Fix selectSparkVersion() to use contains() instead of equals() (#504)

## What changes are proposed in this pull request?

This PR fixes a bug in the `selectSparkVersion()` method in
`ClustersExt.java` where spark version matching doesn't work with real
Databricks Runtime version names.

The current `equals()` implementation fails because real Databricks
Runtime version names contain additional information. For example, the
actual version name is `"13.3 LTS (includes Apache Spark 3.4.1, Scala
2.12)"`, not just `"Apache Spark 3.4.1"`. Both the Go SDK
(`strings.Contains()`) and Python SDK (`in` operator) use substring
matching for this functionality.

Originally reported in PR #229 with real API response data.

## How is this tested?

Added a focused unit test `sparkVersionWithSparkVersionParameter()` that
demonstrates the fix works with realistic API response data. The test
uses a version name in the actual format returned by the Databricks API:
`"13.3 LTS (includes Apache Spark 3.4.1, Scala 2.12)"`.

Fixes #229

---------

Co-authored-by: Claude <[email protected]>

Author: mgyucht

NEXT_CHANGELOG.md

@@ -6,6 +6,8 @@
 
 ### Bug Fixes
 
+* Fixed `selectSparkVersion()` method to use contains() instead of equals() for spark version matching.
+
 ### Documentation
 
 ### Internal Changes

databricks-sdk-java/src/main/java/com/databricks/sdk/mixin/ClustersExt.java

@@ -50,7 +50,7 @@ public String selectSparkVersion(SparkVersionSelector selector) throws IllegalAr
         matches = version.getName().contains("LTS") || version.getKey().contains("-esr-");
       }
       if (matches && selector.sparkVersion != null) {
-        matches = ("Apache Spark " + selector.sparkVersion).equals(version.getName());
+        matches = version.getName().contains("Apache Spark " + selector.sparkVersion);
       }
       if (matches) {
         versions.add(version.getKey());

databricks-sdk-java/src/test/java/com/databricks/sdk/mixin/ClustersExtTest.java

@@ -118,4 +118,29 @@ void nullComparisonTest() {
     String nodeType = clustersExt.selectNodeType(new NodeTypeSelector().withLocalDisk());
     assertEquals("testId1", nodeType);
   }
+
+  private GetSparkVersionsResponse testGetSparkVersionsWithSparkVersion() {
+    Collection<SparkVersion> versions = new ArrayList<>();
+    // Mock realistic Databricks Runtime version based on actual API response format
+    // The key point: version name contains more than just "Apache Spark X.Y.Z"
+    versions.add(
+        new SparkVersion()
+            .setName("13.3 LTS (includes Apache Spark 3.4.1, Scala 2.12)")
+            .setKey("13.3.x-scala2.12"));
+    return new GetSparkVersionsResponse().setVersions(versions);
+  }
+
+  @Test
+  void sparkVersionWithSparkVersionParameter() {
+    ClustersExt clustersExt = new ClustersExt(clustersMock);
+    Mockito.doReturn(testGetSparkVersionsWithSparkVersion()).when(clustersMock).sparkVersions();
+
+    // Test that sparkVersion parameter works with realistic API response format
+    // This tests the contains() fix - the version name is "13.3 LTS (includes Apache Spark 3.4.1,
+    // Scala 2.12)"
+    // not just "Apache Spark 3.4.1", so equals() would fail but contains() works
+    String sparkVersion =
+        clustersExt.selectSparkVersion(new SparkVersionSelector().withSparkVersion("3.4.1"));
+    assertEquals("13.3.x-scala2.12", sparkVersion);
+  }
 }