Validate NX_class against known NeXus class names

src/chexus/nexus_base_classes.py

@@ -0,0 +1,157 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2026 Scipp contributors (https://github.com/scipp)
+"""NeXus base class names.
+
+Source: https://github.com/nexusformat/definitions/tree/main/base_classes
+Regenerated by ``tools/refresh_nexus_base_classes.py``; do not edit by hand.
+
+Includes deprecated classes (NXgeometry, NXorientation, NXshape, NXtranslation),
+which are flagged separately by NX_class_is_legacy.
+"""
+
+nexus_base_classes = frozenset(
+    [
+        "NXaberration",
+        "NXactivity",
+        "NXactuator",
+        "NXaperture",
+        "NXapm_charge_state_analysis",
+        "NXapm_event_data",
+        "NXapm_instrument",
+        "NXapm_measurement",
+        "NXapm_ranging",
+        "NXapm_reconstruction",
+        "NXapm_simulation",
+        "NXatom",
+        "NXattenuator",
+        "NXbeam",
+        "NXbeam_stop",
+        "NXbeam_transfer_matrix_table",
+        "NXbending_magnet",
+        "NXcalibration",
+        "NXcapillary",
+        "NXcg_alpha_complex",
+        "NXcg_cylinder",
+        "NXcg_ellipsoid",
+        "NXcg_face_list_data_structure",
+        "NXcg_grid",
+        "NXcg_half_edge_data_structure",
+        "NXcg_hexahedron",
+        "NXcg_parallelogram",
+        "NXcg_point",
+        "NXcg_polygon",
+        "NXcg_polyhedron",
+        "NXcg_polyline",
+        "NXcg_primitive",
+        "NXcg_roi",
+        "NXcg_tetrahedron",
+        "NXcg_triangle",
+        "NXcg_unit_normal",
+        "NXchemical_composition",
+        "NXcircuit",
+        "NXcite",
+        "NXcollection",
+        "NXcollectioncolumn",
+        "NXcollimator",
+        "NXcomponent",
+        "NXcoordinate_system",
+        "NXcorrector_cs",
+        "NXcrystal",
+        "NXcs_computer",
+        "NXcs_filter_boolean_mask",
+        "NXcs_memory",
+        "NXcs_prng",
+        "NXcs_processor",
+        "NXcs_profiling",
+        "NXcs_profiling_event",
+        "NXcs_storage",
+        "NXcylindrical_geometry",
+        "NXdata",
+        "NXdeflector",
+        "NXdetector",
+        "NXdetector_channel",
+        "NXdetector_group",
+        "NXdetector_module",
+        "NXdisk_chopper",
+        "NXdistortion",
+        "NXebeam_column",
+        "NXelectromagnetic_lens",
+        "NXelectron_detector",
+        "NXelectronanalyzer",
+        "NXem_ebsd",
+        "NXem_eds",
+        "NXem_eels",
+        "NXem_event_data",
+        "NXem_img",
+        "NXem_instrument",
+        "NXem_interaction_volume",
+        "NXem_measurement",
+        "NXem_optical_system",
+        "NXem_simulation",
+        "NXenergydispersion",
+        "NXentry",
+        "NXenvironment",
+        "NXevent_data",
+        "NXfabrication",
+        "NXfermi_chopper",
+        "NXfilter",
+        "NXfit",
+        "NXfit_function",
+        "NXflipper",
+        "NXfresnel_zone_plate",
+        "NXgeometry",
+        "NXgrating",
+        "NXguide",
+        "NXhistory",
+        "NXibeam_column",
+        "NXimage",
+        "NXinsertion_device",
+        "NXinstrument",
+        "NXlog",
+        "NXmanipulator",
+        "NXmirror",
+        "NXmoderator",
+        "NXmonitor",
+        "NXmonochromator",
+        "NXnote",
+        "NXobject",
+        "NXoff_geometry",
+        "NXoptical_lens",
+        "NXoptical_window",
+        "NXorientation",
+        "NXparameters",
+        "NXpdb",
+        "NXpeak",
+        "NXphase",
+        "NXpid_controller",
+        "NXpinhole",
+        "NXpolarizer",
+        "NXpositioner",
+        "NXprocess",
+        "NXprogram",
+        "NXpump",
+        "NXreflections",
+        "NXregistration",
+        "NXresolution",
+        "NXroi_process",
+        "NXroot",
+        "NXrotations",
+        "NXsample",
+        "NXsample_component",
+        "NXscan_controller",
+        "NXsensor",
+        "NXshape",
+        "NXslit",
+        "NXsource",
+        "NXspectrum",
+        "NXspindispersion",
+        "NXsubentry",
+        "NXtransformations",
+        "NXtranslation",
+        "NXunit_cell",
+        "NXuser",
+        "NXvelocity_selector",
+        "NXwaveplate",
+        "NXxraylens",
+    ]
+)

src/chexus/validators.py

@@ -1,7 +1,8 @@
 # SPDX-License-Identifier: BSD-3-Clause
-# Copyright (c) 2023 Scipp contributors (https://github.com/scipp)
+# Copyright (c) 2026 Scipp contributors (https://github.com/scipp)
 import numpy as np
 
+from .nexus_base_classes import nexus_base_classes
 from .tree import Dataset, Group
 from .validate import Validator, Violation
 
