WIP: adding bond types as facet

gsrs-module-substances-core/src/main/java/gsrs/module/substance/indexers/ChemicalSubstanceBondTypeIndexer.java

@@ -0,0 +1,140 @@
+package gsrs.module.substance.indexers;
+
+import java.util.function.Consumer;
+
+import org.springframework.beans.factory.annotation.Autowired;
+import org.springframework.stereotype.Component;
+
+import gov.nih.ncats.molwitch.Bond;
+import gov.nih.ncats.molwitch.Chemical;
+import ix.core.chem.StructureProcessor;
+import ix.core.search.text.IndexValueMaker;
+import ix.core.search.text.IndexableValue;
+import ix.core.search.text.ReflectingIndexValueMaker;
+import ix.ginas.models.v1.ChemicalSubstance;
+import ix.ginas.models.v1.Substance;
+import lombok.extern.slf4j.Slf4j;
+
+/**
+ * Adds chemical bond types as a facet for quick filtering. 
+ * 
+ */
+@Component
+@Slf4j
+public class ChemicalSubstanceBondTypeIndexer implements IndexValueMaker<Substance>{
+    private static final String TEXT_SINGLE_BOND = "Single Bond";
+    private static final String TEXT_DOUBLE_BOND = "Double Bond";
+    private static final String TEXT_TRIPLE_BOND = "Triple Bond";
+    private static final String TEXT_AROMATIC_BOND = "Aromatic Bond";
+    private static final String CHEMICAL_BONDS_FACET_NAME = "Chemical Bonds";
+
+
+    private static final boolean SMILES_LIKE_SYNTAX = false;
+
+
+    @Override
+    public Class<Substance> getIndexedEntityClass() {
+        return Substance.class;
+    }
+
+    //This is the method which does the work
+    @Override
+    public void createIndexableValues(Substance s, Consumer<IndexableValue> consumer) {
+        if(s instanceof ChemicalSubstance){
+            addBondTypes((ChemicalSubstance)s, consumer);
+        }
+    }
+
+    /**
+     * Create a canonical text string for the supplied {@link Bond}. This returns
+     * the atoms in pseudo hill-order, where carbon is always first, hydrogen is always
+     * last, and the other atoms are sorted alphabetically.
+     *  
+     * @param b - The chemical bond to be canonicalized
+     * @return A canonical string for the bond
+     */
+    private String stdBond(Bond b) {
+        String typeText= TEXT_SINGLE_BOND;
+
+        String a1=b.getAtom1().getSymbol();
+        String a2=b.getAtom2().getSymbol();
+        if(a1.equals("C")) {
+
+        }else if(a2.equals("C")) {
+            String t=a1;
+            a1=a2;
+            a2=t;
+        }else if(a1.equals("H")) {
+            String t=a1;
+            a1=a2;
+            a2=t;
+        }else if(a2.equals("H")) {
+
+        }else if(a1.compareTo(a2)>0) {
+            String t=a1;
+            a1=a2;
+            a2=t;
+        }
+
+        String bt="-";
+        switch(b.getBondType().getOrder()) {
+        case 2:
+            bt="=";
+            typeText=TEXT_DOUBLE_BOND;
+            break;
+        case 3:
+            bt="#";
+            typeText=TEXT_TRIPLE_BOND;
+            break;
+        default:
+            bt="-";
+            typeText=TEXT_SINGLE_BOND;
+            break;
+        }
+        if(b.isAromatic()) {
+            bt=":";
+            typeText=TEXT_AROMATIC_BOND;
+        }
+
+
+        //This will return in smiles-like syntax
+        if(SMILES_LIKE_SYNTAX) {
+            return a1 + bt + a2;
+        }else {
+            //this returns in more verbose syntax
+            return a1 + "-" + a2 + ", " + typeText;    
+        }
+    }
+
+
+    /**
+     * Indexes all the unique bonds found in the supplied {@link ChemicalSubstance} by producing
+     * the canonical bond names using {@link #stdBond(Bond)}. This method also adds explicit hydrogens
+     * and aromatizes the structure (transiently) before indexing.
+     * @param s - The {@link ChemicalSubstance} to be indexed
+     * @param consumer - The consumer to accept new {@link IndexableValue} objects
+     */
+    public void addBondTypes(ChemicalSubstance s, Consumer<IndexableValue> consumer) {
+        try{
+            //Quick simple standardization of the supplied molecule
+            Chemical c = s.toChemical();
+
+            //We want hydrogens explicitly populated, this adds them
+            c.makeHydrogensExplicit();
+
+            //Aromatic bonds should be present, not kekulized bonds
+            c.aromatize();
+
+            //for each bond, get canonical representation, sort, get unique forms, add to index
+            c.bonds().map(b->stdBond(b))
+            .sorted()
+            .distinct()
+            .forEach(bt->{
+                consumer.accept(IndexableValue.simpleFacetStringValue(CHEMICAL_BONDS_FACET_NAME, bt));
+            });
+        }catch(Exception e){
+            log.warn("Problem indexing substance for bond types:" + s.uuid, e);
+        }
+    }
+
+}

gsrs-module-substances-core/src/main/resources/substances-core.conf

@@ -447,10 +447,9 @@ gsrs.indexers.list=[
 "indexer" = "gsrs.module.substance.indexers.DeprecatedIndexValueMaker",
 "class" = "ix.ginas.models.v1.Substance",
 },
-# This should be unnecessary right now. May be used later
-#{
-# "indexer" = "gsrs.module.substance.indexers.ChemicalSubstanceStructureHashIndexValueMaker", 
-#},
+{
+"indexer" = "gsrs.module.substance.indexers.ChemicalSubstanceBondTypeIndexer"
+},
 {
 "indexer" = "gsrs.module.substance.indexers.InchiKeyIndexValueMaker",
 },