Elsevier Parser: remove date_published simulation

* ref: ref: cern-sis/issues-scoap3#338
cern-sis · Jul 22, 2024 · cd97ba1 · cd97ba1
1 parent d11f31b
commit cd97ba1
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 11 deletions.
diff --git a/dags/elsevier/metadata_parser.py b/dags/elsevier/metadata_parser.py
@@ -31,12 +31,10 @@ def __init__(self, file_path):
             CustomExtractor(
                 destination="date_published",
                 extraction_function=self._get_published_date,
-                required=True,
             ),
             CustomExtractor(
                 destination="journal_year",
                 extraction_function=self._get_journal_year,
-                required=True,
             ),
             CustomExtractor(
                 destination="collections",
@@ -88,14 +86,11 @@ def _get_published_date(self, article):
             field_name="published_date",
             dois=self.dois,
         )
-        if not date:
-            self.published_date = datetime.now().strftime("%Y-%m-%d")
-            self.year = datetime.now().strftime("%Y")
+        if date:
+            date = datetime.fromisoformat(date[:-1])
+            self.published_date = date.strftime("%Y-%m-%d")
+            self.year = date.strftime("%Y")
             return self.published_date
-        date = datetime.fromisoformat(date[:-1])
-        self.published_date = date.strftime("%Y-%m-%d")
-        self.year = date.strftime("%Y")
-        return self.published_date
 
     def _get_journal_year(self, article):
         return self.year

diff --git a/tests/units/elsevier/test_metadata_parser.py b/tests/units/elsevier/test_metadata_parser.py
@@ -75,7 +75,7 @@ def parsed_articles(parser, article):
             id="test_publication_info",
         ),
         param(
-            ["2023-11-02", "2023-11-02", "2023-02-04", "2023-11-02"],
+            ["", "", "2023-02-04", ""],
             "date_published",
             id="test_published_date",
         ),
@@ -150,7 +150,7 @@ def parsed_articles(parser, article):
 @freeze_time("2023-11-02")
 def test_elsevier_dataset_parsing(parsed_articles, expected, key):
     for (parsed_article, expected_article) in zip(parsed_articles, expected):
-        assert expected_article == parsed_article[key]
+        assert expected_article == parsed_article.get(key, "")
 
 
 @fixture