allow specified chiral features to SSS match unspecified features (rd…

…kit#8115)

Code/GraphMol/Substruct/SubstructMatch.cpp

@@ -246,6 +246,9 @@ bool MolMatchFinalCheckFunctor::operator()(const std::uint32_t q_c[],
     }
     const Atom *mAt = d_mol.getAtomWithIdx(m_c[i]);
     if (!detail::hasChiralLabel(mAt)) {
+      if (d_params.specifiedStereoQueryMatchesUnspecified) {
+        continue;
+      }
       return false;
     }
     if (qAt->getDegree() > mAt->getDegree()) {
@@ -327,10 +330,15 @@ bool MolMatchFinalCheckFunctor::operator()(const std::uint32_t q_c[],
     const Bond *mBnd = d_mol.getBondBetweenAtoms(
         q_to_mol[qBnd->getBeginAtomIdx()], q_to_mol[qBnd->getEndAtomIdx()]);
     CHECK_INVARIANT(mBnd, "Matching bond not found");
-    if (mBnd->getBondType() != Bond::DOUBLE ||
-        qBnd->getStereo() <= Bond::STEREOANY) {
+    if (mBnd->getBondType() != Bond::DOUBLE) {
       continue;
     }
+
+    if (!d_params.specifiedStereoQueryMatchesUnspecified &&
+        mBnd->getStereo() <= Bond::STEREOANY) {
+      return false;
+    }
+
     // don't think this can actually happen, but check to be sure:
     if (mBnd->getStereoAtoms().size() != 2) {
       continue;
@@ -386,14 +394,16 @@ class AtomLabelFunctor {
   AtomLabelFunctor(const ROMol &query, const ROMol &mol,
                    const SubstructMatchParameters &ps)
       : d_query(query), d_mol(mol), d_params(ps) {};
+
   bool operator()(unsigned int i, unsigned int j) const {
     bool res = false;
     if (d_params.useChirality) {
       const Atom *qAt = d_query.getAtomWithIdx(i);
       if (qAt->getChiralTag() == Atom::CHI_TETRAHEDRAL_CW ||
           qAt->getChiralTag() == Atom::CHI_TETRAHEDRAL_CCW) {
         const Atom *mAt = d_mol.getAtomWithIdx(j);
-        if (mAt->getChiralTag() != Atom::CHI_TETRAHEDRAL_CW &&
+        if (!d_params.specifiedStereoQueryMatchesUnspecified &&
+            mAt->getChiralTag() != Atom::CHI_TETRAHEDRAL_CW &&
             mAt->getChiralTag() != Atom::CHI_TETRAHEDRAL_CCW) {
           return false;
         }
@@ -421,6 +431,7 @@ class BondLabelFunctor {
           qBnd->getStereo() > Bond::STEREOANY) {
         const Bond *mBnd = d_mol[j];
         if (mBnd->getBondType() == Bond::DOUBLE &&
+            !d_params.specifiedStereoQueryMatchesUnspecified &&
             mBnd->getStereo() <= Bond::STEREOANY) {
           return false;
         }

Code/GraphMol/Substruct/SubstructMatch.h

@@ -67,7 +67,10 @@ struct RDKIT_SUBSTRUCTMATCH_EXPORT SubstructMatchParameters {
                         //!< match
   unsigned int maxRecursiveMatches =
       1000;  //!< maximum number of matches that the recursive substructure
-             //!< matching should return
+  //!< matching should return
+  bool specifiedStereoQueryMatchesUnspecified =
+      false;  //!< If set, query atoms and bonds with specified stereochemistry
+              //!< will match atoms and bonds with unspecified stereochemistry
   SubstructMatchParameters() {}
 };
 

Code/GraphMol/Substruct/SubstructUtils.cpp

@@ -273,6 +273,7 @@ void updateSubstructMatchParamsFromJSON(SubstructMatchParameters &params,
   PT_OPT_GET(maxMatches);
   PT_OPT_GET(maxRecursiveMatches);
   PT_OPT_GET(numThreads);
+  PT_OPT_GET(specifiedStereoQueryMatchesUnspecified);
 }
 
 std::string substructMatchParamsToJSON(const SubstructMatchParameters &params) {
@@ -287,6 +288,7 @@ std::string substructMatchParamsToJSON(const SubstructMatchParameters &params) {
   PT_OPT_PUT(maxMatches);
   PT_OPT_PUT(maxRecursiveMatches);
   PT_OPT_PUT(numThreads);
+  PT_OPT_PUT(specifiedStereoQueryMatchesUnspecified);
 
   std::stringstream ss;
   boost::property_tree::json_parser::write_json(ss, pt);

