feat(isthmus)!: configurable SQL parser config via ConverterProvider builder

examples/isthmus-api/src/main/java/io/substrait/examples/CustomDialectDynamicFnToSql.java

@@ -5,6 +5,7 @@
 import io.substrait.expression.Expression;
 import io.substrait.extension.DefaultExtensionCatalog;
 import io.substrait.extension.SimpleExtension;
+import io.substrait.isthmus.ConverterProvider;
 import io.substrait.isthmus.DynamicConverterProvider;
 import io.substrait.isthmus.SubstraitToSql;
 import io.substrait.plan.Plan;
@@ -79,7 +80,8 @@ public void run(final String[] args) {
 
     // Convert the plan to SQL
     final SubstraitToSql substraitToSql =
-        new SubstraitToSql(new DynamicConverterProvider(extensions));
+        new SubstraitToSql(
+            new DynamicConverterProvider(ConverterProvider.builder().extensions(extensions)));
 
     System.out.println("\nWith custom SparkSQL SqlDialect::");
     substraitToSql.convert(plan, customSqlDialect()).stream().forEachOrdered(System.out::println);

examples/isthmus-api/src/main/java/io/substrait/examples/DynamicFnToSql.java

@@ -5,6 +5,7 @@
 import io.substrait.expression.Expression;
 import io.substrait.extension.DefaultExtensionCatalog;
 import io.substrait.extension.SimpleExtension;
+import io.substrait.isthmus.ConverterProvider;
 import io.substrait.isthmus.DynamicConverterProvider;
 import io.substrait.isthmus.SubstraitToSql;
 import io.substrait.plan.Plan;
@@ -77,7 +78,8 @@ public void run(final String[] args) {
 
     // Convert the plan to SQL
     final SubstraitToSql substraitToSql =
-        new SubstraitToSql(new DynamicConverterProvider(extensions));
+        new SubstraitToSql(
+            new DynamicConverterProvider(ConverterProvider.builder().extensions(extensions)));
     System.out.println("\nWith default DuckDB SqlDialect::");
     substraitToSql.convert(plan, SqlDialect.DatabaseProduct.DUCKDB.getDialect()).stream()
         .forEachOrdered(System.out::println);

examples/isthmus-api/src/main/java/io/substrait/examples/FromSql.java

@@ -1,6 +1,7 @@
 package io.substrait.examples;
 
 import io.substrait.examples.IsthmusAppExamples.Action;
+import io.substrait.isthmus.ConverterProvider;
 import io.substrait.isthmus.SqlToSubstrait;
 import io.substrait.isthmus.sql.SubstraitCreateStatementParser;
 import io.substrait.plan.Plan;
@@ -10,8 +11,8 @@
 import java.nio.file.Path;
 import java.nio.file.Paths;
 import java.util.List;
+import org.apache.calcite.avatica.util.Casing;
 import org.apache.calcite.prepare.CalciteCatalogReader;
-import org.apache.calcite.sql.SqlDialect;
 import org.apache.calcite.sql.parser.SqlParseException;
 
 /**
@@ -22,7 +23,7 @@
  * <p>1. Create a fully typed schema for the inputs. Within a SQL context this represents the CREATE
  * TABLE commands, which need to be converted to a Calcite Schema.
  *
- * <p>2. Parse the SQL query to convert (in the source SQL dialect).
+ * <p>2. Parse the SQL query to convert.
  *
  * <p>3. Convert the SQL query to Calcite Relations.
  *
@@ -49,22 +50,27 @@ public void run(final String[] args) {
                   "test_result" varchar(15),"test_mileage" int, "postcode_area" varchar(15));
               """);
 
