Merge pull request #7 from psilberk/add-oracle-store

Code clean-up, copyright and license.

Author: psilberk

data/semantickernel-data-jdbc/src/main/java/com/microsoft/semantickernel/data/jdbc/JDBCVectorStoreRecordCollection.java

@@ -195,7 +195,7 @@ protected String getKeyFromRecord(Record data) {
             Field keyField = data.getClass()
                 .getDeclaredField(recordDefinition.getKeyField().getName());
             keyField.setAccessible(true);
-            return (String) keyField.get(data);
+            return keyField.get(data).toString();
         } catch (NoSuchFieldException | IllegalAccessException e) {
             throw new SKException("Failed to get key from record", e);
         }

data/semantickernel-data-oracle/pom.xml

@@ -43,6 +43,17 @@
             <artifactId>jackson-core</artifactId>
             <scope>compile</scope>
         </dependency>
+        <dependency>
+            <groupId>com.fasterxml.jackson.datatype</groupId>
+            <artifactId>jackson-datatype-jsr310</artifactId>
+            <version>2.18.0</version>
+        </dependency>
+<!--
+        <dependency>
+            <groupId>com.github.spotbugs</groupId>
+            <artifactId>spotbugs-annotations</artifactId>
+        </dependency>
+-->
         <dependency>
             <groupId>com.oracle.database.jdbc</groupId>
             <artifactId>ojdbc11</artifactId>
@@ -53,6 +64,7 @@
             <artifactId>ojdbc-provider-jackson-oson</artifactId>
             <version>1.0.4</version>
         </dependency>
+        <!-- Tests use TestContainers to create an Oracle Database -->
         <dependency>
             <groupId>org.junit.jupiter</groupId>
             <artifactId>junit-jupiter</artifactId>
@@ -63,7 +75,6 @@
             <artifactId>junit-jupiter-api</artifactId>
             <scope>test</scope>
         </dependency>
-        <!-- Tests use TestContainers to create an Oracle Database -->
         <dependency>
             <groupId>org.testcontainers</groupId>
             <artifactId>testcontainers</artifactId>

data/semantickernel-data-oracle/src/main/java/com/microsoft/semantickernel/data/jdbc/oracle/OracleDataTypesMapping.java

@@ -1,22 +1,74 @@
+/*
+ ** Semantic Kernel Oracle connector version 1.0.
+ **
+ ** Copyright (c) 2025 Oracle and/or its affiliates.
+ ** Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
+ */
 package com.microsoft.semantickernel.data.jdbc.oracle;
 
 /**
- * Defines oracle database type constants for supported field types.
+ * Defines oracle database type constants for supported java types.
  */
 public class OracleDataTypesMapping {
-    public static final String STRING_VARCHAR = "NVARCHAR2(%s)";
+
+    /**
+     * Oracle database type used when strings are mapped to VARCHAR
+     */
+    public static final String STRING_VARCHAR = "VARCHAR2(%s)";
+    /**
+     * Oracle database type used when strings are mapped to CLOB
+     */
     public static final String STRING_CLOB = "CLOB";
+    /**
+     * Oracle database type used to map booleans
+     */
     public static final String BOOLEAN = "BOOLEAN";
+    /**
+     * Oracle database type used to map bytes
+     */
     public static final String BYTE = "NUMBER(3)";
+    /**
+     * Oracle database type used to map byte arrays
+     */
     public static final String BYTE_ARRAY = "RAW(2000)";
+    /**
+     * Oracle database type used to map shorts
+     */
     public static final String SHORT = "NUMBER(5)";
+    /**
+     * Oracle database type used to map ints
+     */
     public static final String INTEGER = "NUMBER(10)";
+    /**
+     * Oracle database type used to map longs
+     */
     public static final String LONG = "NUMBER(19)";
+    /**
+     * Oracle database type used to map float
+     */
     public static final String FLOAT = "BINARY_FLOAT";
+    /**
+     * Oracle database type used to map double
+     */
     public static final String DOUBLE = "BINARY_DOUBLE";
+    /**
+     * Oracle database type used to map BigDecimal
+     */
     public static final String DECIMAL = "NUMBER";
+    /**
+     * Oracle database type used to map offset date time
+     */
     public static final String OFFSET_DATE_TIME = "TIMESTAMP(7) WITH TIME ZONE";
-    public static final String UUID = "RAW(16)";
+    /**
+     * Oracle database type used to map UUID
+     */
+    public static final String UUID = "VARCHAR2(36)";
+    /**
+     * Oracle database type used to map lists
+     */
     public static final String JSON = "JSON";
+    /**
+     * Oracle database type used to map vectors (the parameter is the dimension of the vector)
+     */
     public static final String VECTOR_FLOAT = "VECTOR(%s, FLOAT32)";
 }

data/semantickernel-data-oracle/src/main/java/com/microsoft/semantickernel/data/jdbc/oracle/OracleVectorStoreFieldHelper.java

