related_items

MITLibraries · Jul 29, 2024 · ac00eaa · ac00eaa
1 parent ad1d046
commit ac00eaa
Show file tree

Hide file tree

Showing 2 changed files with 166 additions and 59 deletions.
diff --git a/tests/sources/xml/test_marc.py b/tests/sources/xml/test_marc.py
@@ -1731,6 +1731,103 @@ def test_get_publishers_transforms_correctly_if_fields_missing():
     assert Marc.get_publishers(source_record) is None
 
 
+def test_get_related_items_success():
+    source_record = create_marc_source_record_stub(
+        datafield_insert=(
+            """
+            <datafield tag="765" ind1="0" ind2=" ">
+                <subfield code="t">Java 2 in plain English.</subfield>
+            </datafield>
+            <datafield tag="770" ind1="1" ind2=" ">
+                <subfield code="t">Geological Society of America data repository</subfield>
+                <subfield code="w">(DLC)sn 86025915</subfield>
+                <subfield code="w">(OCoLC)13535209</subfield>
+            </datafield>
+            <datafield tag="772" ind1="0" ind2="0">
+                <subfield code="a">Earthquake engineering and structural dynamics</subfield>
+                <subfield code="v">v. 14, no. 5</subfield>
+            </datafield>
+            <datafield tag="780" ind1="1" ind2="1">
+                <subfield code="t">Entertainment design</subfield>
+                <subfield code="x">1520-5150</subfield>
+            </datafield>
+            <datafield tag="785" ind1="0" ind2="0">
+                <subfield code="t">Protist</subfield>
+                <subfield code="w">(DLC)sn 98050216</subfield>
+                <subfield code="w">(OCoLC)39018023</subfield>
+                <subfield code="x">1434-4610</subfield>
+            </datafield>
+            <datafield tag="787" ind1="0" ind2=" ">
+                <subfield code="i">Part of:</subfield>
+                <subfield code="t">De historien des Ouden en Nieuwen Testaments</subfield>
+            </datafield>
+            <datafield tag="830" ind1=" " ind2="0">
+                <subfield code="a">Map and chart series (New York State Geological Survey) ;</subfield>
+                <subfield code="0">(DLC)n  84704569</subfield>
+                <subfield code="0">(DLC)n  84704570</subfield>
+                <subfield code="v">no. 53.</subfield>
+                <subfield code="x">0097-3793</subfield>
+            </datafield>
+            <datafield tag="510" ind1="2" ind2=" ">
+                <subfield code="a">Predicasts</subfield>
+            </datafield>
+            """
+        )
+    )
+    assert Marc.get_related_items(source_record) == [
+        timdex.RelatedItem(
+            description="Java 2 in plain English",
+            relationship="Original Language Version",
+        ),
+        timdex.RelatedItem(
+            description=(
+                "Geological Society of America data repository "
+                "(DLC)sn 86025915 "
+                "(OCoLC)13535209"
+            ),
+            relationship="Has Supplement",
+        ),
+        timdex.RelatedItem(
+            description="Earthquake engineering and structural dynamics",
+            relationship="Supplement To",
+        ),
+        timdex.RelatedItem(
+            description=("Entertainment design 1520-5150"), relationship="Previous Title"
+        ),
+        timdex.RelatedItem(
+            description="Protist (DLC)sn 98050216 (OCoLC)39018023 1434-4610",
+            relationship="Subsequent Title",
+        ),
+        timdex.RelatedItem(
+            description="Part of: De historien des Ouden en Nieuwen Testaments",
+            relationship="Not Specified",
+        ),
+        timdex.RelatedItem(
+            description="Map and chart series (New York State Geological Survey) ; no. 53. 0097-3793",
+            relationship="In Series",
+        ),
+        timdex.RelatedItem(description="Predicasts", relationship="In Bibliography"),
+    ]
+
+
+def test_get_related_items_transforms_correctly_if_fields_blank():
+    source_record = create_marc_source_record_stub(
+        datafield_insert=(
+            """
+            <datafield tag="765" ind1="0" ind2=" ">
+                <subfield code="t"></subfield>
+            </datafield>
+            """
+        )
+    )
+    assert Marc.get_related_items(source_record) is None
+
+
+def test_get_related_items_transforms_correctly_if_fields_missing():
+    source_record = create_marc_source_record_stub()
+    assert Marc.get_related_items(source_record) is None
+
+
 def test_marc_record_missing_leader_skips_record(caplog):
     marc_xml_records = Marc.parse_source_file(
         "tests/fixtures/marc/marc_record_missing_leader.xml"

diff --git a/transmogrifier/sources/xml/marc.py b/transmogrifier/sources/xml/marc.py
@@ -105,65 +105,7 @@ def get_optional_fields(self, source_record: Tag) -> dict | None:
         fields["publishers"] = self.get_publishers(source_record)
 
