[python] Add implementation of the MultiscaleImage to main (#3112)

Add the Python implementation of the `MultiscaleImage` class. --------- Co-authored-by: nguyenv <vivian@tiledb.com>
single-cell-data · Oct 2, 2024 · 2dab4e9 · 2dab4e9
1 parent eb89c31
commit 2dab4e9
Show file tree

Hide file tree

Showing 11 changed files with 690 additions and 44 deletions.
diff --git a/apis/python/src/tiledbsoma/__init__.py b/apis/python/src/tiledbsoma/__init__.py
@@ -175,6 +175,7 @@
 )
 from ._indexer import IntIndexer, tiledbsoma_build_index
 from ._measurement import Measurement
+from ._multiscale_image import MultiscaleImage
 from ._point_cloud_dataframe import PointCloudDataFrame
 from ._sparse_nd_array import SparseNDArray, SparseNDArrayRead
 from .options import SOMATileDBContext, TileDBCreateOptions, TileDBWriteOptions
@@ -209,6 +210,7 @@
     "get_storage_engine",
     "IntIndexer",
     "Measurement",
+    "MultiscaleImage",
     "NotCreateableError",
     "open",
     "PointCloudDataFrame",

diff --git a/apis/python/src/tiledbsoma/_arrow_types.py b/apis/python/src/tiledbsoma/_arrow_types.py
@@ -74,6 +74,23 @@
     pa.timestamp("ns"): "tsn:",
 }
 
+_CARROW_TO_PYARROW: Dict[pa.DataType, str] = {
+    "c": pa.int8(),
+    "s": pa.int16(),
+    "i": pa.int32(),
+    "l": pa.int64(),
+    "C": pa.uint8(),
+    "S": pa.uint16(),
+    "I": pa.uint32(),
+    "L": pa.uint64(),
+    "f": pa.float32(),
+    "g": pa.float64(),
+    "tss:": pa.timestamp("s"),
+    "tsm:": pa.timestamp("ms"),
+    "tsu:": pa.timestamp("us"),
+    "tsn:": pa.timestamp("ns"),
+}
+
 # Same as _ARROW_TO_TDB_ATTR, but used for DataFrame indexed columns, aka TileDB Dimensions.
 # Any type system differences from the base-case Attr should be added here.
 _ARROW_TO_TDB_DIM: Dict[Any, Union[str, TypeError]] = _ARROW_TO_TDB_ATTR.copy()
@@ -284,3 +301,10 @@ def pyarrow_to_carrow_type(pa_type: pa.DataType) -> str:
         return _PYARROW_TO_CARROW[pa_type]
     except KeyError:
         raise TypeError(f"Invalid pyarrow type {pa_type}") from None
+
+
+def carrow_type_to_pyarrow(ca_type: str) -> pa.DataType:
+    try:
+        return _CARROW_TO_PYARROW[ca_type]
+    except KeyError:
+        raise TypeError(f"Invalid carrrow type {ca_type}") from None
diff --git a/apis/python/src/tiledbsoma/_constants.py b/apis/python/src/tiledbsoma/_constants.py
@@ -9,7 +9,7 @@
 SOMA_JOINID = "soma_joinid"
 SOMA_GEOMETRY = "soma_geometry"
 SOMA_COORDINATE_SPACE_METADATA_KEY = "soma_coordinate_space"
-SOMA_MULTISCALE_IMAGE_SCHAME = "soma_multiscale_image_schema"
+SOMA_MULTISCALE_IMAGE_SCHEMA = "soma_multiscale_image_schema"
 SOMA_OBJECT_TYPE_METADATA_KEY = "soma_object_type"
 SOMA_ENCODING_VERSION_METADATA_KEY = "soma_encoding_version"
 SOMA_ENCODING_VERSION = "1.1.0"

diff --git a/apis/python/src/tiledbsoma/_factory.py b/apis/python/src/tiledbsoma/_factory.py
@@ -28,6 +28,7 @@
     _dense_nd_array,
     _experiment,
     _measurement,
+    _multiscale_image,
     _point_cloud_dataframe,
     _soma_object,
     _sparse_nd_array,
@@ -216,6 +217,7 @@ def _type_name_to_cls(type_name: str) -> Type[AnySOMAObject]:
             _dense_nd_array.DenseNDArray,
             _experiment.Experiment,
             _measurement.Measurement,
+            _multiscale_image.MultiscaleImage,
             _sparse_nd_array.SparseNDArray,
             _point_cloud_dataframe.PointCloudDataFrame,
         )