@@ -1,7 +1,14 @@
+/*
+ ** Semantic Kernel Oracle connector version 1.0.
+ **
+ ** Copyright (c) 2025 Oracle and/or its affiliates.
+ ** Licensed under the Universal Permissive License v 1.0 as shown at https://oss.oracle.com/licenses/upl/
+ */
 package com.microsoft.semantickernel.data.jdbc.oracle;
 
 import com.microsoft.semantickernel.data.jdbc.oracle.OracleVectorStoreQueryProvider.StringTypeMapping;
 import com.microsoft.semantickernel.data.vectorstorage.definition.VectorStoreRecordDataField;
+import com.microsoft.semantickernel.data.vectorstorage.definition.VectorStoreRecordField;
 import com.microsoft.semantickernel.data.vectorstorage.definition.VectorStoreRecordKeyField;
 import com.microsoft.semantickernel.data.vectorstorage.definition.VectorStoreRecordVectorField;
 import oracle.jdbc.OracleTypes;
@@ -16,24 +23,22 @@
 import java.util.stream.Collectors;
 
 /**
- * Helper class for field operations.
+ * Helper class for field operations. Handles mapping between field java types to DB types and
+ * generating SQL statement to create field indexes.
  */
-public class OracleVectorStoreFieldHelper {
-    private static final Logger LOGGER = Logger.getLogger(OracleVectorStoreQueryProvider.class.getName());
+class OracleVectorStoreFieldHelper {
+
+    /**
+     * The logger
+     */
+    private static final Logger LOGGER = Logger.getLogger(OracleVectorStoreFieldHelper.class.getName());
 
     /**
      * Maps supported key java classes to Oracle database types
      */
     private static final HashMap<Class<?>, String> supportedKeyTypes = new HashMap() {
         {
             put(String.class, String.format(OracleDataTypesMapping.STRING_VARCHAR, 255));
-            put(short.class, OracleDataTypesMapping.SHORT);
-            put(Short.class, OracleDataTypesMapping.SHORT);
-            put(int.class, OracleDataTypesMapping.INTEGER);
-            put(Integer.class, OracleDataTypesMapping.INTEGER);
-            put(long.class, OracleDataTypesMapping.LONG);
-            put(Long.class, OracleDataTypesMapping.LONG);
-            put(UUID .class, OracleDataTypesMapping.UUID);
         }
     };
 
@@ -47,14 +52,6 @@ public class OracleVectorStoreFieldHelper {
             put(Collection.class, OracleDataTypesMapping.VECTOR_FLOAT);
             put(float[].class, OracleDataTypesMapping.VECTOR_FLOAT);
             put(Float[].class, OracleDataTypesMapping.VECTOR_FLOAT);
-/*
-            put(byte[].class,"VECTOR(%s, INT8)");
-            put(Byte[].class,"VECTOR(%s, INT8)");
-            put(double[].class,"VECTOR(%s, FLOAT64)");
-            put(Double[].class,"VECTOR(%s, FLOAT64)");
-            put(boolean[].class,"VECTOR(%s, BINARY)");
-            put(Boolean[].class,"VECTOR(%s, BINARY)");
- */
         }
     };
 
@@ -83,26 +80,12 @@ public class OracleVectorStoreFieldHelper {
             put(byte[].class, OracleDataTypesMapping.BYTE_ARRAY);
             put(List.class, OracleDataTypesMapping.JSON);
         }
-
     };
 
