Add XML based vtk datatype

lib/galaxy/datatypes/constructive_solid_geometry.py

@@ -29,6 +29,8 @@
     FilePrefix,
 )
 from galaxy.datatypes.tabular import Tabular
+from galaxy.datatypes.xml import GenericXml
+from xml.etree import ElementTree as ET
 
 if TYPE_CHECKING:
     from io import TextIOBase
@@ -813,3 +815,63 @@ def get_next_line(fh):
         # Discard the rest of the line
         fh.readline()
     return line.strip()
+
+class Vtu(GenericXml):
+    """Format for defining VTU (VTK Unstructured Grid) data https://docs.vtk.org/"""
+
+    edam_data = "edam:data_3671"
+    edam_format = "edam:format_3621"
+    file_ext = "vtu"
+
+    MetadataElement(name="file_format", default=None, desc="File format", readonly=True, optional=True, visible=True)
+    MetadataElement(name="num_cells", desc="Number of cells in the mesh", param_type="int", optional=True)
+    MetadataElement(name="num_points", desc="Number of points in the mesh", param_type="int", optional=True)
+
+    def set_meta(self, dataset: DatasetProtocol, **kwd) -> None:
+        """Set metadata for VTU files, including number of cells and points."""
+        try:
+            with open(dataset.file_name, "r") as file:
+                tree = ET.parse(file)
+                root = tree.getroot()
+
+                # Suche nach dem UnstructuredGrid-Knoten
+                unstructured_grid = root.find(".//UnstructuredGrid")
+                if unstructured_grid is not None:
+                    # Suche nach Zell- und Punktinformationen
+                    piece = unstructured_grid.find(".//Piece")
+                    if piece is not None:
+                        num_cells = piece.get("NumberOfCells")
+                        num_points = piece.get("NumberOfPoints")
+
+                        # Setze die Metadaten
+                        dataset.metadata.file_format = "VTK Unstructured Grid"
+                        dataset.metadata.num_cells = int(num_cells) if num_cells else None
+                        dataset.metadata.num_points = int(num_points) if num_points else None
+        except Exception as e:
+            # Logge das Problem und setze Metadaten auf None, falls Parsing fehlschlägt
+            dataset.metadata.file_format = None
+            dataset.metadata.num_cells = None
+            dataset.metadata.num_points = None
+            log.warning(f"Fehler beim Setzen der Metadaten für VTU-Datei: {e}")
+
+    def set_peek(self, dataset: DatasetProtocol, **kwd) -> None:
+        """Set the peek and blurb text"""
+        if not dataset.dataset.purged:
+            dataset.peek = data.get_file_peek(dataset.get_file_name())
+            "VTK Unstructured Grid file (VTU)"
+        else:
+            dataset.peek = "file does not exist"
+            dataset.blurb = "file purged from disk"
+
+    def sniff_prefix(self, file_prefix: FilePrefix) -> bool:
+        """ "Checking for keystring - 'VTKFile type="UnstructuredGrid"'.
+
+        >>> from galaxy.datatypes.sniff import get_test_fname
+        >>> fname = get_test_fname( 'data.vtu' )
+        >>> Vtu().sniff( fname )
+        True
+        >>> fname = get_test_fname( 'solid.xml' )
+        >>> Vtu().sniff( fname )
+        False
+        """
+        return self._has_root_element_in_prefix(file_prefix, 'VTKFile type="UnstructuredGrid"')

