use imprints date from tag rather than publication_date

ref: cern-sis/issues-inspire#206

hepcrawl/parsers/elsevier.py

@@ -127,9 +127,12 @@ def parse(self):
             self.builder.add_doi(**doi)
         for keyword in self.keywords:
             self.builder.add_keyword(keyword)
-        self.builder.add_imprint_date(
-            self.publication_date.dumps() if self.publication_date else None
-        )
+        if self.imprints_date:
+            self.builder.add_imprint_date(self.imprints_date)
+        elif self.publication_date:
+            self.builder.add_imprint_date(
+                self.publication_date.dumps()
+            )
         for reference in self.references:
             self.builder.add_reference(reference)
 
@@ -372,6 +375,14 @@ def page_end(self):
         ).extract_first()
         return page_end
 
+    @property
+    def imprints_date(self):
+        imprints_date = self.root.xpath(
+            "string(./RDF/Description/availableOnlineInformation/availableOnline)"
+        ).extract_first()
+        if imprints_date:
+            return PartialDate.parse(imprints_date).dumps()
+
     @property
     def publication_date(self):
         publication_date = None

setup.py

@@ -76,7 +76,7 @@
 }
 
 setup_requires = [
-    'autosemver~=0.2',
+    "autosemver~=0.0,>=0.5.2,<1.0.0",
 ]
 
 extras_require['all'] = []

tests/unit/test_parsers_elsevier.py

@@ -157,3 +157,9 @@ def test_record_should_be_harvested(records):
 def test_record_shouldnt_be_harvested():
     parser = get_parser_by_file("record-that-shouldnt-be-harvested.xml")
     assert not parser.should_record_be_harvested()
+
+
+def test_imprints_date_should_be_taken_from_avaliable_online():
+    parser = get_parser_by_file("j.nima.2023.168018.xml")
+    result = parser.parse()
+    assert result['imprints'] == [{'date': '2023-01-02'}]