#1245 Reaction/Molecule autoloaders don't load SMARTS (#1243)

epam · Aug 24, 2023 · 964da3f · 964da3f
1 parent 490145e
commit 964da3f
Show file tree

Hide file tree

Showing 4 changed files with 39 additions and 5 deletions.
diff --git a/api/tests/integration/ref/ketcher/detect.py.out b/api/tests/integration/ref/ketcher/detect.py.out
@@ -5,3 +5,5 @@ detected as molecule
 detected as query
 detected as molecule
 detected as query
+detected as query
+detected as query reaction
diff --git a/api/tests/integration/tests/ketcher/detect.py b/api/tests/integration/tests/ketcher/detect.py
@@ -56,3 +56,7 @@ def detect(indigo, molstr):
     detect(indigo, mol)
     mfile.close()
     detect(indigo, "[#6]1-[#6]=[#6]-[#6]=[#6]-[#6]=1")
+    # check SMARTS
+    detect(indigo, "[$([CX3]=[OX1]),$([CX3+]-[OX1-])]")
+    # check reaction SMARTS
+    detect(indigo, "([#8:1].[#6:2])>>([#8:1][#6:2])")
diff --git a/core/indigo-core/molecule/src/molecule_auto_loader.cpp b/core/indigo-core/molecule/src/molecule_auto_loader.cpp
@@ -433,20 +433,33 @@ void MoleculeAutoLoader::_loadMolecule(BaseMolecule& mol)
         try
         {
             SmilesLoader loader(*_scanner);
+            long long start = _scanner->tell();
 
             loader.ignore_closing_bond_direction_mismatch = ignore_closing_bond_direction_mismatch;
             loader.stereochemistry_options = stereochemistry_options;
             loader.ignore_cistrans_errors = ignore_cistrans_errors;
             loader.ignore_no_chiral_flag = ignore_no_chiral_flag;
 
             /*
-            If exception is thrown, the string is rather an IUPAC name than a SMILES string
+            If exception is thrown, try the SMARTS, if exception thrown again - the string is rather an IUPAC name than a SMILES string
             We catch it and pass down to IUPAC name conversion
             */
             if (query)
-                loader.loadQueryMolecule((QueryMolecule&)mol);
+            {
+                try
+                {
+                    loader.loadQueryMolecule(static_cast<QueryMolecule&>(mol));
+                }
+                catch (Exception& e)
+                {
+                    _scanner->seek(start, SEEK_SET);
+                    loader.loadSMARTS(static_cast<QueryMolecule&>(mol));
+                }
+            }
             else
-                loader.loadMolecule((Molecule&)mol);
+            {
+                loader.loadMolecule(static_cast<Molecule&>(mol));
+            }
             return;
         }
         catch (Exception& e)

diff --git a/core/indigo-core/reaction/src/reaction_auto_loader.cpp b/core/indigo-core/reaction/src/reaction_auto_loader.cpp
@@ -284,6 +284,7 @@ void ReactionAutoLoader::_loadReaction(BaseReaction& reaction)
     // check for SMILES format
     if (Scanner::isSingleLine(*_scanner))
     {
+        long long pos = _scanner->tell();
         RSmilesLoader loader(*_scanner);
 
         loader.ignore_closing_bond_direction_mismatch = ignore_closing_bond_direction_mismatch;
@@ -292,9 +293,23 @@ void ReactionAutoLoader::_loadReaction(BaseReaction& reaction)
         loader.ignore_bad_valence = ignore_bad_valence;
 
         if (query)
-            loader.loadQueryReaction((QueryReaction&)reaction);
+        {
+            // Try to load query as SMILES, if error occured - try to load as SMARTS
+            try
+            {
+                loader.loadQueryReaction(static_cast<QueryReaction&>(reaction));
+            }
+            catch (Exception& e)
+            {
+                loader.smarts_mode = true;
+                _scanner->seek(pos, SEEK_SET);
+                loader.loadQueryReaction(static_cast<QueryReaction&>(reaction));
+            }
+        }
         else
-            loader.loadReaction((Reaction&)reaction);
+        {
+            loader.loadReaction(static_cast<Reaction&>(reaction));
+        }
     }
 
     // default is Rxnfile format