test(gooddata-pandas): Arrow path ground truth and parity coverage

Backfill column_totals_indexes into all 36 fixture meta.json files; extend
parity tests to cover all four DataFrameMetadata fields
(row_totals_indexes, column_totals_indexes, primary_labels_from_index,
primary_labels_from_columns) and expand for_arrow_table tests from
4 hand-picked cases to the full fixture set.

risk: nonprod

Author: Martozar

packages/gooddata-pandas/src/gooddata_pandas/__init__.py

@@ -1,7 +1,7 @@
 # (C) 2021 GoodData Corporation
 
 from gooddata_pandas._version import __version__
-from gooddata_pandas.arrow_types import TypesMapper
+from gooddata_pandas.arrow_types import ArrowConfig, TypesMapper
 
 try:
     from gooddata_pandas.arrow_convertor import convert_arrow_table_to_dataframe

packages/gooddata-pandas/src/gooddata_pandas/arrow_types.py

@@ -1,6 +1,7 @@
 # (C) 2026 GoodData Corporation
 from __future__ import annotations
 
+from dataclasses import dataclass, field
 from enum import Enum
 
 
@@ -19,3 +20,35 @@ class TypesMapper(Enum):
     DEFAULT = "default"
     ARROW_STRINGS = "arrow_strings"
     CUSTOM = "custom"
+
+
+@dataclass
+class ArrowConfig:
+    """
+    Arrow IPC conversion configuration for DataFrameFactory.
+
+    Controls *how* Arrow data is converted to a pandas DataFrame. Whether to
+    use the Arrow path at all is set via the ``use_arrow`` parameter on
+    ``DataFrameFactory`` itself.
+
+    Set once on the factory; applies to every Arrow-path call (for_exec_def,
+    for_exec_def_arrow, for_arrow_table, for_exec_result_id).
+
+    Attributes:
+        self_destruct: When True, Arrow buffers are freed during conversion,
+            reducing peak native memory at the cost of not being able to reuse
+            the table after the call. Defaults to False.
+        types_mapper: Controls how Arrow types are mapped to pandas dtypes.
+            TypesMapper.DEFAULT (default) — no mapping; float64 and object strings,
+                identical to the JSON execution path.
+            TypesMapper.ARROW_STRINGS — strings use Arrow-backed StringDtype;
+                all numeric types unchanged.
+            TypesMapper.CUSTOM — uses custom_mapping; raises ValueError if
+                custom_mapping is not provided.
+        custom_mapping: Arrow type → pandas dtype mapping dict. Only used when
+            types_mapper=TypesMapper.CUSTOM, ignored otherwise.
+    """
+
+    self_destruct: bool = False
+    types_mapper: TypesMapper = TypesMapper.DEFAULT
+    custom_mapping: dict | None = field(default=None)