Add Support for Ignoring JSON Properties in Schema Comparison in own + vanilla codegen (#589)

* Add jsonignore props to SchemaBuilder, and both vanilla + own codegen

* Use comparison util

Author: li-ukumar

avro-builder/builder/src/main/java/com/linkedin/avroutil1/builder/SchemaBuilder.java

@@ -122,6 +122,12 @@ public static void main(String[] args) throws Exception {
         .defaultsTo("true")
         .describedAs("true/false");
 
+    OptionSpec<String> jsonPropsToIgnoreOpt = parser.accepts(
+            "jsonPropsToIgnore",
+            "Comma-separated JSON property names to ignore during schema equality comparison (e.g., custom props)")
+        .withOptionalArg()
+        .withValuesSeparatedBy(',');
+
     //allow plugins to add CLI options
     for (BuilderPlugin plugin : plugins) {
       plugin.customizeCLI(parser);
@@ -261,6 +267,19 @@ public static void main(String[] args) throws Exception {
       }
     }
 
+    java.util.Set<String> jsonPropsToIgnore = new java.util.HashSet<>();
+    if (options.has(jsonPropsToIgnoreOpt)) {
+      List<String> vals = options.valuesOf(jsonPropsToIgnoreOpt);
+      for (String v : vals) {
+        if (v != null) {
+          String trimmed = v.trim();
+          if (!trimmed.isEmpty()) {
+            jsonPropsToIgnore.add(trimmed);
+          }
+        }
+      }
+    }
+
     //allow plugins to parse and validate their own added options
     for (BuilderPlugin plugin : plugins) {
       plugin.parseAndValidateOptions(options);
@@ -282,7 +301,8 @@ public static void main(String[] args) throws Exception {
         handleAvro702,
         handleUtf8EncodingInPutByIndex,
         skipCodegenIfSchemaOnClasspath,
-        handleUtf8Encoding
+        handleUtf8Encoding,
+        jsonPropsToIgnore
     );
 
     opConfig.validateParameters();

avro-builder/builder/src/main/java/com/linkedin/avroutil1/builder/operations/codegen/CodeGenOpConfig.java

@@ -50,6 +50,8 @@ public class CodeGenOpConfig {
   boolean utf8EncodingPutByIndex;
   boolean skipCodegenIfSchemaOnClasspath;
   boolean enableUtf8Encoding;
+  // names of json properties to ignore during schema equality comparisons (e.g., custom props)
+  java.util.Set<String> jsonPropsToIgnore;
 
   @Deprecated
   public CodeGenOpConfig(
@@ -223,7 +225,8 @@ public CodeGenOpConfig(List<File> inputRoots,
       boolean avro702Handling,
       boolean handleUtf8EncodingInPutByIndex,
       boolean skipCodegenIfSchemaOnClasspath,
-      boolean handleUtf8Encoding) {
+      boolean handleUtf8Encoding,
+      java.util.Set<String> jsonPropsToIgnore) {
     this.inputRoots = inputRoots;
     this.nonImportableSourceRoots = nonImportableSourceRoots;
     this.includeClasspath = includeClasspath;
@@ -240,6 +243,45 @@ public CodeGenOpConfig(List<File> inputRoots,
     this.utf8EncodingPutByIndex = handleUtf8EncodingInPutByIndex;
     this.skipCodegenIfSchemaOnClasspath = skipCodegenIfSchemaOnClasspath;
     this.enableUtf8Encoding = handleUtf8Encoding;
+    this.jsonPropsToIgnore = jsonPropsToIgnore;
+  }
+
+  // Overload preserving original public API (without jsonPropsToIgnore)
+  public CodeGenOpConfig(List<File> inputRoots,
+      List<File> nonImportableSourceRoots,
+      boolean includeClasspath,
+      File outputSpecificRecordClassesRoot,
+      File outputExpandedSchemasRoot,
+      List<File> resolverPath,
+      CodeGenerator generatorType,
+      DuplicateSchemaBehaviour dupBehaviour,
+      List<String> duplicateSchemasToIgnore,
+      StringRepresentation stringRepresentation,
+      StringRepresentation methodStringRepresentation,
+      AvroVersion minAvroVersion,
+      boolean avro702Handling,
+      boolean handleUtf8EncodingInPutByIndex,
+      boolean skipCodegenIfSchemaOnClasspath,
+      boolean handleUtf8Encoding) {
+    this(
+        inputRoots,
+        nonImportableSourceRoots,
+        includeClasspath,
+        outputSpecificRecordClassesRoot,
+        outputExpandedSchemasRoot,
+        resolverPath,
+        generatorType,
+        dupBehaviour,
+        duplicateSchemasToIgnore,
+        stringRepresentation,
+        methodStringRepresentation,
+        minAvroVersion,
+        avro702Handling,
+        handleUtf8EncodingInPutByIndex,
+        skipCodegenIfSchemaOnClasspath,
+        handleUtf8Encoding,
+        java.util.Collections.emptySet()
+    );
   }
 