     /**
-     * Maps vector type to OracleTypes. Only needed if types other than FLOAT_32 are supported.
+     * Suffix added to the effective column name to generate the index name for a vector column.
      */
-    private static final Map<Class<?>, Integer> mapOracleTypeToVector = new HashMap() {
-        {
-            put(float[].class, OracleTypes.VECTOR_FLOAT32);
-            put(Float[].class, OracleTypes.VECTOR_FLOAT32);
-/*
-            put(byte[].class, OracleTypes.VECTOR_INT8);
-            put(Byte[].class, OracleTypes.VECTOR_INT8);
-            put(Double[].class, OracleTypes.VECTOR_FLOAT64);
-            put(double[].class, OracleTypes.VECTOR_FLOAT64);
-            put(Boolean[].class, OracleTypes.VECTOR_BINARY);
-            put(boolean[].class, OracleTypes.VECTOR_BINARY);
-*/
-        }
-    };
+    public static final String VECTOR_INDEX_SUFFIX = "_VECTOR_INDEX";
 
     /**
      * Gets the mapping between the supported Java key types and the Oracle database type.
@@ -120,12 +103,11 @@ public static HashMap<Class<?>, String> getSupportedKeyTypes() {
      */
     public static Map<Class<?>, String> getSupportedDataTypes(
         StringTypeMapping stringTypeMapping, int defaultVarCharLength) {
-
-        if (stringTypeMapping.equals(StringTypeMapping.USE_VARCHAR)) {
-            supportedDataTypes.put(String.class, String.format(OracleDataTypesMapping.STRING_VARCHAR, defaultVarCharLength));
-        } else {
-            supportedDataTypes.put(String.class, OracleDataTypesMapping.STRING_CLOB);
-        }
+        String stringType = stringTypeMapping.equals(StringTypeMapping.USE_VARCHAR)
+            ? String.format(OracleDataTypesMapping.STRING_VARCHAR, defaultVarCharLength)
+            : OracleDataTypesMapping.STRING_CLOB;
+        supportedDataTypes.put(String.class, stringType);
+        LOGGER.finest("Mapping String columns to " + stringType);
         return supportedDataTypes;
     }
 
@@ -176,14 +158,14 @@ public static String getCreateVectorIndexStatement(VectorStoreRecordVectorField
      */
     public static String createIndexForDataField(String collectionTableName, VectorStoreRecordDataField dataField, Map<Class<?>, String> supportedDataTypes) {
         if (supportedDataTypes.get(dataField.getFieldType()) == "JSON") {
-            String dataFieldIndex = "CREATE MULTIVALUE INDEX %s ON %s t (t.%s.%s)";
+            String dataFieldIndex = "CREATE MULTIVALUE INDEX IF NOT EXISTS %s ON %s t (t.%s.%s)";
             return String.format(dataFieldIndex,
                 collectionTableName + "_" + dataField.getEffectiveStorageName(),
                 collectionTableName,
                 dataField.getEffectiveStorageName(),
                 getFunctionForType(supportedDataTypes.get(dataField.getFieldSubType())));
         }  else {
-            String dataFieldIndex = "CREATE INDEX %s ON %s (%s ASC)";
+            String dataFieldIndex = "CREATE INDEX IF NOT EXISTS %s ON %s (%s ASC)";
             return String.format(dataFieldIndex,
                 collectionTableName + "_" + dataField.getEffectiveStorageName(),
                 collectionTableName,
@@ -192,6 +174,50 @@ public static String createIndexForDataField(String collectionTableName, VectorS
         }
     }
 
+    /**
+     * Returns vector columns names and types for CREATE TABLE statement
+     * @param fields list of vector record fields.
+     * @return comma separated list of columns and types for CREATE TABLE statement.
+     */
+    public static String getVectorColumnNamesAndTypes(List<VectorStoreRecordVectorField> fields) {
+        List<String> columns = fields.stream()
+            .map(field -> field.getEffectiveStorageName() + " " +
+                OracleVectorStoreFieldHelper.getTypeForVectorField(field)
+            ).collect(Collectors.toList());
+
+        return String.join(", ", columns);
+    }
+
+    /**
+     * Returns key column names and type for key column for CREATE TABLE statement
+     * @param field the key field.
+     * @return column name and type of the key field for CREATE TABLE statement.
+     */
+    public static String getKeyColumnNameAndType(VectorStoreRecordKeyField field) {
+        return field.getEffectiveStorageName() + " " + supportedKeyTypes.get(field.getFieldType());
+    }
+
+
+    /**
+     * Generates the index name given the field name. by suffixing "_VECTOR_INDEX" to the field name.
+     * @param effectiveStorageName the field name.
+     * @return the index name.
+     */
+    private static String getIndexName(String effectiveStorageName) {
+        return effectiveStorageName + VECTOR_INDEX_SUFFIX;
+    }
+
+    /**
+     * Gets the type of the vector given the field definition. This method is not needed if only
+     *
+     * @param field the vector field definition.
+     * @return returns the type of vector for the given field type.
+     */
+    private static String getTypeForVectorField(VectorStoreRecordVectorField field) {
+        String dimension = field.getDimensions() > 0 ? String.valueOf(field.getDimensions()) : "*";
+        return String.format(supportedVectorTypes.get(field.getFieldType()), dimension);
+    }
+
     /**
      * Gets the function that allows to return the function that converts the JSON value to the
      * data type.
@@ -218,90 +244,4 @@ private static String getFunctionForType(String jdbcType) {
         }
     }
 