Code/GraphMol/Substruct/catch_tests.cpp

@@ -710,7 +710,7 @@ TEST_CASE("pickling HasPropWithValue queries") {
       REQUIRE(pklmol.getAtomWithIdx(0)->hasQuery());
       REQUIRE(pklmol.getBondWithIdx(0)->hasQuery());
       CHECK(SubstructMatch(*target, pklmol, ps).size() == 1);
-      // make sure we are idempotent in pickling      
+      // make sure we are idempotent in pickling
       CHECK(SubstructMatch(*target, *mol, ps).size() == 1);
     }
     {
@@ -720,7 +720,7 @@ TEST_CASE("pickling HasPropWithValue queries") {
       REQUIRE(pklmol.getAtomWithIdx(0)->hasQuery());
       REQUIRE(pklmol.getBondWithIdx(0)->hasQuery());
       CHECK(SubstructMatch(*target, pklmol, ps).size() == 0);
-      // make sure we are idempotent in pickling      
+      // make sure we are idempotent in pickling
       CHECK(SubstructMatch(*target, mol2, ps).size() == 0);
     }
     {
@@ -730,8 +730,58 @@ TEST_CASE("pickling HasPropWithValue queries") {
       REQUIRE(pklmol.getAtomWithIdx(0)->hasQuery());
       REQUIRE(pklmol.getBondWithIdx(0)->hasQuery());
       CHECK(SubstructMatch(*target, pklmol, ps).size() == 0);
-      // make sure we are idempotent in pickling      
+      // make sure we are idempotent in pickling
       CHECK(SubstructMatch(*target, mol3, ps).size() == 0);
     }
   }
 }
+
+TEST_CASE("specified query matches unspecified atom") {
+  SECTION("atom basics") {
+    auto q = "F[C@](Cl)(Br)C"_smarts;
+    REQUIRE(q);
+
+    auto m1 = "F[C@](Cl)(Br)C"_smiles;
+    REQUIRE(m1);
+    auto m2 = "FC(Cl)(Br)C"_smiles;
+    REQUIRE(m2);
+    auto m3 = "F[C@@](Cl)(Br)C"_smiles;
+    REQUIRE(m3);
+
+    SubstructMatchParameters ps;
+    ps.useChirality = true;
+    CHECK(SubstructMatch(*m1, *q, ps).size() == 1);
+    CHECK(SubstructMatch(*m2, *q, ps).empty());
+    CHECK(SubstructMatch(*m3, *q, ps).empty());
+
+    ps.specifiedStereoQueryMatchesUnspecified = true;
+    CHECK(SubstructMatch(*m1, *q, ps).size() == 1);
+    CHECK(SubstructMatch(*m2, *q, ps).size() == 1);
+    CHECK(SubstructMatch(*m3, *q, ps).empty());
+  }
+  SECTION("bond basics") {
+    auto q = "F/C=C/Br"_smarts;
+    REQUIRE(q);
+
+    auto m1 = "F/C=C/Br"_smiles;
+    REQUIRE(m1);
+    auto m2 = "FC=CBr"_smiles;
+    REQUIRE(m2);
+    auto m3 = "F/C=C\\Br"_smiles;
+    REQUIRE(m3);
+
+    SubstructMatchParameters ps;
+    ps.useChirality = true;
+    CHECK(SubstructMatch(*m1, *q, ps).size() == 1);
+    CHECK(SubstructMatch(*m2, *q, ps).empty());
+    CHECK(SubstructMatch(*m3, *q, ps).empty());
+
+    ps.specifiedStereoQueryMatchesUnspecified = true;
+    std::cerr << "m1" << std::endl;
+    CHECK(SubstructMatch(*m1, *q, ps).size() == 1);
+    std::cerr << "m2" << std::endl;
+    CHECK(SubstructMatch(*m2, *q, ps).size() == 1);
+    std::cerr << "m3" << std::endl;
+    CHECK(SubstructMatch(*m3, *q, ps).empty());
+  }
+}

Code/GraphMol/Wrap/Mol.cpp

@@ -334,12 +334,17 @@ struct mol_wrapper {
             "0 selects the number of concurrent threads supported by the"
             "hardware. negative values are added to the number of concurrent"
             "threads supported by the hardware.")
+        .def_readwrite(
+            "bondProperties", &RDKit::SubstructMatchParameters::bondProperties,
+            "bond properties that must be equivalent in order to match.")
         .def_readwrite(
             "atomProperties", &RDKit::SubstructMatchParameters::atomProperties,
             "atom properties that must be equivalent in order to match.")
         .def_readwrite(
-            "bondProperties", &RDKit::SubstructMatchParameters::bondProperties,
-            "bond properties that must be equivalent in order to match.")
+            "specifiedStereoQueryMatchesUnspecified",
+            &RDKit::SubstructMatchParameters::
+                specifiedStereoQueryMatchesUnspecified,
+            "If set, query atoms and bonds with specified stereochemistry will match atoms and bonds with unspecified stereochemistry.")
         .def("setExtraFinalCheck", setSubstructMatchFinalCheck,
              python::with_custodian_and_ward<1, 2>(),
              python::args("self", "func"),