Upload test_khazna.py

tests/test_khazna.py

@@ -42,6 +42,7 @@ DB = [
     ro("a4", "Solar desalination membranes", 2023, typ="thesis", abstract="solar desalination", ku=("Ali", "Hassan", "p2")),
     ro("a5", "Graphene oxide for desalination", 2022, abstract="graphene membranes desalination", ku=("Jane", "Smith", "p1"), ext=("x1", "x2")),
     ro("a6", "Urban PM2.5 sensor readings", 2021, typ="dataset", abstract="pollution sensor readings Abu Dhabi", ext=()),
+    ro("a7", "Deep learning for seismic imaging", 2024, abstract="seismic inversion with neural networks", ku=("Kumar", "Padmanabhan", "k1"), ext=()),
 ]
 EXT_ORGS = {"x1": {"uuid": "x1", "name": {"en_GB": "Massachusetts Institute of Technology"}, "type": {"term": {"en_GB": "University"}},
                    "address": {"country": {"term": {"en_GB": "United States"}}}},
@@ -61,6 +62,12 @@ PERSONS = {
            "staffOrganizationAssociations": [{"jobTitle": {"term": {"en_GB": "Associate Professor"}}, "organization": {"uuid": "o1"}}],
            "keywordGroups": [{"keywords": [{"freeKeywords": ["membranes", "desalination"]}]}]},
     "p2": {"uuid": "p2", "name": {"firstName": "Ali", "lastName": "Hassan"}, "visibility": {"key": "FREE"}},
+    # several "Kumar"s in different name positions, plus a prefix-only match
+    "k1": {"uuid": "k1", "name": {"firstName": "Kumar", "lastName": "Padmanabhan"}, "visibility": {"key": "FREE"},
+           "staffOrganizationAssociations": [{"jobTitle": {"term": {"en_GB": "Professor"}}, "organization": {"uuid": "o1"}}]},
+    "k2": {"uuid": "k2", "name": {"firstName": "Rajesh", "lastName": "Kumar"}, "visibility": {"key": "FREE"}},
+    "k3": {"uuid": "k3", "name": {"firstName": "Anil Kumar", "lastName": "Sharma"}, "visibility": {"key": "FREE"}},
+    "k4": {"uuid": "k4", "name": {"firstName": "Kumaran", "lastName": "Iyer"}, "visibility": {"key": "FREE"}},
 }
 ORGS = {"o1": {"uuid": "o1", "name": {"en_GB": "Department of Chemical Engineering"}, "type": {"term": {"en_GB": "Department"}}, "visibility": {"key": "FREE"}},
         "o2": {"uuid": "o2", "name": {"en_GB": "Center for Membranes and Advanced Water Technology"}, "type": {"term": {"en_GB": "Research Centre"}}, "visibility": {"key": "FREE"}}}
@@ -269,6 +276,7 @@ def test_endpoint_scopes_widget_ratelimit_and_dataset_403():
         assert d["intent"] == "expert_search" and d["search_results"][0]["kind"] == "person"
         assert "email" not in json.dumps(d)
         assert "/en/persons/?search=desalination" in d["portal_search_url"]
+        assert "searchBy" not in d["portal_search_url"]
         d = tc.post("/agent", json={"question": "membranes", "scope": "organisations"}).json()
         assert d["search_results"][0]["kind"] == "organisation"
         DATASETS_ENABLED["on"] = False
@@ -429,3 +437,42 @@ def test_agent_new_scopes_end_to_end():
         assert d["search_results"][0]["kind"] == "collaborator" and d["search_results"][0]["joint_outputs"] == 2
         d = tc.post("/agent", json={"question": "hydrogen storage", "scope": "publications"}).json()
         assert {f["scope"] for f in d["follow_up_suggestions"]} == {"persons", "theses", "projects", "collaborations"}
+
+
+def test_match_people_positions_and_prefix():
+    from pure_client import normalize_person
+    people = [normalize_person(PERSONS[k], PORTAL) for k in ("k1", "k2", "k3", "k4")]
+    matches, full = pipeline.match_people("KUMAR", people)
+    assert [p["uuid"] for p in matches] == ["k2", "k3", "k1"] or {p["uuid"] for p in matches} == {"k1", "k2", "k3"}
+    assert matches[0]["uuid"] == "k2"                      # surname matches are listed first
+    assert "k4" not in {p["uuid"] for p in matches}        # 'Kumaran' only used when no exact match exists
+    assert full == []
+    m, f = pipeline.match_people("Kumar Padmanabhan", people)
+    assert [p["uuid"] for p in f] == ["k1"]
+    m, f = pipeline.match_people("Kumara", people)          # prefix fallback
+    assert [p["uuid"] for p in m] == ["k4"]
+
+
+def test_ambiguous_name_asks_to_choose_then_explores_choice():
+    os.environ["PURE_API_KEY"] = "test-key"
+    import app as appmod
+    from fastapi.testclient import TestClient
+    with TestClient(appmod.app) as tc:
+        appmod.pure = client()
+        appmod._hits.clear()
+        d = tc.post("/agent", json={"question": "Kumar", "scope": "persons"}).json()
+        assert d["intent"] == "author_clarify" and d["clarify"] is True
+        assert d["portal_search_url"] == "https://khazna.ku.ac.ae/en/persons/?search=Kumar"
+        assert {x["uuid"] for x in d["search_results"]} == {"k1", "k2", "k3"}
+        assert all(x["choose"] for x in d["search_results"]) and "3 KU researchers" in d["answer"]
+        assert d["search_results"][[x["uuid"] for x in d["search_results"]].index("k1")]["organisations"][0]["name"].startswith("Department")
+        # the user picks one → that exact person, no second clarification
+        d2 = tc.post("/agent", json={"question": d["original_question"], "scope": "persons", "person_uuid": "k1"}).json()
+        assert d2["intent"] == "author_search" and d2["person"]["uuid"] == "k1"
+        assert [x["uuid"] for x in d2["search_results"]] == ["a7"]
+        # a full unique name goes straight through
+        d3 = tc.post("/agent", json={"question": "Kumar Padmanabhan", "scope": "persons"}).json()
+        assert d3["intent"] == "author_search" and d3["person"]["uuid"] == "k1"
+        # a non-matching / single name still works as before
+        d4 = tc.post("/agent", json={"question": "Jane Smith", "scope": "persons"}).json()
+        assert d4["person"]["name"] == "Jane Smith"