-    /**
-     * Gets the type of the vector given the field definition. This method is not needed if only
-     *
-     * @param field the vector field definition.
-     * @return returns the type of vector for the given field type.
-     */
-    public static String getTypeForVectorField(VectorStoreRecordVectorField field) {
-        String dimension = field.getDimensions() > 0 ? String.valueOf(field.getDimensions()) : "*";
-        return String.format(supportedVectorTypes.get(field.getFieldType()), dimension);
-/* Not needed since all types are FLOAT32
-        if (field.getFieldSubType() != null) {
-            String vectorType;
-            switch (field.getFieldSubType().getName()) {
-                case "java.lang.Double":
-                    vectorType = "FLOAT64";
-                    break;
-                case "java.lang.Byte":
-                    vectorType = "INT8";
-                    break;
-                case "java.lang.Boolean":
-                    vectorType = "BINARY";
-                    break;
-                default:
-                    vectorType = "FLOAT32";
-            }
-            return String.format(supportedVectorTypes.get(field.getFieldType()), dimension, vectorType);
-        } else {
-            return String.format(supportedVectorTypes.get(field.getFieldType()), dimension);
-        }
- */
-    }
-
-    /**
-     * Gets the JDBC oracle of the vector field definition.
-     * @param field the vector field definition.
-     * @return the JDBC oracle type.
-     */
-    public static int getOracleTypeForField(VectorStoreRecordVectorField field) {
-        if (field.getFieldSubType() == null) {
-            return mapOracleTypeToVector.get(field.getFieldType()).intValue();
-        } else {
-            switch (field.getFieldSubType().getName()) {
-                case "java.lang.Double":
-                    return OracleTypes.VECTOR_FLOAT64;
-                case "java.lang.Byte":
-                    return OracleTypes.VECTOR_INT8;
-                case "java.lang.Boolean":
-                    return OracleTypes.VECTOR_BINARY;
-                default:
-                    return OracleTypes.VECTOR_FLOAT32;
-            }
-        }
-    }
-
-    /**
-     * Generates the index name given the field name. by suffixing "_VECTOR_INDEX" to the field name.
-     * @param effectiveStorageName the field name.
-     * @return the index name.
-     */
-    private static String getIndexName(String effectiveStorageName) {
-        return effectiveStorageName + "_VECTOR_INDEX";
-    }
-
-    /**
-     * Returns vector columns names and types for CREATE TABLE statement
-     * @param fields list of vector record fields.
-     * @return comma separated list of columns and types for CREATE TABLE statement.
-     */
-    public static String getVectorColumnNamesAndTypes(List<VectorStoreRecordVectorField> fields) {
-        List<String> columns = fields.stream()
-            .map(field -> field.getEffectiveStorageName() + " " +
-                OracleVectorStoreFieldHelper.getTypeForVectorField(field)
-            ).collect(Collectors.toList());
-
-        return String.join(", ", columns);
-    }
-
-    /**
-     * Returns key column names and type for key column for CREATE TABLE statement
-     * @param field the key field.
-     * @return column name and type of the key field for CREATE TABLE statement.
-     */
-    public static String getKeyColumnNameAndType(VectorStoreRecordKeyField field) {
-        return field.getEffectiveStorageName() + " " + supportedKeyTypes.get(field.getFieldType());
-    }
-
 }