   /**
@@ -357,6 +399,10 @@ public boolean isUtf8EncodingEnabled() {
     return enableUtf8Encoding;
   }
 
+  public java.util.Set<String> getJsonPropsToIgnore() {
+    return jsonPropsToIgnore;
+  }
+
   private void validateInput(Collection<File> files, String desc) {
     for (File f : files) {
       if (!f.exists()) {

avro-builder/builder/src/main/java/com/linkedin/avroutil1/builder/operations/codegen/own/AvroUtilOperationContextBuilder.java

@@ -7,24 +7,26 @@
 package com.linkedin.avroutil1.builder.operations.codegen.own;
 
 import com.linkedin.avroutil1.builder.operations.OperationContext;
+import com.linkedin.avroutil1.builder.operations.SchemaSet;
 import com.linkedin.avroutil1.builder.operations.codegen.CodeGenOpConfig;
 import com.linkedin.avroutil1.builder.operations.codegen.OperationContextBuilder;
 import com.linkedin.avroutil1.builder.operations.codegen.util.AvscFileFinderUtil;
 import com.linkedin.avroutil1.builder.operations.codegen.vanilla.ClasspathSchemaSet;
 import com.linkedin.avroutil1.builder.operations.codegen.vanilla.ResolverPathSchemaSet;
-import com.linkedin.avroutil1.builder.operations.SchemaSet;
+import com.linkedin.avroutil1.builder.operations.codegen.util.SchemaComparisonUtil;
 import com.linkedin.avroutil1.compatibility.AvroCompatibilityHelper;
 import com.linkedin.avroutil1.compatibility.AvscGenerationConfig;
-import com.linkedin.avroutil1.compatibility.SchemaComparisonConfiguration;
 import com.linkedin.avroutil1.model.AvroNamedSchema;
 import com.linkedin.avroutil1.model.AvroSchema;
 import com.linkedin.avroutil1.model.SchemaOrRef;
 import com.linkedin.avroutil1.parser.avsc.AvroParseContext;
 import com.linkedin.avroutil1.parser.avsc.AvscParseResult;
 import com.linkedin.avroutil1.parser.avsc.AvscParser;
-import com.linkedin.avroutil1.util.ConfigurableAvroSchemaComparator;
-import com.linkedin.avroutil1.writer.avsc.AvscSchemaWriter;
-import com.linkedin.avroutil1.writer.avsc.AvscWriterConfig;
+import org.apache.avro.Schema;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+
 import java.io.File;
 import java.util.Collections;
 import java.util.HashSet;
@@ -34,9 +36,6 @@
 import java.util.StringJoiner;
 import java.util.stream.Collectors;
 import java.util.stream.Stream;
-import org.apache.avro.Schema;
-import org.slf4j.Logger;
-import org.slf4j.LoggerFactory;
 
