From 22f6da540f017df08b96490b68a23f19f3a9c57c Mon Sep 17 00:00:00 2001
From: flange <38500-flange@users.noreply.git.rwth-aachen.de>
Date: Tue, 12 Nov 2024 16:51:28 +0100
Subject: [PATCH] exclude deprecated disciplines
---
project/dalia/curation/suggest/disciplines.py | 17 ++++++++--
.../curation/suggest/test_disciplines.py | 31 +++++++++++++++++--
2 files changed, 44 insertions(+), 4 deletions(-)
diff --git a/project/dalia/curation/suggest/disciplines.py b/project/dalia/curation/suggest/disciplines.py
index efad3c0..172aba7 100644
--- a/project/dalia/curation/suggest/disciplines.py
+++ b/project/dalia/curation/suggest/disciplines.py
@@ -1,10 +1,16 @@
from typing import List, Tuple
-from rdflib import Literal, RDF, SKOS, URIRef, Variable
+from rdflib import Literal, OWL, RDF, SKOS, URIRef, Variable
from project.dalia.api_models.api_models import CurationSuggestDisciplinesResultItem
from project.dalia.query.utils import query_ontologies_dataset
-from project.dalia.query_builder.query_builder import FILTER, FunctionExpressions, Operators, QueryBuilder
+from project.dalia.query_builder.query_builder import (
+ FILTER,
+ FILTER_EXISTS,
+ FunctionExpressions,
+ Operators,
+ QueryBuilder,
+)
# Note: This algorithm is quite inefficient and results in a lot of SPARQL queries.
# Another idea would be to pull all disciplines and their narrower/broader nodes at once, e.g. via
@@ -19,6 +25,9 @@ from project.dalia.query_builder.query_builder import FILTER, FunctionExpression
# ?broaderDiscipline skos:narrower ?discipline .
# }
# ?discipline a skos:Concept .
+# FILTER NOT EXISTS {
+# ?discipline owl:deprecated true .
+# }
# ?discipline skos:prefLabel ?label .
# FILTER(LANG(?label) = "en") .
# ?discipline skos:notation ?notation .
@@ -57,6 +66,10 @@ def _prepare_query_to_get_disciplines_metadata(discipline_selection_bgp: Tuple)
).WHERE(
discipline_selection_bgp,
(var_discipline, RDF.type, SKOS.Concept),
+ FILTER_EXISTS(
+ (var_discipline, OWL.deprecated, Literal(True)),
+ state=False
+ ),
(var_discipline, SKOS.prefLabel, var_label),
FILTER(
Operators.EQ(
diff --git a/tests/project/dalia/curation/suggest/test_disciplines.py b/tests/project/dalia/curation/suggest/test_disciplines.py
index 8dd4c38..8625a45 100644
--- a/tests/project/dalia/curation/suggest/test_disciplines.py
+++ b/tests/project/dalia/curation/suggest/test_disciplines.py
@@ -1,10 +1,16 @@
+from typing import List
+
from django.urls import reverse
from rdflib import URIRef
from rest_framework import status
from project.dalia.api_models.api_models import CurationSuggestDisciplinesResultItem
-from project.dalia.curation.suggest.disciplines import get_child_disciplines_of_discipline, get_top_disciplines, \
- prepare_query_to_get_all_narrower_disciplines_of_discipline, prepare_query_to_get_all_top_disciplines
+from project.dalia.curation.suggest.disciplines import (
+ get_child_disciplines_of_discipline,
+ get_top_disciplines,
+ prepare_query_to_get_all_narrower_disciplines_of_discipline,
+ prepare_query_to_get_all_top_disciplines,
+)
from project.dalia.serializers import CurationSuggestDisciplinesResultItemSerializer
from tests.project.dalia.utils import dedent_and_normalize, normalize
@@ -17,6 +23,9 @@ def test_prepare_query_to_get_all_top_disciplines():
WHERE {
?discipline <http://www.w3.org/2004/02/skos/core#topConceptOf> <https://w3id.org/kim/hochschulfaechersystematik/scheme> .
?discipline <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2004/02/skos/core#Concept> .
+ FILTER NOT EXISTS {
+ ?discipline <http://www.w3.org/2002/07/owl#deprecated> "true"^^<http://www.w3.org/2001/XMLSchema#boolean> .
+ } .
?discipline <http://www.w3.org/2004/02/skos/core#prefLabel> ?label .
FILTER ( LANG ( ?label ) = "en" ) .
?discipline <http://www.w3.org/2004/02/skos/core#notation> ?notation .
@@ -33,6 +42,9 @@ def test_prepare_query_to_get_all_narrower_disciplines_of_discipline():
WHERE {
<abc> <http://www.w3.org/2004/02/skos/core#narrower> ?discipline .
?discipline <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2004/02/skos/core#Concept> .
+ FILTER NOT EXISTS {
+ ?discipline <http://www.w3.org/2002/07/owl#deprecated> "true"^^<http://www.w3.org/2001/XMLSchema#boolean> .
+ } .
?discipline <http://www.w3.org/2004/02/skos/core#prefLabel> ?label .
FILTER ( LANG ( ?label ) = "en" ) .
?discipline <http://www.w3.org/2004/02/skos/core#notation> ?notation .
@@ -142,3 +154,18 @@ def test_get_on_CurationSuggestDisciplinesView_returns_200_and_the_disciplines_t
assert n242.label == "Interior Architecture"
assert n242.value == "https://w3id.org/kim/hochschulfaechersystematik/n242"
assert len(n242.children) == 0
+
+ assert _has_item_with_uri(data, "https://w3id.org/kim/hochschulfaechersystematik/n077")
+ assert _has_item_with_uri(data, "https://w3id.org/kim/hochschulfaechersystematik/n194")
+
+ # deprecated items are not in the returned data
+ assert not _has_item_with_uri(data, "https://w3id.org/kim/hochschulfaechersystematik/n241")
+ assert not _has_item_with_uri(data, "https://w3id.org/kim/hochschulfaechersystematik/n237")
+
+
+# DFS in nested data structure
+def _has_item_with_uri(data: List[CurationSuggestDisciplinesResultItem], uri: str) -> bool:
+ for item in data:
+ if item.value == uri or _has_item_with_uri(item.children, uri):
+ return True
+ return False
--
GitLab