+      // The unquoted identifier casing applied while parsing is configurable via a
+      // ConverterProvider. The same provider is used for both the schema and the query so that
+      // identifier casing stays consistent end-to-end. Casing.UNCHANGED preserves identifiers as
+      // written, matching the lower-case names used in the CREATE TABLE statements above.
+      final ConverterProvider converterProvider =
+          ConverterProvider.builder().unquotedCasing(Casing.UNCHANGED).build();
+
       final CalciteCatalogReader catalogReader =
-          SubstraitCreateStatementParser.processCreateStatementsToCatalog(createSqlStatements);
+          SubstraitCreateStatementParser.processCreateStatementsToCatalog(
+              converterProvider, createSqlStatements);
 
-      // Query that needs to be converted; again this could be in a variety of SQL
-      // dialects
+      // Query that needs to be converted
       final String sqlQuery =
           """
           SELECT vehicles.colour, count(*) as colourcount FROM vehicles INNER JOIN tests
               ON vehicles.vehicle_id=tests.vehicle_id WHERE tests.test_result = 'P'
               GROUP BY vehicles.colour ORDER BY count(*)
           """;
-      final SqlToSubstrait sqlToSubstrait = new SqlToSubstrait();
 
-      // choose DuckDB as an example dialect
-      final SqlDialect dialect = SqlDialect.DatabaseProduct.DUCKDB.getDialect();
-      final Plan substraitPlan = sqlToSubstrait.convert(sqlQuery, catalogReader, dialect);
+      final SqlToSubstrait sqlToSubstrait = new SqlToSubstrait(converterProvider);
+      final Plan substraitPlan = sqlToSubstrait.convert(sqlQuery, catalogReader);
 
       // Create the proto plan to display to stdout - as it has a better format
       final PlanProtoConverter planToProto = new PlanProtoConverter();

examples/isthmus-api/src/main/java/io/substrait/examples/ToOptimizedSql.java

@@ -40,7 +40,8 @@ public void run(String[] args) {
       final Plan substraitPlan = new ProtoPlanConverter().from(proto);
 
       // Configure Isthmus Utilities
-      final SubstraitToCalcite substraitToCalcite = new SubstraitToCalcite(new ConverterProvider());
+      final SubstraitToCalcite substraitToCalcite =
+          new SubstraitToCalcite(ConverterProvider.DEFAULT);
 
       // Configure Calcite Utilities
       final SqlDialect sqlDialect = SqlDialect.DatabaseProduct.MYSQL.getDialect();

isthmus-cli/src/main/java/io/substrait/isthmus/cli/IsthmusEntryPoint.java

@@ -15,7 +15,6 @@
 import java.util.concurrent.Callable;
 import org.apache.calcite.avatica.util.Casing;
 import org.apache.calcite.prepare.Prepare;
-import org.apache.calcite.sql.parser.SqlParser;
 import picocli.CommandLine;
 import picocli.CommandLine.Command;
 import picocli.CommandLine.Option;
@@ -60,15 +59,6 @@ enum OutputFormat {
       description = "Calcite's casing policy for unquoted identifiers: ${COMPLETION-CANDIDATES}")
   private Casing unquotedCasing = Casing.TO_UPPER;
 
-  private ConverterProvider converterProvider() {
-    return new ConverterProvider() {
-      @Override
-      public SqlParser.Config getSqlParserConfig() {
-        return super.getSqlParserConfig().withUnquotedCasing(unquotedCasing);
-      }
-    };
-  }
-
   /**
    * Standard Java Main method invoked by the isthmus CLI command.
    *
@@ -96,16 +86,17 @@ public static void main(String... args) {
 