@@ -73,6 +74,30 @@ def validate(self, node: Dataset | Group) -> Violation | None:
             return Violation(node.name, f"NX_class {nx_class} is deprecated")
 
 
+# Application/contributed classes expected in raw ESS files.
+# Add to this set if a new application class is encountered.
+nexus_application_classes = frozenset(["NXcanSAS"])
+
+valid_nx_classes = nexus_base_classes | nexus_application_classes
+
+
+class NX_class_invalid(Validator):
+    def __init__(self) -> None:
+        super().__init__(
+            "NX_class_invalid",
+            "NX_class is not a known NeXus class name "
+            "(typos like NXPositioner instead of NXpositioner)",
+        )
+
+    def applies_to(self, node: Dataset | Group) -> bool:
+        return isinstance(node, Group) and "NX_class" in node.attrs
+
+    def validate(self, node: Dataset | Group) -> Violation | None:
+        nx_class = node.attrs["NX_class"]
+        if nx_class not in valid_nx_classes:
+            return Violation(node.name, f"Unknown NX_class {nx_class!r}")
+
+
 class group_has_units(Validator):
     def __init__(self) -> None:
         super().__init__("group_has_units", "Group should not have units attribute")
@@ -497,6 +522,7 @@ def base_validators(*, has_scipp=True):
         mask_has_units(),
         non_numeric_dataset_has_units(),
         NX_class_attr_missing(),
+        NX_class_invalid(),
         NX_class_is_legacy(),
         transformation_depends_on_missing(),
         transformation_offset_units_missing(),

tests/validators_test.py

@@ -178,6 +178,33 @@ def test_NX_class_attr_missing():
     assert result.name == "x"
 
 
+@pytest.mark.parametrize(
+    "nx_class",
+    ["NXpositioner", "NXdisk_chopper", "NXgeometry", "NXcanSAS"],
+)
+def test_NX_class_invalid_accepts_known(nx_class: str):
+    group = chexus.Group(name="x", attrs={"NX_class": nx_class})
+    assert chexus.validators.NX_class_invalid().applies_to(group)
+    assert chexus.validators.NX_class_invalid().validate(group) is None
+
+
+@pytest.mark.parametrize(
+    "nx_class",
+    ["NXPositioner", "NXDetector", "NXfoo", "positioner"],
+)
+def test_NX_class_invalid_rejects_unknown(nx_class: str):
+    group = chexus.Group(name="x", attrs={"NX_class": nx_class})
+    assert chexus.validators.NX_class_invalid().applies_to(group)
+    result = chexus.validators.NX_class_invalid().validate(group)
+    assert isinstance(result, chexus.Violation)
+    assert result.name == "x"
+
+
+def test_NX_class_invalid_does_not_apply_without_attr():
+    group = chexus.Group(name="x", attrs={})
+    assert not chexus.validators.NX_class_invalid().applies_to(group)
+
+
 def test_NX_class_is_legacy():
     good = chexus.Group(name="x", attrs={"NX_class": "NXtransformations"})
     assert chexus.validators.NX_class_is_legacy().validate(good) is None

tools/refresh_nexus_base_classes.py

@@ -0,0 +1,77 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2026 Scipp contributors (https://github.com/scipp)
+"""Regenerate ``src/chexus/nexus_base_classes.py`` from the upstream NeXus
+definitions repository.
+
+Requires the GitHub CLI (``gh``) to be installed and authenticated.
+
+Usage:
+    python tools/refresh_nexus_base_classes.py
+"""
+
+from __future__ import annotations
+
+import subprocess
+import sys
+from pathlib import Path
+
+REPO = "nexusformat/definitions"
+DIRECTORY = "base_classes"
+SUFFIX = ".nxdl.xml"
+TARGET = (
+    Path(__file__).resolve().parent.parent / "src" / "chexus" / "nexus_base_classes.py"
+)
+
+HEADER = '''# SPDX-License-Identifier: BSD-3-Clause
+# Copyright (c) 2026 Scipp contributors (https://github.com/scipp)
+"""NeXus base class names.
+
+Source: https://github.com/nexusformat/definitions/tree/main/base_classes
+Regenerated by ``tools/refresh_nexus_base_classes.py``; do not edit by hand.
+
+Includes deprecated classes (NXgeometry, NXorientation, NXshape, NXtranslation),
+which are flagged separately by NX_class_is_legacy.
+"""
+
+nexus_base_classes = frozenset(
+    [
+'''
+
+FOOTER = """    ]
+)
+"""
+
+
+def fetch_names() -> list[str]:
+    result = subprocess.run(
+        ["gh", "api", f"repos/{REPO}/contents/{DIRECTORY}", "--jq", ".[].name"],
+        capture_output=True,
+        text=True,
+        check=True,
+    )
+    names = [
+        line.removesuffix(SUFFIX)
+        for line in result.stdout.splitlines()
+        if line.endswith(SUFFIX)
+    ]
+    return sorted(names)
+
+
+def render(names: list[str]) -> str:
+    body = "".join(f'        "{name}",\n' for name in names)
+    return HEADER + body + FOOTER
+
+
+def main() -> int:
+    names = fetch_names()
+    if not names:
+        print("No base classes returned from gh api", file=sys.stderr)
+        return 1
+    TARGET.write_text(render(names))
+    print(f"Wrote {len(names)} base classes to {TARGET}")
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())