Extend HCS label parsing (#189)

* extend hcs label parsing

* bugfix

* fix p=1 name fetching

* bugfixes

* split three parts in the numerical labels

also strips leading zeros
so that the FOV names start with '0'
and is compatible with napari-ome-zarr

* fix import sorting

* better ndtiff reading them p_idx=0 missing

* revert hcs position label parsing

* select hcs datasets

* simplify _get_pos_names

---------

Co-authored-by: Ziwen Liu <[email protected]>

iohub/convert.py

@@ -307,16 +307,13 @@ def _get_pos_names(self):
         """Append a list of pos names in ascending order
         (order in which they were acquired).
         """
-        if self.p > 1:
-            self.pos_names = []
-            for p in range(self.p):
-                try:
-                    name = self.reader.stage_positions[p]["Label"]
-                except (IndexError, KeyError):
-                    name = p
-                self.pos_names.append(name)
-        else:
-            self.pos_names = ["0"]
+        self.pos_names = []
+        for p in range(self.p):
+            try:
+                name = self.reader.stage_positions[p]["Label"]
+            except (IndexError, KeyError):
+                name = str(p)
+            self.pos_names.append(name)
 
     def _get_image_array(self, p: int, t: int, c: int, z: int):
         try:

iohub/reader_base.py

@@ -1,3 +1,5 @@
+from typing import Union
+
 import zarr
 from numpy.typing import DTypeLike, NDArray
 
@@ -11,7 +13,7 @@ def __init__(self):
         self.width: int = None
         self.dtype: DTypeLike = None
         self._mm_meta: dict = None
-        self._stage_positions: list[dict] = []
+        self._stage_positions: list[dict[str, Union[str, float]]] = []
         self.z_step_size: float = None
         self.channel_names: list[str] = None
 
@@ -112,24 +114,39 @@ def get_num_positions(self) -> int:
     @property
     def hcs_position_labels(self):
         """Parse plate position labels generated by the HCS position generator,
-        e.g. 'A1-Site_0', and split into row, column, and FOV names.
+        e.g. 'A1-Site_0' or '1-Pos000_000', and split into row, column, and
+        FOV names.
 
         Returns
         -------
         list[tuple[str, str, str]]
-            FOV name paths, e.g. ('A', '1', '0')
+            FOV name paths, e.g. ('A', '1', '0') or ('0', '1', '000000')
         """
         if not self.stage_positions:
             raise ValueError("Stage position metadata not available.")
         try:
+            # Look for "'A1-Site_0', 'H12-Site_1', ... " format
             labels = [
                 pos["Label"].split("-Site_") for pos in self.stage_positions
             ]
             return [(well[0], well[1:], fov) for well, fov in labels]
         except Exception:
-            labels = [pos.get("Label") for pos in self.stage_positions]
-            raise ValueError(
-                "HCS position labels are in the format of "
-                "'A1-Site_0', 'H12-Site_1', ... "
-                f"Got labels {labels}"
-            )
+            try:
+                # Look for "'1-Pos000_000', '2-Pos000_001', ... "
+                # and split into ('1', '000_000'), ...
+                labels = [
+                    pos["Label"].split("-Pos") for pos in self.stage_positions
+                ]
+                # remove underscore from FOV name, i.e. '000_000'
+                # collect all wells in row '0' so output is
+                # ('0', '1', '000000')
+                return [
+                    ("0", col, fov.replace("_", "")) for col, fov in labels
+                ]
+            except Exception:
+                labels = [pos.get("Label") for pos in self.stage_positions]
+                raise ValueError(
+                    "HCS position labels are in the format of "
+                    "'A1-Site_0', 'H12-Site_1', or '1-Pos000_000' "
+                    f"Got labels {labels}"
+                )

tests/conftest.py

@@ -86,6 +86,27 @@ def setup_mm2gamma_ome_tiffs():
     yield test_data, one_folder, rand_folder
 
