From 74282d6b76e8efbbceee7fc8823c3d7ae15c916c Mon Sep 17 00:00:00 2001
From: Johannes Kasimir <johanneskasimir@gmail.com>
Date: Thu, 23 Apr 2026 10:05:41 +0200
Subject: [PATCH] feat: verify type of event_index has 8 bytes

---
 src/chexus/validators.py | 18 ++++++++++++++++++
 tests/validators_test.py | 24 ++++++++++++++++++++++++
 2 files changed, 42 insertions(+)

diff --git a/src/chexus/validators.py b/src/chexus/validators.py
index 77ee418..dd7dc2c 100644
--- a/src/chexus/validators.py
+++ b/src/chexus/validators.py
@@ -469,6 +469,24 @@ def validate(self, node: Dataset | Group) -> Violation | None:
                 return Violation(node.name, "NXlog must have a value")
 
 
+class event_index_is_eight_bytes(Validator):
+    def __init__(self) -> None:
+        super().__init__(
+            "event_index_too_small", "event_index needs to be an 8 byte type"
+        )
+
+    def applies_to(self, node: Dataset | Group) -> bool:
+        return (
+            isinstance(node, Dataset)
+            and node.parent.attrs.get("NX_class") == "NXevent_data"
+            and node.name == 'event_index'
+        )
+
+    def validate(self, node: Dataset | Group) -> Violation | None:
+        if np.dtype(node.dtype).itemsize < 8:
+            return Violation(node.name, "event_index type too small")
+
+
 def base_validators(*, has_scipp=True):
     validators = [
         depends_on_missing(),
diff --git a/tests/validators_test.py b/tests/validators_test.py
index deb3104..8cfe612 100644
--- a/tests/validators_test.py
+++ b/tests/validators_test.py
@@ -798,3 +798,27 @@ def test_NXdetector_pixel_offsets_are_unambiguous_2d_ids_allow_axis_attr() -> No
         chexus.validators.NXdetector_pixel_offsets_are_unambiguous().validate(det)
         is None
     )
+
+
+def test_event_index_is_eight_bytes() -> None:
+    parent = chexus.Group(name="event_data", attrs={"NX_class": "NXevent_data"})
+    good = chexus.Dataset(
+        name='event_index',
+        value=[1, 2, 3],
+        shape=(3,),
+        dtype='int64',
+        parent=parent,
+    )
+    bad = chexus.Dataset(
+        name='event_index',
+        value=[1, 2, 3],
+        shape=(3,),
+        dtype='int32',
+        parent=parent,
+    )
+    assert chexus.validators.event_index_is_eight_bytes().applies_to(good)
+    assert chexus.validators.event_index_is_eight_bytes().applies_to(bad)
+    assert chexus.validators.event_index_is_eight_bytes().validate(good) is None
+    assert isinstance(
+        chexus.validators.event_index_is_eight_bytes().validate(bad), chexus.Violation
+    )