeighbors(QueryMolecule& qmol, std::list<std::unordered_set<int>>& externNeib)
+{
+    std::unordered_map<int, std::unordered_set<int>> componentAtoms;
+    for (int i = 0; i < qmol.components.size(); ++i)
+    {
+        int componentId = qmol.components[i];
+        if (componentId > 0)
+        { // vertice[i] belongs to component #Id
+            componentAtoms[componentId].insert(i);
+        }
+    }
+    for (auto elem : componentAtoms)
+    {
+        auto atoms = elem.second;
+        if (atoms.size() > 1)
+            externNeib.emplace_back(atoms);
+    }
+}
+
 void RSmilesLoader::_loadReaction()
 {
     _brxn->clear();
@@ -83,6 +102,10 @@ void RSmilesLoader::_loadReaction()
     std::unique_ptr<BaseMolecule>& rcnt = mols[0];
     std::unique_ptr<BaseMolecule>& ctlt = mols[1];
     std::unique_ptr<BaseMolecule>& prod = mols[2];
+    std::list<std::unordered_set<int>> mols_extNeibs[3];
+    std::list<std::unordered_set<int>>& rcnt_extNeibs = mols_extNeibs[0];
+    std::list<std::unordered_set<int>>& ctlt_extNeibs = mols_extNeibs[1];
+    std::list<std::unordered_set<int>>& prod_extNeibs = mols_extNeibs[2];
 
     QS_DEF(Array<int>, rcnt_aam);
     QS_DEF(Array<int>, ctlt_aam);
@@ -124,6 +147,7 @@ void RSmilesLoader::_loadReaction()
     {
         rcnt = std::make_unique<QueryMolecule>();
         r_loader.loadQueryMolecule(static_cast<QueryMolecule&>(*rcnt));
+        _createComponenetsExternalNeighbors(static_cast<QueryMolecule&>(*rcnt), rcnt_extNeibs);
     }
     rcnt_aam.copy(rcnt->reaction_atom_mapping);
 
@@ -138,11 +162,6 @@ void RSmilesLoader::_loadReaction()
         buf.push(c);
     }
 
-    if (_rxn != 0)
-        ctlt = std::make_unique<Molecule>();
-    else
-        ctlt = std::make_unique<QueryMolecule>();
-
     BufferScanner c_scanner(buf);
     SmilesLoader c_loader(c_scanner);
 
@@ -162,6 +181,7 @@ void RSmilesLoader::_loadReaction()
     {
         ctlt = std::make_unique<QueryMolecule>();
         c_loader.loadQueryMolecule(static_cast<QueryMolecule&>(*ctlt));
+        _createComponenetsExternalNeighbors(static_cast<QueryMolecule&>(*ctlt), ctlt_extNeibs);
     }
     ctlt_aam.copy(ctlt->reaction_atom_mapping);
 
@@ -201,6 +221,7 @@ void RSmilesLoader::_loadReaction()
     {
         prod = std::make_unique<QueryMolecule>();
         p_loader.loadQueryMolecule(static_cast<QueryMolecule&>(*prod));
+        _createComponenetsExternalNeighbors(static_cast<QueryMolecule&>(*prod), prod_extNeibs);
     }
     prod_aam.copy(prod->reaction_atom_mapping);
 
@@ -209,9 +230,9 @@ void RSmilesLoader::_loadReaction()
     QS_DEF(Array<int>, p_fragments);
     Array<int>* fragments[] = {&r_fragments, &c_fragments, &p_fragments};
 
-    r_fragments.clear_resize(rcnt->countComponents());
-    c_fragments.clear_resize(ctlt->countComponents());
-    p_fragments.clear_resize(prod->countComponents());
+    r_fragments.clear_resize(rcnt->countComponents(rcnt_extNeibs));
+    c_fragments.clear_resize(ctlt->countComponents(ctlt_extNeibs));
+    p_fragments.clear_resize(prod->countComponents(prod_extNeibs));
 
     for (int i = 0; i < r_fragments.size(); i++)
         r_fragments[i] = i;

core/indigo-core/tests/tests/reaction.cpp

@@ -91,3 +91,12 @@ TEST_F(IndigoCoreReactionTest, aliases_complex)
         saverReactionJson(reaction).c_str());
     */
 }
+
+TEST_F(IndigoCoreReactionTest, smarts_reaction)
+{
+    QueryReaction qr;
+    std::string smarts_in = "([#8:1].[#6:2])>>([#8:1].[#6:2])";
+    loadQueryReaction(smarts_in.c_str(), qr);
+    ASSERT_EQ(qr.reactantsCount(), 1);
+    ASSERT_EQ(qr.productsCount(), 1);
+}