 
+@pytest.fixture(scope="function")
+def setup_mm2gamma_ome_tiff_hcs():
+    test_data = pjoin(
+        os.getcwd(), ".pytest_temp", "test_data", "MM20_ome-tiffs"
+    )
+
+    subfolders = [
+        f for f in os.listdir(test_data) if os.path.isdir(pjoin(test_data, f))
+    ]
+    # select datasets with multiple positioons; here they all have 4 positions
+    hcs_subfolders = [f for f in subfolders if '4p' in f]
+
+    # specific folder
+    one_folder = pjoin(test_data, hcs_subfolders[0])
+    # random folder
+    rand_folder = pjoin(test_data, random.choice(hcs_subfolders))
+    # return path to unzipped folder containing test images
+    # as well as specific folder paths
+    yield test_data, one_folder, rand_folder
+
+
 @pytest.fixture(scope="function")
 def setup_mm2gamma_ome_tiffs_incomplete():
     """

tests/test_converter.py

@@ -86,16 +86,6 @@ def test_converter_ometiff(
         assert intensity == raw_array.sum()
 
 
-def test_converter_ometiff_hcs_not_available(
-    setup_test_data, setup_mm2gamma_ome_tiffs
-):
-    _, _, data = setup_mm2gamma_ome_tiffs
-    with TemporaryDirectory() as tmp_dir:
-        output = os.path.join(tmp_dir, "converted.zarr")
-        with pytest.raises(ValueError, match="position"):
-            _ = TIFFConverter(data, output, hcs_plate=True)
-
-
 @pytest.fixture(scope="function")
 def mock_hcs_ome_tiff_reader(
     setup_mm2gamma_ome_tiffs, monkeypatch: pytest.MonkeyPatch
@@ -117,6 +107,26 @@ def mock_hcs_ome_tiff_reader(
     return data, expected_ngff_name
 
 
+@pytest.fixture(scope="function")
+def mock_non_hcs_ome_tiff_reader(
+    setup_mm2gamma_ome_tiffs, monkeypatch: pytest.MonkeyPatch
+):
+    all_ometiffs, _, _ = setup_mm2gamma_ome_tiffs
+    # dataset with 4 positions without HCS site names
+    data = os.path.join(all_ometiffs, "mm2.0-20201209_4p_2t_5z_1c_512k_1")
+    mock_stage_positions = [
+        {"Label": "0"},
+        {"Label": "1"},
+        {"Label": "2"},
+        {"Label": "3"},
+    ]
+    monkeypatch.setattr(
+        "iohub.convert.MicromanagerOmeTiffReader.stage_positions",
+        mock_stage_positions,
+    )
+    return data
+
+
 def test_converter_ometiff_mock_hcs(setup_test_data, mock_hcs_ome_tiff_reader):
     data, expected_ngff_name = mock_hcs_ome_tiff_reader
     with TemporaryDirectory() as tmp_dir:
@@ -129,6 +139,28 @@ def test_converter_ometiff_mock_hcs(setup_test_data, mock_hcs_ome_tiff_reader):
             }
 
 
+def test_converter_ometiff_mock_non_hcs(mock_non_hcs_ome_tiff_reader):
+    data = mock_non_hcs_ome_tiff_reader
+    with TemporaryDirectory() as tmp_dir:
+        output = os.path.join(tmp_dir, "converted.zarr")
+        with pytest.raises(ValueError, match="HCS position labels"):
+            TIFFConverter(data, output, hcs_plate=True)
+
+
+def test_converter_ometiff_hcs_numerical(
+    setup_test_data, setup_mm2gamma_ome_tiff_hcs
+):
+    _, data, _ = setup_mm2gamma_ome_tiff_hcs
+    with TemporaryDirectory() as tmp_dir:
+        output = os.path.join(tmp_dir, "converted.zarr")
+        converter = TIFFConverter(data, output, hcs_plate=True)
+        converter.run()
+        with open_ome_zarr(output, mode="r") as plate:
+            for name, _ in plate.positions():
+                for segment in name.split("/"):
+                    assert segment.isdigit()
+
+
 @given(**CONVERTER_TEST_GIVEN)
 @settings(CONVERTER_TEST_SETTINGS)
 def test_converter_ndtiff(