Exported images fail to render in qupath (#137)

* Custom removal of macro/label in OME metadata

* Unittest the remove_ome_image_metadata function

* PR miscs

Author: ktsitsi

tests/__init__.py

@@ -8,6 +8,7 @@
 from tiledb.bioimg import ATTR_NAME
 from tiledb.cc import WebpInputFormat
 from tiledb.bioimg.helpers import merge_ned_ranges
+import xml.etree.ElementTree as ET
 
 DATA_DIR = Path(__file__).parent / "data"
 
@@ -112,3 +113,96 @@ def generate_test_case(num_axes, num_ranges, max_value):
     expected_output = merge_ned_ranges(input_ranges)
 
     return input_ranges, expected_output
+
+
+def generate_xml(has_macro=True, has_label=True, root_tag="OME", num_images=1):
+    """Generate synthetic XML strings with options to include 'macro' and 'label' images."""
+
+    # Create the root element
+    ome = ET.Element(
+        root_tag,
+        {
+            "xmlns": "http://www.openmicroscopy.org/Schemas/OME/2016-06",
+            "xmlns:xsi": "http://www.w3.org/2001/XMLSchema-instance",
+            "Creator": "tifffile.py 2023.7.4",
+            "UUID": "urn:uuid:40348664-c1f8-11ee-a19b-58112295faaf",
+            "xsi:schemaLocation": "http://www.openmicroscopy.org/Schemas/OME/2016-06 http://www.openmicroscopy.org/Schemas/OME/2016-06/ome.xsd",
+        },
+    )
+
+    # Create an instrument element
+    instrument = ET.SubElement(ome, "Instrument", ID="Instrument:95")
+    objective = ET.SubElement(
+        instrument, "Objective", ID="Objective:95", NominalMagnification="40.0"
+    )
+
+    # Create standard image elements
+    for i in range(num_images):
+        image = ET.SubElement(ome, "Image", ID=f"Image:{i}", Name=f"Image{i}")
+        pixels = ET.SubElement(
+            image,
+            "Pixels",
+            DimensionOrder="XYCZT",
+            ID=f"Pixels:{i}",
+            SizeC="3",
+            SizeT="1",
+            SizeX="86272",
+            SizeY="159488",
+            SizeZ="1",
+            Type="uint8",
+            Interleaved="true",
+            PhysicalSizeX="0.2827",
+            PhysicalSizeY="0.2827",
+        )
+        channel = ET.SubElement(
+            pixels, "Channel", ID=f"Channel:{i}:0", SamplesPerPixel="3"
+        )
+        tiffdata = ET.SubElement(pixels, "TiffData", PlaneCount="1")
+
+    # Conditionally add 'macro' and 'label' images
+    if has_label:
+        label_image = ET.SubElement(ome, "Image", ID="Image:label", Name="label")
+        pixels = ET.SubElement(
+            label_image,
+            "Pixels",
+            DimensionOrder="XYCZT",
+            ID="Pixels:label",
+            SizeC="3",
+            SizeT="1",
+            SizeX="604",
+            SizeY="594",
+            SizeZ="1",
+            Type="uint8",
+            Interleaved="true",
+            PhysicalSizeX="43.0",
+            PhysicalSizeY="43.0",
+        )
+        channel = ET.SubElement(
+            pixels, "Channel", ID="Channel:label:0", SamplesPerPixel="3"
+        )
+        tiffdata = ET.SubElement(pixels, "TiffData", IFD="1", PlaneCount="1")
+
+    if has_macro:
+        macro_image = ET.SubElement(ome, "Image", ID="Image:macro", Name="macro")
+        pixels = ET.SubElement(
+            macro_image,
+            "Pixels",
+            DimensionOrder="XYCZT",
+            ID="Pixels:macro",
+            SizeC="3",
+            SizeT="1",
+            SizeX="604",
+            SizeY="1248",
+            SizeZ="1",
+            Type="uint8",
+            Interleaved="true",
+            PhysicalSizeX="43.0",
+            PhysicalSizeY="43.0",
+        )
+        channel = ET.SubElement(
+            pixels, "Channel", ID="Channel:macro:0", SamplesPerPixel="3"
+        )
+        tiffdata = ET.SubElement(pixels, "TiffData", IFD="2", PlaneCount="1")
+
+    # Convert the ElementTree to a string
+    return ET.tostring(ome, encoding="unicode")

tests/unit/test_helpers.py

@@ -1,3 +1,5 @@
+import xml.etree.ElementTree as ET
+
 import numpy as np
 import pytest
 
@@ -8,9 +10,10 @@
     get_rgba,
     iter_color,
     merge_ned_ranges,
+    remove_ome_image_metadata,
 )
 
