Add Spark 4.0 support via deequ:2.0.14-spark-4.0

[m-aciek]

The upstream deequ library released 2.0.14-spark-4.0 on March 23, 2026, adding official Apache Spark 4.0 support (see awslabs/deequ#676, awslabs/deequ#678).

pydeequ currently does not support Spark 4 because:

1. SPARK_TO_DEEQU_COORD_MAPPING in pydeequ/configs.py only maps up to Spark 3.5
2. The PySpark optional dependency in pyproject.toml is capped at <3.4.0
3. Spark 4 uses Scala 2.13, which removed scala.collection.JavaConversions and changed how Seq.empty is accessed via reflection — both used in pydeequ internals

Required changes

• pydeequ/configs.py: add "4.0": "com.amazon.deequ:deequ:2.0.14-spark-4.0" to SPARK_TO_DEEQU_COORD_MAPPING
• pyproject.toml: widen pyspark optional dep from >=2.4.7,<3.4.0 to >=2.4.7,<5.0.0
• pydeequ/scala_utils.py: replace removed JavaConversions with JavaConverters (iterableAsScalaIterableConverter, mapAsJavaMapConverter)
• pydeequ/profiles.py: same JavaConversions → JavaConverters fix
• pydeequ/analyzers.py + pydeequ/checks.py: replace scala.collection.Seq.empty() (inaccessible via Py4J in Scala 2.13) with an empty java.util.ArrayList converted via to_scala_seq
• .github/workflows/base.yml: add Spark 4.0.0 to the test matrix with Java 17 (required by Spark 4); use include: matrix style to pair each Spark version with its Java version

Additional context

• deequ Spark 4 feature request: [FEATURE] Support for Spark 4 deequ#670
• Maven artifact: com.amazon.deequ:deequ:2.0.14-spark-4.0
• Spark 4 requires Java 17 (not Java 11 used for Spark 3.x)
• A working implementation with all 99 tests passing on Spark 4.0.0 is available as a PR

---

Issue authored with assistance from Claude Code