deposits: fix PDF date extraction

* Fixes an error with the date property when the PDF contains an host document. * Closes #603. Co-Authored-by: Sébastien Délèze <sebastien.deleze@rero.ch>
rero · Jul 21, 2021 · 2f96691 · 2f96691
1 parent dc8eb1a
commit 2f96691
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 1 deletion.
diff --git a/sonar/modules/pdf_extractor/utils.py b/sonar/modules/pdf_extractor/utils.py
@@ -157,7 +157,10 @@ def format_extracted_data(data):
                     '@from'] + '-' + item['@to']
 
         if monogr['imprint'].get('date').get('@when'):
-            publication['year'] = monogr['imprint']['date']['@when']
+            match = re.search(r'^([0-9]{4}).*$',
+                              monogr['imprint']['date']['@when'])
+            if match:
+                formatted_data['documentDate'] = match.group(1)
 
     if publication:
         formatted_data['publication'] = publication

diff --git a/tests/ui/pdf_extractor/test_pdf_extractor_utils.py b/tests/ui/pdf_extractor/test_pdf_extractor_utils.py
@@ -74,3 +74,4 @@ def test_format_extracted_data(app):
         assert formatted_data['publication'][
             'publishedIn'] == 'Frontiers in Earth Science'
         assert formatted_data['publication']['volume'] == '7'
+        assert formatted_data['documentDate'] == '2019'