From 2f7a924622ded9bdf8cbb98a62ae248849aad86d Mon Sep 17 00:00:00 2001
From: anon <tim.treis@helmholtz-munich.de>
Date: Mon, 9 Nov 2026 19:11:11 +0100
Subject: [PATCH] feat(query): filter on the tissue and disease roll-up columns

query() rejected tissue_system, tissue_organ, disease_class and disease_group, although the hub
cut publishes them and the MCP connector's snippets filter on them, e.g.
Catalog(version="2026-11-08").query(tissue_organ="brain", validation=None). They are now facets.
Each row carries its anchor, so equality on an anchor includes every tissue or disease under it.
---
 src/sddb/_catalog_schema.py    |  5 +++++
 tests/test_facets_study_doi.py | 29 +++++++++++++++++++++++++++++
 2 files changed, 34 insertions(+)

diff --git a/src/sddb/_catalog_schema.py b/src/sddb/_catalog_schema.py
index c476bad..a0d025a 100644
--- a/src/sddb/_catalog_schema.py
+++ b/src/sddb/_catalog_schema.py
@@ -193,7 +193,12 @@
     "doi",
     "organism",
     "tissue",
+    # Ontology roll-ups the hub cut writes: each row has its anchor, so equality includes everything under it.
+    "tissue_system",
+    "tissue_organ",
     "disease",
+    "disease_class",
+    "disease_group",
     "development_stage",
     "technology",
     "vendor",
diff --git a/tests/test_facets_study_doi.py b/tests/test_facets_study_doi.py
index d125551..fa79a4a 100644
--- a/tests/test_facets_study_doi.py
+++ b/tests/test_facets_study_doi.py
@@ -1,4 +1,5 @@
 import pandas as pd
+import pytest
 
 from sddb import _catalog_schema as schema
 from sddb.catalog import Catalog
@@ -27,3 +28,31 @@ def test_query_by_study_id(tmp_path):
     cat = Catalog.from_file(p)
     assert list(cat.query(study_id="oliveira_2025").to_df()["uid"]) == ["u1"]
     assert list(cat.query(doi="10.2/y").to_df()["uid"]) == ["u2"]
+
+
+def test_query_by_tissue_and_disease_rollup(tmp_path):
+    df = pd.DataFrame(
+        {
+            "uid": ["u1", "u2", "u3"],
+            "technology": ["Xenium"] * 3,
+            "assay": ["spatial"] * 3,
+            "organism": ["Mus musculus"] * 3,
+            "tissue": ["cerebellum", "hippocampal formation", "liver"],
+            "tissue_system": ["nervous system", "nervous system", "digestive system"],
+            "tissue_organ": ["brain", "brain", "liver"],
+            "disease_class": ["neurodegenerative disease", None, None],
+            "disease_group": ["Alzheimer disease", None, None],
+            "validation_status": ["pass"] * 3,
+            "zarr_url": ["s3://b/1.zarr", "s3://b/2.zarr", "s3://b/3.zarr"],
+        }
+    )
+    p = tmp_path / "catalog.parquet"
+    df.to_parquet(p)
+    cat = Catalog.from_file(p)
+    # the form the MCP connector emits
+    assert list(cat.query(tissue_organ="brain", validation=None).to_df()["uid"]) == ["u1", "u2"]
+    assert list(cat.query(tissue_system="Nervous System").to_df()["uid"]) == ["u1", "u2"]
+    assert list(cat.query(disease_group="Alzheimer disease", tissue_organ="brain").to_df()["uid"]) == ["u1"]
+    assert list(cat.query(disease_class="neurodegenerative disease").to_df()["uid"]) == ["u1"]
+    with pytest.raises(ValueError, match="did you mean"):
+        cat.query(tissue_organ="brian")