Custom removal of macro/label in OME metadata

tiledb/bioimg/converters/base.py

@@ -48,6 +48,7 @@
     iter_levels_meta,
     iter_pixel_depths_meta,
     open_bioimg,
+    remove_ome_image_metadata,
     resolve_path,
     validate_ingestion,
 )
@@ -526,6 +527,12 @@ 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

@@ -32,7 +32,12 @@
 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 ImageConverter, ImageReader, ImageWriter
 from .io import as_array

tiledb/bioimg/helpers.py

@@ -3,6 +3,7 @@
 import logging
 import os
 import sys
+import xml.etree.ElementTree as ET
 from pathlib import Path
 from typing import (
     Any,
@@ -14,6 +15,7 @@
     Optional,
     Sequence,
     Tuple,
+    Union,
 )
 from urllib.parse import urlparse
 
@@ -456,3 +458,28 @@ 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]:
+
+    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
+    return ET.tostring(root, encoding="unicode")