Merge pull request #208 from MITLibraries/TIMX-355-control-field-index

Address liteary form control field bug

Author: ghukill

pyproject.toml

@@ -4,7 +4,7 @@ line-length = 90
 [tool.mypy]
 disallow_untyped_calls = true
 disallow_untyped_defs = true
-exclude = ["tests/"]
+exclude = ["tests/", "output/"]
 
 [tool.pytest.ini_options]
 log_level = "INFO"

tests/sources/xml/test_marc.py

@@ -1349,6 +1349,16 @@ def test_get_literary_form_transforms_correctly_if_char_positions_blank():
     assert Marc.get_literary_form(source_record) is None
 
 
+def test_get_literary_form_returns_none_if_control_field_too_short(caplog):
+    caplog.set_level("DEBUG")
+    source_record = create_marc_source_record_stub(
+        control_field_insert='<controlfield tag="008">220613s     '
+        "|||||o||||||||||||d</controlfield>",
+    )
+    assert Marc.get_literary_form(source_record) is None
+    assert "could not parse literary form" in caplog.text
+
+
 def test_get_links_success():
     source_record = create_marc_source_record_stub(
         datafield_insert=(

transmogrifier/sources/xml/marc.py

@@ -586,10 +586,20 @@ def get_literary_form(cls, source_record: Tag) -> str | None:
         and Leader/07 (Bibliographic level) contains code
         a (Monographic component part), c (Collection), d (Subunit),
         or m (Monograph).
+
+        If control field 008 is shorter than 34 characters, return None as we cannot
+        accurately determine.
         """
         leader_field = cls._get_leader_field(source_record)
         control_field = cls._get_control_field(source_record)
         if leader_field[6] in "at" and leader_field[7] in "acdm":
+            if len(control_field) <= 33:  # noqa: PLR2004
+                message = (
+                    f"Record ID '{cls.get_source_record_id(source_record)}' has less than"
+                    "34 characters for control field 008, could not parse literary form."
+                )
+                logger.debug(message)
+                return None
             if control_field[33] in "0se":
                 return "Nonfiction"
             return "Fiction"