Include chunk shape as a parameter in stream resource for HDF dataset

src/ophyd_async/core/_hdf_dataset.py

@@ -19,6 +19,7 @@ class HDFDataset:
     dtype_numpy: str = ""
     multiplier: int = 1
     swmr: bool = False
+    chunk_size: int | None = None
 
 
 SLICE_NAME = "AD_HDF5_SWMR_SLICE"
@@ -65,6 +66,7 @@ def __init__(
                     "dataset": ds.dataset,
                     "swmr": ds.swmr,
                     "multiplier": ds.multiplier,
+                    "chunk_size": ds.chunk_size,
                 },
                 uid=None,
                 validate=True,

src/ophyd_async/epics/adcore/_core_io.py

@@ -135,4 +135,5 @@ def __init__(self, prefix: str, name="") -> None:
         self.array_size0 = epics_signal_r(int, prefix + "ArraySize0")
         self.array_size1 = epics_signal_r(int, prefix + "ArraySize1")
         self.create_directory = epics_signal_rw(int, prefix + "CreateDirectory")
+        self.num_frames_chunks = epics_signal_rw(int, prefix + "NumFramesChunks")
         super().__init__(prefix, name)

src/ophyd_async/epics/adcore/_hdf_writer.py

@@ -83,6 +83,9 @@ async def open(self, multiplier: int = 1) -> Dict[str, DataKey]:
         self._multiplier = multiplier
         outer_shape = (multiplier,) if multiplier > 1 else ()
 
+        # Determine number of frames that will be saved per HDF chunk
+        frames_per_chunk = await self.hdf.num_frames_chunks.get_value()
+
         # Add the main data
         self._datasets = [
             HDFDataset(
@@ -91,6 +94,7 @@ async def open(self, multiplier: int = 1) -> Dict[str, DataKey]:
                 shape=detector_shape,
                 dtype_numpy=np_dtype,
                 multiplier=multiplier,
+                chunk_size=frames_per_chunk,
             )
         ]
         # And all the scalar datasets

tests/epics/adaravis/test_aravis.py

@@ -123,6 +123,7 @@ async def test_can_collect(
         "dataset": "/entry/data/data",
         "swmr": False,
         "multiplier": 1,
+        "chunk_size": 0,
     }
     assert docs[1][0] == "stream_datum"
     stream_datum = docs[1][1]

tests/epics/adkinetix/test_kinetix.py

@@ -19,6 +19,9 @@ async def test_adkinetix(
     async with DeviceCollector(mock=True):
         test_adkinetix = adkinetix.KinetixDetector("KINETIX:", static_path_provider)
 
+    # Set number of frames per chunk to something reasonable
+    set_mock_value(test_adkinetix.hdf.num_frames_chunks, 10)
+
     return test_adkinetix
 
 
@@ -105,6 +108,7 @@ async def test_can_collect(
         "dataset": "/entry/data/data",
         "swmr": False,
         "multiplier": 1,
+        "chunk_size": 10,
     }
     assert docs[1][0] == "stream_datum"
     stream_datum = docs[1][1]

tests/epics/advimba/test_vimba.py

@@ -117,6 +117,7 @@ async def test_can_collect(
         "dataset": "/entry/data/data",
         "swmr": False,
         "multiplier": 1,
+        "chunk_size": 0,
     }
     assert docs[1][0] == "stream_datum"
     stream_datum = docs[1][1]

tests/fastcs/panda/test_hdf_panda.py

@@ -171,6 +171,7 @@ def assert_resource_document():
                     "dataset": f"/{dataset_name}",
                     "swmr": False,
                     "multiplier": 1,
+                    "chunk_size": None,
                 },
             }
             assert "test-panda.h5" in stream_resource["uri"]

tests/fastcs/panda/test_writer.py

@@ -201,7 +201,12 @@ def assert_resource_document(name, resource_doc):
             "data_key": name,
             "mimetype": "application/x-hdf5",
             "uri": "file://localhost" + str(tmp_path / "mock_panda" / "data.h5"),
-            "parameters": {"dataset": f"/{name}", "swmr": False, "multiplier": 1},
+            "parameters": {
+                "dataset": f"/{name}",
+                "swmr": False,
+                "multiplier": 1,
+                "chunk_size": None,
+            },
         }
         assert "mock_panda/data.h5" in resource_doc["uri"]