   @Override
   public Integer call() throws Exception {
+    ConverterProvider provider = ConverterProvider.builder().unquotedCasing(unquotedCasing).build();
     // Isthmus image is parsing SQL Expression if that argument is defined
     if (sqlExpressions != null) {
-      SqlExpressionToSubstrait converter = new SqlExpressionToSubstrait(converterProvider());
+      SqlExpressionToSubstrait converter = new SqlExpressionToSubstrait(provider);
       ExtendedExpression extendedExpression = converter.convert(sqlExpressions, createStatements);
       printMessage(extendedExpression);
     } else { // by default Isthmus image are parsing SQL Query
-      SqlToSubstrait converter = new SqlToSubstrait(converterProvider());
+      SqlToSubstrait converter = new SqlToSubstrait(provider);
       Prepare.CatalogReader catalog =
           SubstraitCreateStatementParser.processCreateStatementsToCatalog(
-              createStatements.toArray(String[]::new));
+              provider, createStatements);
       Plan plan = new PlanProtoConverter().toProto(converter.convert(sql, catalog));
       printMessage(plan);
     }

isthmus/src/main/java/io/substrait/isthmus/AutomaticDynamicFunctionMappingConverterProvider.java

@@ -47,9 +47,14 @@ public class AutomaticDynamicFunctionMappingConverterProvider extends ConverterP
    *
    * <p>Uses {@link DefaultExtensionCatalog#DEFAULT_COLLECTION} for extensions and {@link
    * SubstraitTypeSystem#TYPE_FACTORY} for type operations.
+   *
+   * @deprecated Use {@link
+   *     #AutomaticDynamicFunctionMappingConverterProvider(ConverterProvider.Builder)} with {@link
+   *     ConverterProvider#builder()} instead.
    */
+  @Deprecated
   public AutomaticDynamicFunctionMappingConverterProvider() {
-    this(DefaultExtensionCatalog.DEFAULT_COLLECTION, SubstraitTypeSystem.TYPE_FACTORY);
+    this(ConverterProvider.builder());
   }
 
   /**
@@ -58,10 +63,15 @@ public AutomaticDynamicFunctionMappingConverterProvider() {
    * <p>Uses {@link SubstraitTypeSystem#TYPE_FACTORY} for type operations.
    *
    * @param extensions the extension collection containing function definitions
+   * @deprecated Use {@link
+   *     #AutomaticDynamicFunctionMappingConverterProvider(ConverterProvider.Builder)} instead, e.g.
+   *     {@code new AutomaticDynamicFunctionMappingConverterProvider(
+   *     ConverterProvider.builder().extensions(extensions))}.
    */
+  @Deprecated
   public AutomaticDynamicFunctionMappingConverterProvider(
       SimpleExtension.ExtensionCollection extensions) {
-    this(extensions, SubstraitTypeSystem.TYPE_FACTORY);
+    this(ConverterProvider.builder().extensions(extensions));
   }
 
   /**
@@ -72,10 +82,27 @@ public AutomaticDynamicFunctionMappingConverterProvider(
    *
    * @param extensions the extension collection containing function definitions
    * @param typeFactory the type factory for creating and managing Calcite data types
+   * @deprecated Use {@link
+   *     #AutomaticDynamicFunctionMappingConverterProvider(ConverterProvider.Builder)} instead, e.g.
+   *     {@code new AutomaticDynamicFunctionMappingConverterProvider(
+   *     ConverterProvider.builder().extensions(extensions).typeFactory(typeFactory))}.
    */
+  @Deprecated
   public AutomaticDynamicFunctionMappingConverterProvider(
       SimpleExtension.ExtensionCollection extensions, RelDataTypeFactory typeFactory) {
-    super(extensions, typeFactory);
+    this(ConverterProvider.builder().extensions(extensions).typeFactory(typeFactory));
+  }
+
+  /**
+   * Creates a new provider from a {@link ConverterProvider.Builder}, seeding base state from the
+   * builder and then installing the automatically generated dynamic function mappings and operator
+   * table. This is the seam through which the other constructors — and subclasses, via {@code
+   * super(builder)} — route.
+   *
+   * @param builder the builder carrying the configured components
+   */
+  public AutomaticDynamicFunctionMappingConverterProvider(ConverterProvider.Builder builder) {
+    super(builder);
 
     List<SqlOperator> dynamicScalarOperators = getDynamicScalarOperators();
     this.scalarFunctionConverter = createScalarFunctionConverter(dynamicScalarOperators);