✨ ALTO: Count Layout/Page/* elements

qurator/modstool/lib.py

@@ -198,7 +198,7 @@ class TagGroup:
         counts = {}
         for e in self.group:
             for x in e.iter():
-                tag = ET.QName(x).localname
+                tag = ET.QName(x.tag).localname
                 key = f"{tag}-count"
                 counts[key] = counts.get(key, 0) + 1
         return counts

qurator/modstool/tests/test_alto.py

@@ -0,0 +1,39 @@
+import xml.etree.ElementTree as ET
+
+
+from qurator.modstool.altotool import alto_to_dict
+from qurator.modstool.lib import flatten
+
+
+def dict_fromstring(x):
+   return flatten(alto_to_dict(ET.fromstring(x)))
+
+def test_Page_counts():
+    """
+    Elements below Layout/Page should be counted
+    """
+    d = dict_fromstring("""
+    <alto xmlns="http://www.loc.gov/standards/alto/ns-v2#">
+      <Layout>
+        <Page ID="Page1" PHYSICAL_IMG_NR="1">
+            <TextBlock ID="Page1_Block1">
+              <TextLine>
+                <String STYLE="bold" WC="0.8937500119" CONTENT="Staatsbibliothek" />
+              </TextLine>
+              <TextLine>
+                <String STYLE="bold" WC="0.8899999857" CONTENT="zu" />
+                <String STYLE="bold" WC="0.9866666794" CONTENT="Berlin" />
+              </TextLine>
+              <TextLine>
+                <String STYLE="bold" WC="1." CONTENT="WM" />
+                <String STYLE="bold" WC="0.8927272558" CONTENT="Preußischer" />
+                <String STYLE="bold" WC="0.9058333039" CONTENT="Kulturbesitz" />
+              </TextLine>
+            </TextBlock>
+        </Page>
+      </Layout>
+    </alto>
+    """)
+    assert d['Layout_Page_TextBlock-count'] == 1
+    assert d['Layout_Page_TextLine-count'] == 3
+    assert d['Layout_Page_String-count'] == 6