[Data] New Catalog Abstraction and Unity Catalog Read Consolidation

python/ray/data/BUILD.bazel

@@ -2200,6 +2200,20 @@ py_test(
     ],
 )
 
+py_test(
+    name = "test_catalog",
+    size = "small",
+    srcs = ["tests/test_catalog.py"],
+    tags = [
+        "exclusive",
+        "team:data",
+    ],
+    deps = [
+        ":conftest",
+        "//:ray_lib",
+    ],
+)
+
 py_test(
     name = "test_iter_jax",
     size = "small",

python/ray/data/__init__.py

@@ -83,6 +83,12 @@
     read_videos,
     read_webdataset,
 )
+from ray.data.catalog import (
+    Catalog,
+    ReaderFormat,
+    ResolvedSource,
+    UnityCatalog,
+)
 
 # Module-level cached global functions for callable classes. It needs to be defined here
 # since it has to be process-global across cloudpickled funcs.
@@ -192,6 +198,10 @@
     "read_unity_catalog",
     "read_videos",
     "read_webdataset",
+    "Catalog",
+    "ReaderFormat",
+    "ResolvedSource",
+    "UnityCatalog",
     "KafkaAuthConfig",
     "Preprocessor",
 ]