GH-48254: [Python][Parquet] Support extension types in read_schema

Author: Kuinox

python/pyarrow/parquet/core.py

@@ -262,9 +262,9 @@ class ParquetFile:
     page_checksum_verification : bool, default False
         If True, verify the checksum for each page read from the file.
     arrow_extensions_enabled : bool, default True
-        If True, read Parquet logical types as Arrow extension types where possible,
-        (e.g., read JSON as the canonical `arrow.json` extension type or UUID as
-        the canonical `arrow.uuid` extension type).
+        If True, read Parquet logical types as Arrow extension types where
+        possible (e.g., read JSON as the canonical `arrow.json` extension type
+        or UUID as the canonical `arrow.uuid` extension type).
 
     Examples
     --------
@@ -2347,6 +2347,10 @@ def read_metadata(where, memory_map=False, decryption_properties=None,
         If nothing passed, will be inferred based on path.
         Path will try to be found in the local on-disk filesystem otherwise
         it will be parsed as an URI to determine the filesystem.
+    arrow_extensions_enabled : bool, default True
+        If True, read Parquet logical types as Arrow extension types where
+        possible (e.g. UUID as the canonical `arrow.uuid` extension type).
+        If False, use the underlying storage types instead.
 
     Returns
     -------
@@ -2382,7 +2386,7 @@ def read_metadata(where, memory_map=False, decryption_properties=None,
 
 
 def read_schema(where, memory_map=False, decryption_properties=None,
-                filesystem=None):
+                filesystem=None, arrow_extensions_enabled=True):
     """
     Read effective Arrow schema from Parquet file metadata.
 
@@ -2397,6 +2401,9 @@ def read_schema(where, memory_map=False, decryption_properties=None,
         If nothing passed, will be inferred based on path.
         Path will try to be found in the local on-disk filesystem otherwise
         it will be parsed as an URI to determine the filesystem.
+    arrow_extensions_enabled : bool, default True
+        If True, read Parquet logical types as Arrow extension types where
+        possible (e.g., UUID as the canonical `arrow.uuid` extension type).
 
     Returns
     -------
@@ -2422,11 +2429,15 @@ def read_schema(where, memory_map=False, decryption_properties=None,
 
     with file_ctx:
         file = ParquetFile(
-            where, memory_map=memory_map,
-            decryption_properties=decryption_properties)
+            where,
+            memory_map=memory_map,
+            decryption_properties=decryption_properties,
+            arrow_extensions_enabled=arrow_extensions_enabled,
+        )
+        if arrow_extensions_enabled:
+            return file.schema_arrow
         return file.schema.to_arrow_schema()
 
-
 __all__ = (
     "ColumnChunkMetaData",
     "ColumnSchema",