lib/galaxy/datatypes/test/data.vtu

@@ -0,0 +1,41 @@
+<?xml version="1.0"?>
+<VTKFile type="UnstructuredGrid" version="0.1" byte_order="LittleEndian" header_type="UInt32" compressor="vtkZLibDataCompressor">
+  <UnstructuredGrid>
+    <Piece NumberOfPoints="21" NumberOfCells="21">
+      <PointData>
+        <DataArray type="Float64" Name="P_BC_FLOW" format="binary" RangeMin="-2" RangeMax="0">
+          AQAAAACAAACoAAAAFgAAAA==eJxjYCAKHCBAE1JPqvno4gwABYYGwQ==
+        </DataArray>
+        <DataArray type="UInt64" Name="bulk_node_ids" format="binary" RangeMin="0" RangeMax="70">
+          AQAAAACAAACoAAAAOwAAAA==eJwtxRECgDAAAMBJkDQomQySBk2SYNKg/v+kQXdyIfySJ2fP3r34cHTx6tObqy/fbn7c/frzAL7YAwQ=
+        </DataArray>
+      </PointData>
+      <CellData>
+      </CellData>
+      <Points>
+        <DataArray type="Float64" Name="Points" NumberOfComponents="3" format="binary" RangeMin="42.42640687119285" RangeMax="94.33981132056604">
+          AQAAAACAAAD4AQAAYwAAAA==eJxjYAABOwcw1RAIoeEAXVzlAIo4gx929XBxqHq4fjT1GOLo6j1xqPdEVQ+31wWHe2Di6OajqccQRzffDlU9hjiaerh5Jmjm+KGKY4QDurgnqjiGv9DVu+BQbwcXBwAoLCke
+          <InformationKey name="L2_NORM_RANGE" location="vtkDataArray" length="2">
+            <Value index="0">
+              42.426406871
+            </Value>
+            <Value index="1">
+              94.339811321
+            </Value>
+          </InformationKey>
+        </DataArray>
+      </Points>
+      <Cells>
+        <DataArray type="Int64" Name="connectivity" format="binary" RangeMin="0" RangeMax="20">
+          AQAAAACAAACoAAAAMQAAAA==eJwtxUECQCAAALAQRSLx/7d2sF0Wwm/y7MXRqzcnZ+8+XHy6+vLt5sfdrz8PMMgA0w==
+        </DataArray>
+        <DataArray type="Int64" Name="offsets" format="binary" RangeMin="1" RangeMax="21">
+          AQAAAACAAACoAAAAMgAAAA==eJwtxbcBgCAAADArFsAC8v+rDiRLxqGbPHvx6uDNuw+fjk7Ovnz78evi6s/NPzgAAOg=
+        </DataArray>
+        <DataArray type="UInt8" Name="types" format="binary" RangeMin="1" RangeMax="1">
+          AQAAAACAAAAVAAAACwAAAA==eJxjZMQCAAD8ABY=
+        </DataArray>
+      </Cells>
+    </Piece>
+  </UnstructuredGrid>
+</VTKFile>

lib/galaxy/datatypes/xml.py

@@ -270,37 +270,4 @@ def sniff_prefix(self, file_prefix: FilePrefix) -> bool:
         """
         Checking for keyword - '<sbml' in the first 200 lines.
         """
-        return file_prefix.search(SBML_MARKER)
-
-class Vtu(GenericXml):
-    """Format for defining VTU (VTK Unstructured Grid) data https://docs.vtk.org/"""
-
-    edam_data = "edam:data_3671"
-    edam_format = "edam:format_3621"
-    file_ext = "vtu"
-
-    MetadataElement(name="file_format", default=None, desc="File format", readonly=True, optional=True, visible=True)
-    MetadataElement(name="num_cells", desc="Number of cells in the mesh", param_type="int", optional=True)
-    MetadataElement(name="num_points", desc="Number of points in the mesh", param_type="int", optional=True)
-
-    def set_peek(self, dataset: DatasetProtocol, **kwd) -> None:
-        """Set the peek and blurb text"""
-        if not dataset.dataset.purged:
-            dataset.peek = data.get_file_peek(dataset.get_file_name())
-            "VTK Unstructured Grid file (VTU)"
-        else:
-            dataset.peek = "file does not exist"
-            dataset.blurb = "file purged from disk"
-
-    def sniff_prefix(self, file_prefix: FilePrefix) -> bool:
-        """ "Checking for keystring - 'VTKFile type="UnstructuredGrid"'.
-
-        >>> from galaxy.datatypes.sniff import get_test_fname
-        >>> fname = get_test_fname( 'output.vtu' )
-        >>> Vtu().sniff( fname )
-        True
-        >>> fname = get_test_fname( 'solid.xml' )
-        >>> Vtu().sniff( fname )
-        False
-        """
-        return self._has_root_element_in_prefix(file_prefix, 'VTKFile type="UnstructuredGrid"')
+        return file_prefix.search(SBML_MARKER)