Move the GridKind and Index documentation directly into the docs

Previously they documented actual values in the code, but now that the
generic parameters are more ephemeral, there was nothing to document.

Author: mx-moth

docs/api/conventions/interface.rst

@@ -7,6 +7,8 @@ Convention interface
 .. contents::
    :local:
 
+.. currentmodule:: emsarray.conventions
+
 Each supported convention implements
 the :class:`~emsarray.conventions.Convention` interface.
 
@@ -25,8 +27,37 @@ the :class:`~emsarray.conventions.Convention` interface.
 .. autoclass:: emsarray.conventions.SpatialIndexItem
    :members:
 
-.. autodata:: emsarray.conventions._base.GridKind
-.. autodata:: emsarray.conventions._base.Index
+.. type:: GridKind
+
+    Some type that can enumerate the different :ref:`grid types <grids>`
+    present in a dataset.
+    This can be an :class:`enum.Enum` listing each different kind of grid.
+
+    :type:`Index` values will be included in the feature properties
+    of exported geometry from :mod:`emsarray.operations.geometry`.
+    If the index type includes the grid kind,
+    the grid kind needs to be JSON serializable.
+    The easiest way to achieve this is to make your GridKind type subclass :class:`str`:
+
+    .. code-block:: python
+
+        class MyGridKind(str, enum.Enum):
+            face = 'face'
+            edge = 'edge'
+            node = 'node'
+
+    For cases where the convention only supports a single grid,
+    a singleton enum can be used.
+
+    More esoteric cases involving datasets with a potentially unbounded numbers of grids
+    can use a type that supports this instead.
+
+.. type:: Index
+
+    An :ref:`index <indexing>` to a specific point on a grid in this convention.
+    For conventions with :ref:`multiple grids <grids>` (e.g. cells, edges, and nodes),
+    this should be a tuple whos first element is :type:`.GridKind`.
+    For conventions with a single grid, :type:`.GridKind` is not required.
 
 .. autoclass:: emsarray.conventions.Specificity
    :members:

src/emsarray/conventions/_base.py

@@ -38,37 +38,6 @@
 logger = logging.getLogger(__name__)
 
 
-#: Some type that can enumerate the different :ref:`grid types <grids>`
-#: present in a dataset.
-#: This can be an :class:`enum.Enum` listing each different kind of grid.
-#:
-#: :data:`Index` values will be included in the feature properties
-#: of exported geometry from :mod:`emsarray.operations.geometry`.
-#: If the index type includes the grid kind,
-#: the grid kind needs to be JSON serializable.
-#: The easiest way to achieve this is to make your GridKind type subclass :class:`str`:
-#:
-#: .. code-block:: python
-#:
-#:     class MyGridKind(str, enum.Enum):
-#:         face = 'face'
-#:         edge = 'edge'
-#:         node = 'node'
-#:
-#: For cases where the convention only supports a single grid,
-#: a singleton enum can be used.
-#:
-#: More esoteric cases involving datasets with a potentially unbounded numbers of grids
-#: can use a type that supports this instead.
-# GridKind = TypeVar("GridKind")
-
-#: An :ref:`index <indexing>` to a specific point on a grid in this convention.
-#: For conventions with :ref:`multiple grids <grids>` (e.g. cells, edges, and nodes),
-#: this should be a tuple whos first element is :data:`.GridKind`.
-#: For conventions with a single grid, :data:`.GridKind` is not required.
-# Index = TypeVar("Index")
-
-
 @dataclasses.dataclass
 class SpatialIndexItem[Index]:
     """Information about an item in the :class:`~shapely.strtree.STRtree`