         # related_items
-        related_item_marc_fields = [
-            {
-                "tag": "765",
-                "subfields": "abcdghikmnorstuwxyz",
-                "relationship": "Original Language Version",
-            },
-            {
-                "tag": "770",
-                "subfields": "abcdghikmnorstuwxyz",
-                "relationship": "Has Supplement",
-            },
-            {
-                "tag": "772",
-                "subfields": "abcdghikmnorstuwxyz",
-                "relationship": "Supplement To",
-            },
-            {
-                "tag": "780",
-                "subfields": "abcdghikmnorstuwxyz",
-                "relationship": "Previous Title",
-            },
-            {
-                "tag": "785",
-                "subfields": "abcdghikmnorstuwxyz",
-                "relationship": "Subsequent Title",
-            },
-            {
-                "tag": "787",
-                "subfields": "abcdghikmnorstuwxyz",
-                "relationship": "Not Specified",
-            },
-            {
-                "tag": "830",
-                "subfields": "adfghklmnoprstvwx",
-                "relationship": "In Series",
-            },
-            {
-                "tag": "510",
-                "subfields": "abcx",
-                "relationship": "In Bibliography",
-            },
-        ]
-        for related_item_marc_field in related_item_marc_fields:
-            for datafield in source_record.find_all(
-                "datafield", tag=related_item_marc_field["tag"]
-            ):
-                if related_item_value := (
-                    self.create_subfield_value_string_from_datafield(
-                        datafield,
-                        related_item_marc_field["subfields"],
-                        " ",
-                    )
-                ):
-                    fields.setdefault("related_items", []).append(
-                        timdex.RelatedItem(
-                            description=related_item_value.rstrip(" ."),
-                            relationship=related_item_marc_field["relationship"],
-                        )
-                    )
+        fields["related_items"] = self.get_related_items(source_record)
 
         # rights not used in MARC
 
@@ -992,6 +934,74 @@ def get_publishers(cls, source_record: Tag) -> list[timdex.Publisher] | None:
             )
         return publishers or None
 
+    @classmethod
+    def get_related_items(cls, source_record: Tag) -> list[timdex.RelatedItem] | None:
+        related_items = []
+        related_item_marc_fields = [
+            {
+                "tag": "765",
+                "subfields": "abcdghikmnorstuwxyz",
+                "relationship": "Original Language Version",
+            },
+            {
+                "tag": "770",
+                "subfields": "abcdghikmnorstuwxyz",
+                "relationship": "Has Supplement",
+            },
+            {
+                "tag": "772",
+                "subfields": "abcdghikmnorstuwxyz",
+                "relationship": "Supplement To",
+            },
+            {
+                "tag": "780",
+                "subfields": "abcdghikmnorstuwxyz",
+                "relationship": "Previous Title",
+            },
+            {
+                "tag": "785",
+                "subfields": "abcdghikmnorstuwxyz",
+                "relationship": "Subsequent Title",
+            },
+            {
+                "tag": "787",
+                "subfields": "abcdghikmnorstuwxyz",
+                "relationship": "Not Specified",
+            },
+            {
+                "tag": "830",
+                "subfields": "adfghklmnoprstvwx",
+                "relationship": "In Series",
+            },
+            {
+                "tag": "510",
+                "subfields": "abcx",
+                "relationship": "In Bibliography",
+            },
+        ]
+        for related_item_marc_field in related_item_marc_fields:
+            related_items.extend(
+                [
+                    timdex.RelatedItem(
+                        description=related_item_value.rstrip(" ."),
+                        relationship=related_item_marc_field["relationship"],
+                    )
+                    for datafield in source_record.find_all(
+                        "datafield", tag=related_item_marc_field["tag"]
+                    )
+                    if (
+                        related_item_value := (
+                            cls.create_subfield_value_string_from_datafield(
+                                datafield,
+                                related_item_marc_field["subfields"],
+                                " ",
+                            )
+                        )
+                    )
+                ]
+            )
+        return related_items or None
+
     @staticmethod
     def get_main_titles(xml: Tag) -> list[str]:
         """