-from .. import generate_test_case
+from .. import generate_test_case, generate_xml
 
 
 def test_color_iterator():
@@ -56,3 +59,35 @@ def test_get_pixel_depth():
 def test_validate_ingestion(num_axes, num_ranges, max_value):
     input_ranges, expected_output = generate_test_case(num_axes, num_ranges, max_value)
     assert merge_ned_ranges(input_ranges) == expected_output
+
+
+@pytest.mark.parametrize("macro", [True, False])
+@pytest.mark.parametrize("has_label", [True, False])
+@pytest.mark.parametrize("num_images", [1, 2, 3])
+@pytest.mark.parametrize("root_tag", ["OME", "InvalidRoot"])
+def test_remove_ome_image_metadata(macro, has_label, num_images, root_tag):
+    original_xml_string = generate_xml(
+        has_macro=macro, has_label=has_label, num_images=1, root_tag=root_tag
+    )
+
+    excluded_metadata = remove_ome_image_metadata(original_xml_string)
+    if root_tag == "OME":
+        parsed_excluded = ET.fromstring(excluded_metadata)
+
+        # Assert if "label" subelement is present
+        assert (
+            parsed_excluded.find(
+                ".//{http://www.openmicroscopy.org/Schemas/OME/2016-06}Image[@ID='Image:label'][@Name='label']"
+            )
+            is None
+        )
+
+        # Assert if "macro" subelement is present
+        assert (
+            parsed_excluded.find(
+                ".//{http://www.openmicroscopy.org/Schemas/OME/2016-06}Image[@ID='Image:macro'][@Name='macro']"
+            )
+            is None
+        )
+    else:
+        assert remove_ome_image_metadata(original_xml_string) is None

tiledb/bioimg/converters/base.py

@@ -51,6 +51,7 @@
     iter_levels_meta,
     iter_pixel_depths_meta,
     open_bioimg,
+    remove_ome_image_metadata,
     resolve_path,
     validate_ingestion,
 )
@@ -512,6 +513,12 @@ def to_tiledb(
 
             if not exclude_metadata:
                 original_metadata = reader.original_metadata
+            else:
+                if ome_xml := reader.original_metadata.get("ome_metadata"):
+                    pruned_metadata = remove_ome_image_metadata(ome_xml)
+                    original_metadata = (
+                        {"ome_metadata": pruned_metadata} if pruned_metadata else {}
+                    )
 
         with rw_group:
             rw_group.w_group.meta.update(

tiledb/bioimg/converters/ome_tiff.py

@@ -34,7 +34,12 @@
 from tiledb.highlevel import _get_ctx
 
 from .. import ATTR_NAME, EXPORT_TILE_SIZE, WHITE_RGBA
-from ..helpers import get_decimal_from_rgba, get_logger_wrapper, get_rgba, iter_color
+from ..helpers import (
+    get_decimal_from_rgba,
+    get_logger_wrapper,
+    get_rgba,
+    iter_color,
+)
 from .axes import Axes
 from .base import ImageConverterMixin
 from .io import as_array

tiledb/bioimg/helpers.py

@@ -2,7 +2,9 @@
 
 import logging
 import os
+import re
 import sys
+import xml.etree.ElementTree as ET
 from pathlib import Path
 from typing import (
     Any,
@@ -14,6 +16,7 @@
     Optional,
     Sequence,
     Tuple,
+    Union,
 )
 from urllib.parse import urlparse
 
@@ -456,3 +459,44 @@ def merge_ned_ranges(
     merged_ranges_per_axis = [merge_ranges(ranges) for ranges in ranges_per_axis]
 
     return tuple(merged_ranges_per_axis)
+
+
+def remove_ome_image_metadata(xml_string: str) -> Union[str, Any]:
+    """
+    This functions parses an OME-XML file and removes the `macro` and `label` metadata
+    of the image that it accompanies.
+    :param xml_string: OME-XML string to remove metadata from
+    :return: OME-XML string with metadata removed
+    """
+    if not xml_string.lstrip().startswith("<OME") or not xml_string:
+        return None
+
+    # Parse the XML string
+    root = ET.fromstring(xml_string)
+
+    # Extract the namespace from the root element's tag
+    namespace = root.tag.split("}")[0].strip("{")  # Extract namespace
+    ns = {"ome": namespace}
+
+    # Find all images
+    images = root.findall("ome:Image", ns)
+
+    # Iterate over images and remove those with Name 'macro' or 'label'
+    for image in images:
+        name = image.attrib.get("Name")
+        if name in ["macro", "label"]:
+            root.remove(image)
+
+    # Return the modified XML as a string
+    # Regular expression pattern to match 'ns0', 'ns0:', or ':ns0'
+    pattern = r"ns0:|:ns0|ns0"
+
+    # Substitute the matches with an empty string
+    return re.sub(
+        pattern,
+        "",
+        ET.tostring(
+            root,
+            encoding="unicode",
+        ),
+    )