 
 public class AvroUtilOperationContextBuilder implements OperationContextBuilder {
@@ -122,8 +121,7 @@ public OperationContext buildOperationContext(CodeGenOpConfig config) throws Exc
           if (cpSchema != null) {
             // check if the schema on classpath is the same as the one we are trying to generate
             AvroSchema avroSchemaFromClasspath = (new AvscParser()).parse(cpSchema.toString()).getTopLevelSchema();
-            boolean areEqual = ConfigurableAvroSchemaComparator.equals(avroSchemaFromClasspath, schema,
-                SchemaComparisonConfiguration.STRICT);
+            boolean areEqual = SchemaComparisonUtil.equalsAvroSchema(avroSchemaFromClasspath, schema, config.getJsonPropsToIgnore());
             if (!areEqual) {
               throw new IllegalStateException("Schema with name " + fullName
                   + " is defined in the filesystem and on the classpath, but the two schemas are not equal.");
@@ -155,28 +153,26 @@ public OperationContext buildOperationContext(CodeGenOpConfig config) throws Exc
             System.err.println("WARNING: schema " + fqcn + " found in 2+ places: " + allFilesString);
             break;
           case FAIL_IF_DIFFERENT:
-            String baseSchema = null;
+            AvroNamedSchema baseNamed = null;
             AvscParseResult baseSchemaResult = null;
+            // Use SchemaComparisonUtil to build the comparison configuration
             for (AvscParseResult duplicateParseResult : duplicateEntry.getValue()) {
-              if (baseSchema == null) {
-                baseSchema = new AvscSchemaWriter().generateAvsc(duplicateParseResult.getDefinedSchema(fqcn),
-                    AvscWriterConfig.CORRECT_MITIGATED);
+              AvroNamedSchema currentNamed = duplicateParseResult.getDefinedSchema(fqcn);
+              if (baseNamed == null) {
+                baseNamed = currentNamed;
                 baseSchemaResult = duplicateParseResult;
                 continue;
               }
-              String currSchema = new AvscSchemaWriter().generateAvsc(duplicateParseResult.getDefinedSchema(fqcn),
-                  AvscWriterConfig.CORRECT_MITIGATED);
-
-              // TODO: compare canonical forms when canonicalization work is complete
+              // Compare using SchemaComparisonUtil for consistent schema comparison
+              boolean equal = SchemaComparisonUtil.equalsAvroSchema(baseNamed, currentNamed, config.getJsonPropsToIgnore());
               long baseLineNumber = baseSchemaResult.getDefinedSchema(fqcn).getCodeLocation().getEnd().getLineNumber();
-              long duplicateLineNumber =
-                  duplicateParseResult.getDefinedSchema(fqcn).getCodeLocation().getEnd().getLineNumber();
-              String errorMsg = "schema " + fqcn + " found DIFFERENT in 2+ places: " + baseSchemaResult.getURI() + "#L"
+              long duplicateLineNumber = currentNamed.getCodeLocation().getEnd().getLineNumber();
+              String msg = "schema " + fqcn + " found DIFFERENT in 2+ places: " + baseSchemaResult.getURI() + "#L"
                   + baseLineNumber + " and " + duplicateParseResult.getURI() + "#L" + duplicateLineNumber;
-              if (!baseSchema.equals(currSchema)) {
-                throw new RuntimeException("ERROR: " + errorMsg);
+              if (!equal) {
+                throw new RuntimeException("ERROR: " + msg);
               } else {
-                System.err.println("WARNING: " + errorMsg);
+                System.err.println("WARNING: " + msg);
               }
             }
             break;

avro-builder/builder/src/main/java/com/linkedin/avroutil1/builder/operations/codegen/util/SchemaComparisonUtil.java

@@ -0,0 +1,43 @@
+/*
+ * Copyright 2025 LinkedIn Corp.
+ * Licensed under the BSD 2-Clause License (the "License").
+ * See License in the project root for license information.
+ */
+
+package com.linkedin.avroutil1.builder.operations.codegen.util;
+
+import com.linkedin.avroutil1.compatibility.ConfigurableSchemaComparator;
+import com.linkedin.avroutil1.compatibility.SchemaComparisonConfiguration;
+import com.linkedin.avroutil1.model.AvroSchema;
+import java.util.Set;
+import org.apache.avro.Schema;
+
+/**
+ * Centralized helpers for schema equality checks used across builder flows.
+ */
+public final class SchemaComparisonUtil {
+
+  private SchemaComparisonUtil() { }
+
+  public static SchemaComparisonConfiguration buildConfig(Set<String> jsonPropsToIgnore) {
+    SchemaComparisonConfiguration config = SchemaComparisonConfiguration.STRICT;
+    if (jsonPropsToIgnore != null && !jsonPropsToIgnore.isEmpty()) {
+      config = config.jsonPropNamesToIgnore(jsonPropsToIgnore);
+    }
+    return config;
+  }
+
+  public static boolean equalsApacheSchema(Schema a, Schema b, Set<String> jsonPropsToIgnore) {
+    if (a == null || b == null) {
+      return a == b;
+    }
+    return ConfigurableSchemaComparator.equals(a, b, buildConfig(jsonPropsToIgnore));
+  }
+
+  public static boolean equalsAvroSchema(AvroSchema a, AvroSchema b, Set<String> jsonPropsToIgnore) {
+    if (a == null || b == null) {
+      return a == b;
+    }
+    return com.linkedin.avroutil1.util.ConfigurableAvroSchemaComparator.equals(a, b, buildConfig(jsonPropsToIgnore));
+  }
+}